Apex Legends uses a friction system to create "momentum." Mechanically:

• Sprint velocity converts into a low-friction movement state
• Friction coefficient drops sharply
• Downhill slopes amplify velocity
• Jumping preserves slide momentum

This creates the famous loop: Sprint -> Slide -> Jump -> Re-sprint -> Slide. It also means Map Size / Avg. Character Speed = Traversal Time (TT), where higher TT means more combat instances, as the "zones" on the squad collapse on high and non-empty loot locations.

The game has built on momentum over time, creating more "flow" states by allowing players to chain speed boosts after correctly timing a mini-time event after vaulting an object. The game still touches its core mechanic more than 5 years after launch! Even its newest character throws "slide boosts," making the action hero's flow slide even more central to the game.

Its recent feature allows respawns on death boxes rather than the banner-retrieve -> revive point loop that the game first made famous. It's another accessibility move at a time when the game is at its healthiest point since 2024.

This IS the optimal strategy in a portfolio (or any ad market): pair the highest-relevant LTV game to the ad viewer. Remember, "those who win the auction, win the market."

The question has always been what happens on the UA side, and match games regularly hit D365-720+ payback windows. The "eligible" player pool has been tapped out, and thus, we see lower revenue per install in Royal Kingdom than in Royal Match: they are acquiring marginal users.

The stacked revenue chart is key: Match displaced by Kingdom, but total portfolio revenue hardly tells the story that looking at each game does. A prima facie read says that Dream spends $200m in launch budget to maybe add $5m in incremental portfolio revenue. If that's the best a titan can perform, with every Xscapes rots in soft launch, and Match Villains + Truck Star "husking" poor KPIs, then Match 3 really is dead.

The goal of the game designer is to maximize the return on that wage over a life cycle (i.e., generating new spend or extending the player's current spend trajectory). We also know that unlocking most of the content will increase churn. There's a jump in churn rate as players reach the endgame, so there is a looming cost to pushing players farther down the funnel. The combination of the two results creates a backward-bending labor supply curve. Paying players an unlimited progression amount at the start of the game increases churn, turning the marginal wage into effectively zero (there's no more progression to ration). he player is not rejecting a high wage because they are rich; the game has destroyed the future wage schedule. There is no remaining progression to sell, earn, or ration. On the other hand, giving players only one unit of progression might lead to high churn, as it effectively amounts to zero.

This causes mechanics like Clash Royale's daily chest quantity-limiting system to effectively function as a wage cap. This means that the labor supply drops and shortages emerge, ultimately manifesting as lower engagement. Of course, this makes sense if it rations players in a way that also drives purchases.

The logarithmic wage solutions to progression wage caps make more sense to me than hard caps: diminishing returns; maybe after the 4th box, reward the next box at 10 wins, and the one after that at 30. In some ways, the auxiliary systems already do this: while lucky boxes drop off as a wage, trophies, card achievements, and battle pass remain.

In this same light, a purchase is the difference in payoff between the "free" wage rate free versus the "paid" wage rate. A great example is battle pass, which effectively increases the wage rate per unit hour of progression earned by giving additional rewards. Holding all else constant, the effect of a higher wage is more engagement (up until the backward-bending part!)

It's important to note that many underexamined questions about the game economy concern how much progression comes from which areas of the game; each source of progression pays its own wage per unit of engagement. Should PvP ranked ladder or login rewards trigger more progression? Each affects different audience segments differently, which is why the system's design reinforces the game design intent: higher wages signal what the game rewards engagement with.

Listen on Spotify | Watch on YouTube

We talk about:

• Why mobile gaming revenue is increasingly concentrated in puzzle and 4X games, and what that means for smaller or independent developers.
• The three forces that define the mobile gaming market.
• What the growth of merge means for casual economics, and the coming "Royal Match" moment for the genre.
• Whether hyper-casual mechanics are being absorbed into deeper core loops rather than remaining a standalone category.
• The connection between LTV and UA bidding, and whether gains from better matching accrue to studios or ad networks.
• What Android and iOS install patterns reveal about mobile marketing health and app-store gatekeeping.
• Whether third-party web shops and direct-to-consumer monetization can meaningfully reduce dependence on platform fees.
• When mobile gaming will finally stop being framed primarily through the post-ATT recovery lens.
• How Warbonds represent a shift away from traditional battle passes.

Figma Economy Flow Builder v2.0 is live. This is the biggest update yet, with a ton of bug fixes, cleaner JSON logic, and more robust collision logic. It also adds ChatGPT and Claude Deep Research capabilities (look out for v 2.5 for what else this could do!).

Create and manage complex economy flow diagrams with JSON-powered generation, two-way sync, auto-layout, and customizable styling. Perfect for game designers and economists who need clear, readable flowcharts.

• Added the new v2 economy flow renderer with explicit stages, lanes, nodes, and edges.
• Redesigned layouts to be more compact and readable for large game economy diagrams.
• Improved connector routing so lines stay behind cards and avoid running through node content.
• Added clearer edge styling for normal, value, final, and cross-lane relationships.
• Migrated bundled examples to the v2 schema, including heavier Apex Legends and Rainbow Six Siege examples.
• Improved validation for route/card intersections, final-stage outcomes, child layer ordering, and compact layout bounds.

Download here.

On U.S. iOS, 4X plus puzzle may well pass 50% of revenue this year, while year-on-year growth shows it as one of the few bright spots amid the fall-off in other genres. RPG has been hit hard, and Casino is hurting too. Remember Eric Seufert's move to the middle thesis?

The thesis is looking better now, but there was also an organism-level adaptation or a Kuhnian crisis: 4X not only casualized onboarding, but also casualized art, through games like Tiles Survive and Kingshot. RPG, in the meantime, has been hit hard.

I'm unsure what the sportsbook effect will be, but so far there does not appear to be any obvious decline in disposable income for games.

For all the grumbling about the East's inevitability in the West, Overwatch seems to have struck back, and this may become video games' modern Miracle on Ice story. When considering Blizz.net active users, OW is on top for the first time since Rivals launch.

As I've written about previously, the Stadium mode perks, improvements to Live Ops, and Narrative Focus have all maximized game productivity relative to their content costs. Now, a single character maintains a strong volume of retentive units and a stream of servicable content. The cosmetic system has improved with a wider economy array. Something that even the original Overwatch did well, with things like sprays and voice lines on player banners, and also by modularizing weapon cosmetics relative to proprietary character-only single equip cosmetics.

Marvel Rivals is still a Western IP, but it's built by a Chinese team that also fired its Western support studio. It's hard to ignore the project's core supply-side thesis: releasing characters every 4 to 6 weeks is a sustainable content volume, and, by and large, Marvel Rivals has proved it. That hasn't led to compounding gains, with DAU stuck between 1M and 2M. It's hard to know if the game is generally considered successful, considering scope and IP fees. Prior analysis I've done suggests it's kinda meh, not far and away a smashing success (why this game never released for mobile is puzzling).

Mobile may be susceptible to copying, and ad auction economics may strip the price down to CPI. On the other hand, the markets are not nearly as fluid on PC console, and the Chinese will struggle here. So far, Delta Force and Genshin remain strong breach heads, but that's not nearly the pace of something like 4x growth.

Phillip Black, Christopher Kaczmarczyk-Smith, and Eric Guan talk cozy Pokemon habitats, merge-game economics multipliers, AI labor-market irony, and Edward Castronova's foundational work on virtual worlds.

Listen on Castos | Watch on YouTube

We talk:

• Pokopia as the Pokemon cozy game:
  • short production chains
  • daily-quest systems without free-to-play monetization pressure
• Multipliers as the key merge-game innovation:
  • a gas pedal on spend per hour
  • faster energy drain
  • faster story progression
  • a new way to price acceleration
• Narrative as reward subsidy or tax:
  • story can make the next meta milestone worth chasing
  • story can also break flow for players who just want the core loop
• AI and the game-industry labor market:
  • layoffs look more like a post-2021 correction
  • longer unemployment spells in information work may be the cleaner AI signal
  • AI may suppress hiring before it shows up as direct separations
• Castronova's virtual-world economics:
  • challenge
  • labor-leisure tradeoffs
  • property rights
  • platform dictatorships
  • price controls
  • why MMOs looked like the future in 2003
• The player contract:
  • games rarely grant formal property rights
  • players still behave as if they own skins, items, and progress
  • developers often compensate players even when the legal right is weak

There's a serious question: do the metas matter in any of these casual games? Remember the model for both match and merge is wed to engagement: playing lots of levels to hit fail screens and convert, or completing lots of merges to draw down energy and trigger another purchase. Anything that inhibits my time dedicated to merging or matching effectively reduces this equation and, in turn, revenue.

Spending time decorating in Gardenscapes or completing townspeople's dialogue in Gossip Harbor lowers revenue in this model! This lesson hasn't been lost on the hypercasual crop of titles, as saga maps and metas have been chipped away at in future designs on this basis. Many games will simply have a home screen with a level number (see: Toon Blast).

There's certainly something to be said for these narratives, acquiring users, but do they really matter as much on level 1000 as they do on level 1? Do the benefits of the dialogue tree really outweigh the benefits of more level-end or merge attempts? It's a tough sell, and so is the argument that players are completing the level ends to begin with so that they can unfold more of the narrative. It's certainly a testable hypothesis, yet it's often the narrative supporters who shy away from examining it this way.

There are middle grounds that I'm fond of, too, like letting players default to skipping dialogue or choosing one of the decoration options automatically, similar to auto-battle and squad RPGs. Even something as simple as increasing animation speed for anything done outside of the core matching or merging activity.

The other design puzzle is: what if there is a way to save meta and make it more productive? In many ways, they already feel nonsensical. I'm completing an endless series of food orders for the townspeople, yet they keep coming back asking for more. There is very little gratitude. Instead, why can't I match animal food to feed strays? The meta doesn't give the player a core purpose, identity, or aspiration. Some of the relationships in Gossip Harbor and the newer match/merge games are starting to play with this, but the integration feels adjacent rather than dovetailed.

Division: Resurgence is good! Ubisoft and Level Infinite did not ship some broken embarrassment; it's a polished, high-fidelity mobile looter that replicated, and actually IMPROVES on The Division. Yet, it's a KPI disaster, and a waste of four years of development. For all hope lost, there's a pivot that might make sense.

Some years ago, I wrote that Call of Duty Mobile posed a puzzle no one had solved. CODM tempted everyone into believing Western mobile shooters had crossed the chasm. Maybe virtual dual analog sticks were finally good enough! Maybe Western players had simply needed one giant franchise to bless the format! Maybe the rest of the category would follow! All wrong.

On China iOS alone, Delta Force posts roughly $429M over the trailing twelve months. Arena Breakout adds another $39.6M. Meanwhile, on U.S. iOS, Call of Duty Mobile still does about $130M, and everything else is a waste of time: among top franchises, CODM accounts for about 86.8% of revenue.

China did not "solve Western mobile shooters." China solved mobile shooters in China! That sounds obvious, but the industry keeps refusing to absorb it. The ultimate reality is the Chinese market, where mobile is still much closer to the primary gaming device, where the genre has long-standing roots, and where players are willing to tolerate the ergonomics and session patterns that Western shooter fans mostly are not. Consoles were banned in China, and gaming cafe culture is a response to PC gaming expense and constraints that the West doesn't have.

I don't think all hope is lost here because Level Infinite will be far better at live-opsing the title than Ubi's HD attempts. Any Western penetration attempt should instead be framed on PC in the same way Diablo Immortal launches on BlizzNet. Given that the title is built in Unreal 5 and the acceleration of mobile fidelity, I actually think this is becoming a far more feasible strategy.

While that's a long shot, if you are building a mobile shooter around Western IP, your only credible mass-market strategy is China-first. If you do not have a strong China plan, then you might as well light up the gas grill and start burning cash.

So now, the puzzle persists: how has CODM invented, but not expanded, Western mobile shooters? It was the exception that fooled everyone into thinking the market had already arrived.

Last week saw the annual parade of Valley leaders proclaiming the importance of games and the fundamental business lessons they supposedly teach. Apparently, games like StarCraft are hidden MBAs, training players in resource allocation, thinking on their feet, and a variety of other hand-wavy, pro-MBA attributes. Pavel Durov claims, "These games taught me planning, timing, resource allocation, and risk management." Brian Armstrong tells us that games help him realize, "that entrepreneurship and business is the ultimate game."

Of course, this is all just a bunch of horseshit. Valley leaders post to be ironic. If that's not the case and I'm wrong, then the lesson is very simple: aggressively recruit and hire top-tier StarCraft players. Or, even better, fund more game universities. Maybe encourage it to be taught in MBA programs, given that half of these leaders sit on university boards and yet do nothing.

While Valley leaders should get away with a slap on the wrist, the real disappointment lies with the industry's own leaders. Tim Sweeney made an extremely generous preservation of North Carolina's forests, but he's invested jack diddly squat into building out a games program at N.C. State. At this point, I'm convinced @Mark Pincus doesn't even care about games and more or less tripped into Zynga. He rarely talks about them, instead focusing on right-wing politics. I'm curious to read his new book: does Pincus pick up as someone who actually bleeds for games? I have a funny feeling that won't be the case, but I hope I'm wrong.

It's more surprising that we don't see Chinese firms like NetEase and Tencent investing in game universities. If nothing else, building a strong hiring pipeline would put downward pressure on wages while helping combat China's skyrocketing youth unemployment.

This is why I wrote earlier this year about the loss of @Jason Citron as CEO of Discord - he was "one of us." We know this because he wasted millions building a MOBA, a practical rite of passage for anyone who actually cares about making games.

A $100k annual salary in USD is not unheard of, nor is $1M USD for star employees. Competition for game talent is fierce, which is why the whole Lilith-FunPlus-who-knows-who-else-start-a-game-as-a-company model is emerging. In practice, this is structured at the level of an individual game: employees receive equity-like stakes in a specific project, often with participation in dividends tied directly to that game's performance. Not only does it let employees share in the success, but it's also a retention mechanic.

Of course, rising wages mean a changing comparative advantage: it is simply better to outsource parts of development, with Vietnam playing a key emerging role. All real, as far as I can tell.

2. No, you don't need a VPN to access global Steam

Personally confirmed. Everyone I asked was about as confused as I am. How is it that the CCP is allowing Chinese users to access the global version of Steam without a VPN? I got about three different origin stories or speculations on why this is the case, but nothing lined up.

This bypasses requirements such as having a domestic publishing partner and even brushes up against censorship concerns. It's even more striking when you consider that the CCP takes many board seats on key firms, which should be crying foul for protectionism from foreign distributors. Don't forget this is also the same Politburo that called gaming "opium of the people."

3. Rising video card and RAM prices will hurt China's domestic HD demand faster than they will in the U.S.

Wages may be rising in the game industry, but in China, growth has started to slow, while the United States continues to see far greater real wage gains. 4X may be dominated by Chinese publishers, but it is increasingly dominated by U.S. consumers.

On the margin, AI chip demand will start to affect domestic demand for high-fidelity games. All this suggests is a doubling down on mobile and lower-quality graphics.

4. Chinese firms openly access Western AI models via VPN

The CCP has stepped up packet sniffing, and simple NordVPN or ExpressVPN no longer reliably gets you through the Great Firewall. Despite this, nearly all firms openly use VPNs to access virtually everything, including Western AI models.

While much of the conversation around, say, Anthropic's models has focused on early access being used to penetrate Western security, this is already effectively guaranteed to happen on a lag via VPN access. We'll see the same dynamic in game development. The conversation, particularly on the Patel-Jensen podcast about chip controls, is starting to ask whether AI VPN controls are far more important.

5. Chinese economic policy looks a lot more like Sweden

At a high level, the system resembles a blend of welfare state and industrial policy, and the comparison becomes clearer when you look at how labor and capital are taxed. China's top marginal income tax reaches 45%, while employer-side social contributions typically land somewhere in the mid-teens to mid-twenties depending on the city. Dividend income, by contrast, is often taxed around 20%.

Sweden follows a similar shape, but generally pushes harder. Top marginal income taxes run higher once municipal and national layers are combined, and employer payroll taxes sit just above 31%. Dividend taxation can also land around 20% within certain limits, though the system becomes more punitive outside them. The result is that Sweden likely taxes more, but not by an order of magnitude, and the structure of the system feels surprisingly comparable.

Where the analogy becomes more interesting is in state capacity. Both systems are capable of actually executing on policy. Sweden does it through a rules-based, predictable framework; China does it through a mix of central direction and local discretion. The mechanisms differ, but the outcome is similar: when the state decides to prioritize something, whether it's industrial upgrading or social stability, it tends to happen.

In that sense, China can look like a lower-tax, less uniform version of Sweden, one that trades consistency for flexibility while still maintaining a high degree of coordination between the state and private firms.

6. The high-speed train should be branded like a dragon

Much has been said about China's rapid expansion of high-speed trains, particularly the Fuxing high-speed trains. Yet they seem to have missed the most obvious win: redesigning the nose of the train so it looks like a dragon. It's a free brand marketing win and a reinforcement of Chinese national identity.

Deductively, the most useful definition is "the unfolding of a game's systems design." It captures a wide range of experiences and centers the system's design as the key progression element. It also does not presuppose progression as any form of "linear" movement of elements.

These games don't have a cognitively intensive core loop like Clash Royale or even a 4x game. Instead, action "unfolds" with largely illusionist player agency. This grants long lifecycles, and steady engagement - a good deal of research indicated that match player maintain more fixed gaming budgets more than their midcore counterparts. Simply reduce friction to prevent any risk of losing an LTV stream. These days, the first 100 match-3 levels will barely hit 1.1 attempts per success (APS), with bailout difficulty easing on subsequent attempts.

I pin similar cognitive resource requirements to the struggle to elicit the "other" puzzle category. Tile/Match 3D looks like they'll be one or two category winners, but on a casual GDP scale, it looks irrelevant compared to Merger's real potential to overtake match revenue and become a plurality by the end of 26'.

It's also something that appears to be affecting Pixel Flow, which glows up Scopely BD's performance-based aptitude.

Listen on Spotify | Watch on YouTube

What to do, what to do about Marathon? And how about that Fortnite extraction mode? The crew considers radical shock therapy with surprising answers and solutions. Deadlock pushes systems design to its logical extreme in a few shooters that are just starting to play with, but does it actually have an audience? Take bets now.

Shooter Monthly #6 is here with Chris Sides, Christopher Anjos, and Feras Musmar.

We talk about:

• Deadlock's systems-first design philosophy: layered stat architecture across weapon, vitality, and spirit; scaling curves that turn every component into a progression vector; and the role of the Action and X-Y-Z plane crosshairs.
• Marathon's core tension, retention without acquisition: strong engagement signals versus weak top-of-funnel growth; why extraction shooters remain a constrained TAM; Arc Raiders reframing the genre through accessibility; and why tuning economy variables will not solve an acquisition problem.
• Extraction as a product, not a feature: why the loop requires full-system commitment, not a side mode; the importance of PvE readability and social cooperation; and session design, server economics, and continuous player insertion.

It's just that our supply chain location has transformed. Epic's recent layoffs don't change that, although it is a blow to one of our shining stars.

I've seen a couple of posts about it online and messages about on Joost van Dreunen's recent Substack piece, "Epic decline". He writes:

"One major contributor now showing its true impact is the encroaching, disproportionate power wielded by platform holders. Over the ten years leading up to 2025, platform revenue, from app stores, console marketplaces, and digital storefronts, jumped from $14 billion to $41 billion, a 191 percent increase. By comparison, game publishers saw their revenue rise from $65 billion to $128 billion, a more modest 98 percent increase."

And cites Roblox's unprofitability (?) as a case example. In fact, so much of that data point is driven by Roblox's growth, one of the few examples we can cite. Roblox is certainly raising its rates and is no open platform (sorry, metaversers), but that's the fee to fuel this growth. The fact that they're unprofitable is merely evidence of their own reinvestment, which they spend an enormous amount on both infrastructure and paying developers.

It's certainly true that if players age up and stay on the platform, Roblox will capture an increasing share of the gaming LTV, but recently it's been on the decline. Bad for platform power!

The more significant reduction has been the introduction of web stores and alternative PC platforms. Remember, even Valve ceded ground, implementing a reverse marginal tax to combat EA Play and UPlay (the only time anyone's written threateningly about either of those products).

Google and Apple have been beaten back in many countries, and D2G is already 25% of revenue for many firms, sometimes more. Even just a 5% pp shave in platform fee is significant! However, that mostly gets reinvested at equilibrium back into marketing. Marginal propensity for marketing or something of the sort. Yet, the companies best positioned to capture this are American firms like Apple, Meta, Applovin, and soon Roblox, with all its new ad tools. Not to mention Discord, Twitch, and one day, Valve.

Cost is cited as a reason for the stunted growth of American development, with NetEase recently pulling back and even closing a NA-based Marvel Rivals studio. On the other hand, that reality is that cost is driven by American comparative advantage wages. The AI boom revitalized San Francisco, and OpenAI recently closed what is reportedly the biggest round ever.

At the top of all value chains sit platforms because they can maximize economic rent and, by virtue of their scale, capture a % of the ecosystem. Advertising platforms are the biggest, but remember this also includes dominate American engine firms like Unity and Unreal. Where will the inevitable future AI-driven game development platform emerge? I take a large bet on the US.

Read More:

• Epic's Fortnite Mishaps Come Back to Bite All of Us
• Laffer's Roblox Platform Econ Take
• Roblox & Now the Alarming Trojan Horse

Over the last 24 months, Match Merge 2 generated about $738 million in United States iOS revenue and about $1.71 billion worldwide. That means the United States alone accounted for roughly 43% of worldwide iOS revenue in the genre. Match Merge 2 is still smaller than Match 3 on this cut of the data, but it is no longer trivial relative to it: Match Merge 2 equaled about 22% of United States Match 3 iOS revenue and about 26% of worldwide Match 3 iOS revenue over the same window.

That is why I think the United States consumer is in the driver's seat. A single country contributing that much of global genre revenue ends up exerting outsized influence over what kinds of content, pricing, pacing, and live-ops structures get rewarded. The supply base may be international, but the commercial feedback loop is still heavily American.

What is striking is that the firms capturing that revenue are mostly not American. Over the last 12 complete months in this fixed iOS cohort, China captured about 58% of United States Match Merge 2 iOS revenue, Israel about 18%, Singapore about 9%, and Finland about 9%. By the latest month in the cut, China alone represented about 66% of United States revenue and about 67% of worldwide revenue. So the genre's demand center is the United States, but much of the supply-side execution is coming from overseas publisher bases.

The pipeline data points in the same direction. On the known-region portion of the last 12 months of releases in the current Match Merge 2 iOS cohort, Asia accounted for 70 new titles and Europe accounted for 31. The Americas contributed 4 and the Middle East contributed 4. In share terms, Asia represented 62.5% of the tracked known-region pipeline, while Europe represented 27.7%.

Put differently: the United States wallet is setting the prize, while Asia is doing most of the incremental building.

That is an important market structure. It suggests the next wave of Match Merge 2 competition is less about whether the genre has proven product-market fit and more about which publisher ecosystems can best translate global production into products that monetize with the United States consumer. This is an iOS-only cut, Match 3 is proxied here with Sensor Tower's Match Swap subgenre, the publisher-country chart is built from the current cohort rather than a historical point-in-time cohort, and the grouped release chart excludes titles whose publisher region is unknown. Even with those caveats, the directional story is hard to miss: global supply, American demand center.

• What do generator level-up drop tables actually buy?
• When can efficient generators offset energy-intensive orders?
• How does multiplier scaling hit a ceiling once output starts clipping?

Generator Level-Up Drop-Table Shifts

The big efficiency lever is the generator's drop table for items of a given level: generator level-ups shift what the generator can roll in the first place.

At generator level [latex]h[/latex], the generator rolls item levels with different probabilities. The expected level for any generation is given by:

[latex display='true']\text{average rolled level from one generation at generator level } h = \sum_s p(s \mid h)\cdot s[/latex]

Here [latex]s[/latex] is the possible rolled item level, and the weights [latex]p(s \mid h)[/latex] change as generator level [latex]h[/latex] changes.

Yet, because each merge level doubles value, the thing that actually matters is the number of level-1 equivalents produced by one generation:

[latex display='true']\text{expected level-1 equivalents from one generation at generator level } h = \sum_s p(s \mid h)\cdot 2^{s-1}[/latex]

This is what generator upgrades are really doing: they shift the drop table upward so that each generation carries more level-1 equivalents, making the generator more efficient per generation and per unit of energy spent. Unpegging this value, as multipliers do, opens a large design space that could push merge-2 further toward the slots domain.

If one generation from that generator costs [latex]e_h[/latex] energy, then:

[latex display='true']\text{expected level-1 equivalents per 1 energy at generator level } h = \frac{\text{expected level-1 equivalents from one generation at generator level } h}{e_h}[/latex]

[latex display='true']\text{expected energy needed to produce a target that requires } D \text{ level-1 units} = \frac{e_h\cdot D}{\text{expected level-1 equivalents from one generation at generator level } h}[/latex]

[latex display='true']\text{expected energy needed to produce one level-}L \text{ item} = \frac{e_h\cdot 2^{L-1}}{\text{expected level-1 equivalents from one generation at generator level } h}[/latex]

That is the efficiency gain from the drop table itself. A higher-level drop table, and thus more level-1 equivalents per energy, lowers the expected energy cost of orders because it gives the player more level-1 equivalents each time the generator fires.

Power Progression

Raising requested item levels directly raises the energy needed to clear the order.

[latex display='true']\text{expected energy needed for one order} = \frac{\text{level-1 equivalent demand in the order}}{\text{expected level-1 equivalents per 1 energy from the required generators}}[/latex]

Pinning one level-1 item to one energy is a foundational merge-economy design choice. If higher generator levels raise the expected number of level-1 equivalents per unit of energy spent, then the economy gains a long-run efficiency layer. These are permanent generator-efficiency upgrades. It is a huge risk for designers to change this, since it affects the player's entire lifecycle: the board doesn't reset like in match-3. Ever.

[latex display='true']\text{orders completed per day} = \frac{\text{daily energy budget} \cdot \text{expected level-1 equivalents per 1 energy}}{\text{level-1 equivalent demand per order}}[/latex]

That is the heart of the power progression economy. If order energy demand grows faster than generator efficiency, progression slows. If drop-table upgrades push more level-1 equivalents into each unit of energy at roughly the same pace that order demand rises, more efficient generators can offset more energy-intensive orders.

This is also where merge starts to look like a power progression economy. Designers can keep raising order demand while moving generator drop tables further to the right. If those two curves rise together, the player can keep seeing higher-level requests while expected energy per order stays roughly constant. If they drift apart, the economy turns punitive fast. This is also how slot economies manage inflation, punishing players for holding assets over time. As order energy requests increase, the purchasing power of energy decreases as each unit satisfies fewer orders, and that means there are fewer coins or less meta progression per energy unit.

Another suggestion that falls out of the model is something match-3 adopted, which is spiking newly released meta content costs, that is, more meta coins to progress each day, then decreasing them as the content ages and the player state increases for new players who experience it.

Clipping & Deadweight Loss

The multiplier discussion gets more interesting once the output ladder stops being infinite. Consider that the multiplier is applied after the generation roll occurs on each generator. In the clean version of the system:

• [latex]1x[/latex] adds [latex]0[/latex] levels
• [latex]2x[/latex] adds [latex]1[/latex] level
• [latex]4x[/latex] adds [latex]2[/latex] levels
• [latex]8x[/latex] adds [latex]3[/latex] levels

That means:

[latex display='true']\text{added levels from multiplier } \mu = \log_2(\mu)[/latex]

[latex display='true']\text{realized item level after multiplier and clipping} = \min(\text{rolled item level} + \log_2(\mu), \text{maximum available item level})[/latex]

[latex display='true']\text{realized level-1 equivalents from one generation after clipping} = 2^{\text{realized item level after multiplier and clipping} - 1}[/latex]

[latex display='true']\text{expected level-1 equivalents per 1 energy with clipping} = \frac{\text{weighted expected clipped level-1 equivalents from one generation}}{\mu}[/latex]

If rolled level plus added levels exceeds the maximum available item level, the item level gets "clipped" to the maximum level. Part of the paid multiplier produces no additional useful output: a deadweight loss in some sense. If clipping pushes the player onto an item level no active order currently accepts, the player can pay for output that neither improves per-energy efficiency nor helps queue progress.

In the no-cap generator level-up efficiency model, level-1 equivalents per energy stays flat. The downward bend only appears once the multiplier starts pushing rolled items past the maximum available level.

The design implications are straightforward. Anytime clipping occurs, the game should automatically refund the wasted share of energy. Additionally, players should be able to turn multiplier behavior on or off for particular generators instead of being forced into one blunt mode.

What the Genre Should Try Next

The deeper generator-side lesson is not just that merge games sell more energy. They sell more efficient energy, and that creates a wider design space than the genre currently uses.

• Per-generator or per-item boosters are better than blunt board-wide boosts because they let players buy down the exact generator where exponential demand is binding.
• Clipping should refund wasted energy automatically.
• Players should be able to turn multiplier behavior on or off for particular generators.
• Merge games should experiment with selling MetaCoins directly, not just energy, cooldown skips, and item shortcuts.
• A battle pass should grant an additional active order slot, because a wider queue increases board efficiency by increasing the chance that current inventory or current work-in-progress lines up with at least one order.
• Merge games should experiment with a parachute after zero energy, where the player re-enters on a cooldown with a limited amount of playable energy. That would push the genre further toward the social-casino safety-net playbook without forcing a full fail state.

Social casino already has two cleaner spend-velocity mechanics worth borrowing. Safety Net gives the player a rebound grant after hitting zero, then re-enables that claim on a cooldown. That makes full balance drains part of the rhythm instead of the end of the session. Star Power rewards players for raising average transaction size, which turns higher spend velocity into a small progression track with built-in crescendo moments. Merge and match-3 should probably borrow more from that playbook than they currently do.

There is also an experimentation layer sitting underneath all of this. If Experiment With Seeds, Damn It is right that seeds are one of the least acknowledged but most powerful design variables, merge games should treat generator drop tables and order-table rolls the same way. Both systems are seeded randomness. Teams should experiment with per-player lifetime seeds or seeded order-and-drop-table regimes, measure which seed families retain best over time, and then promote the strongest seeds into defaults.

Of course, the question of more conditional weighting on order queues remains the most pressing area merge-2 games need to explore.

That became harder to ignore earlier in 2026, when Gossip Harbor outgrossed Candy Crush Saga.

In a match-3 economy, a useful approximation is:

[latex display='true']\text{revenue} = [(\text{levels played} \cdot \text{expected fail screens per level played}) \cdot \text{p of converting after fail screen}] \cdot \text{price per conversion}[/latex]

Revenue grows when players play more levels, see more fail screens, convert at a higher rate after failure, or spend more each time they convert.

Merge games are a radical departure with no core fail states. They replace fail states with an energy-based stop state, where players stop when energy or a production bottleneck impedes them. It ends up looking a lot like social casino, where energy functions as coins while also gating gameplay.

The genre compresses time with multipliers and higher-level drops, and introduces much more price variability through energy refills, cooldown skips, direct item purchases, and occasional board-space relief. Order cost and production are the right place to start because geometric growth is a constraining factor of "merging two." The genre brilliantly manages costs and production, creating deep economic loops in item management. Once that queue slows down, everything else slows down with it: coins, story progress, feature unlocks, and generator progression.

Meta Progress

A useful literal way to write the first step is:

[latex display='true']\text{coins earned on day } t = \text{orders completed on day } t \cdot \text{average coins per order on day } t[/latex]

[latex display='true']\text{meta progress on day } t = \frac{\text{coins earned on day } t}{\text{coins required for the next meta step on day } t}[/latex]

This keeps two things separate: 1) reward flow from orders and 2) meta cost. Merge games can slow progress by lowering reward flow, raising mission coin cost requirements, or doing both.

Exponential Demand

The game functions by merging two of the exact same item to create a new item at +1 higher level.

Orders are collections of particular items at various item levels, which are satisfied by spending an energy unit to generate a single item. This means we can express all orders as the number of level-1 copies required to fulfill an order.

[latex display='true']\text{level-1 equivalent items for order item of level } L = \text{distinct items requested} \cdot 2^{L - 1}[/latex]

Level-1 equivalent means the number of base items the player would have needed if they had to build the requested item from scratch. If an order asks for two Level 8 items, that is [latex]2\cdot 2^{7}=256[/latex] level-1 items of demand.

For a whole generator, add up the level-1 equivalent demand of every requested item in the active order queue that comes from that same generator.

[latex display='true']\text{level-1 equivalent demand on generator in order queue} = \sum_{\text{requested order items from generator}} \text{level-1 equivalent number of order items}[/latex]

So if the queue asks for two bakery items and one Orange Tree item, you do not average them together. You total the bakery demand on the bakery generator, total the Orange Tree demand on the Orange Tree generator, and then compare each generator's own demand against its own supply. These are the valuable constraints and levers in the merge economy.

Increasing a requested item's level raises demand geometrically. If a designer bumps a requested item from Level 10 to Level 11, that is not a small adjustment. It doubles base-unit demand.

Since energy has a hard currency price, and thus a USD price, we can calculate the USD cost per item level on a strictly energy expenditure basis.

The genre's core gameplay system is an exponential cost curve inside order requests. To manage this, the genre introduced multipliers, or an increase in the level of an item generated.

The Multiplier

There are two separate objects to look at: level-1 equivalents from one generation, and level-1 equivalents per energy spent.

Level-1 equivalents means the number of base items one generation is economically worth after accounting for the higher levels it may skip to.

At multiplier [latex]\mu[/latex]:

[latex display='true']\text{energy spent on generation at multiplier } \mu = \mu\cdot \text{1x energy spend}[/latex]

[latex display='true']\text{level-1 equivalents from generation at multiplier } \mu = \mu\cdot \text{level-1 equivalents from generation at 1x}[/latex]

[latex display='true']\text{level-1 equivalents per energy at multiplier } \mu = \frac{\mu\cdot \text{level-1 equivalents from one generation at 1x}}{\mu\cdot \text{energy spent on one generation at 1x}}[/latex]

And for any fixed energy budget:

[latex display='true']\text{generations available from a generator} = \frac{\text{energy budget}}{\mu \cdot \text{energy spent on one generation at 1x}}[/latex]

Under clean linear scaling, mean energy efficiency is unchanged; what changes is the number of taps required to reach the order item level. A [latex]4x[/latex] generation can hand the player what is effectively a much higher-level item, but that higher-level object simply embeds the low-level outputs and merges that would have been required under [latex]1x[/latex] play.

Higher multiplier means:

• fewer generations
• fewer spawned objects (thus fewer merges)
• less board clutter (thus better board efficiency)
• faster burn of the player's energy balance

That last part draws the genre closer to social casino. It makes the active session more expensive in real time, effectively increasing the cost per hour, but with the benefit of increased acceleration. And just like slots, it lets players choose variance. Because the level and item generated from a generator come from a drop table, hitting an [latex]8x[/latex] multiplier on a rarer generated item is a more significant boost in terms of forgone taps.

If the player has some current energy balance and is using generators at a steady pace, then expected time until the player hits the stop state is:

[latex display='true']\text{minutes until 0 energy balance} = \frac{\text{current energy balance}}{\text{generations per minute} \cdot (\mu\ \cdot \text{1x generation energy cost})}[/latex]

The expected value per energy unit may not change, but the player's bankroll gets consumed faster per minute of play.

Energy obviously has an income side too. For the purposes of this model, it's simple: energy regenerates over time until it hits the maximum stored balance and then gets spent back into production. Events, inbox gifts, and side rewards create a much richer energy-income system in practice, but I don't dive into that here.

Cooldown Chains

Not every order is primarily energy-gated; some are wait-gated.

This matters because cooldown chains, like the Orange Tree item merge chain, still pay coins and still move the meta, but they do so on a different efficiency frontier. They themselves will only generate a fixed amount of items until they go on production cooldown, and less hard currency is spent to reset them.

[latex display='true']\text{expected merged item level per cooldown cycle} = \sum_s \text{p of item level } s \cdot s[/latex]

[latex display='true']\text{expected level-1 equivalents per cooldown cycle} = \sum_s \text{chance of item level } s \cdot 2^{s-1}[/latex]

[latex display='true']\text{expected level-1 equivalents per hour from cooldown chain } c = \frac{\text{expected level-1 equivalents from one cooldown cycle}}{\text{cooldown cycle length in hours}}[/latex]

[latex display='true']\text{cooldown-time burden of an order} = \sum_{\text{cooldown chains in the order}} \text{required level-1 equivalent demand} \cdot \frac{\text{time per cycle}}{\text{expected level-1 equivalents per cycle}}[/latex]

If a cooldown cycle drops a bundle instead of one item, the same math still works. Just add up the level-1 equivalents inside the bundle before taking the expectation.

That is why cooldown generators matter even if storage is not the main monetization vector. They change the mix of wait time and energy time inside the queue; they slow the player through waiting rather than pure energy spend, and may price time separately. This is really important for live ops events. Designers may want to charge a different currency price per cooldown generator reset.

This is also how we can speak about the order queue: some orders are mostly energy merge problems (time + merge action), while some orders are wait-time problems (time).

Board Efficiency

Board efficiency determines how much of the player's board actually advances order progress. A board space taken by a Level 10 Dog no one wants reduces efficiency.

[latex display='true']\text{board efficiency on day } t = \frac{\text{level-1 equivalent production that actually advances current orders}}{\text{total level-1 equivalent production created on the board}}[/latex]

If the player creates 100 level-1 equivalents and only 70 level-1 equivalents actually move active orders forward, board efficiency is [latex]0.70[/latex]. When the board is clogged with low-level scraps, blocked merges, side-chain junk, or sold throwaways, board efficiency falls below [latex]1.0[/latex]. When the board is clean and current production lines up with the queue, it moves closer to [latex]1.0[/latex].

That is the same pain players are describing in threads about boards being filled with generator pieces or always full. Board efficiency is also a function of the relative items the order queue is asking for. As board efficiency rises, a larger share of the merges the player makes actually advance active orders, so expected order completions per merge rises too.

The board size is 9 x 7 or 63 tiles. If we assume one item per order, fix the item order level, set the number of distinct item types to 5, and assume all 63 tiles are occupied by items, we can show how additional orders increase board efficiency.

Order Completion

With those pieces in place, order completion is just constrained production:

[latex display='true']\text{orders completed on day } t = \text{board efficiency on day } t \cdot \min_{\text{across required generators}} \left( \frac{\text{level-1 equivalent supply produced by that generator on day } t}{\text{level-1 equivalent demand placed on that generator on day } t} \right)[/latex]

In the baseline case:

[latex display='true']\text{level-1 equivalent supply from generator on day } t = \text{energy spent on generator} \cdot \text{expected level-1 item equivalents per energy spent}[/latex]

This is just the same generator idea as above. If the order queue asks for several bakery items and one Orange Tree item, those are different generators, and each generator gets its own level-1 equivalent demand total. This is a better way to talk about variety.

Variety does not just mean more items, it means more generators in the denominator, which raises the odds that one weak generator throttles order completion, and thus slows the queue. The order slot cannot turn over to new item requests that reward gold and clear board space.

It also makes the bottleneck logic explicit. The player progresses at the speed of the most constrained generator in the item order request set, rather than the speed of the average generator.

The Queue

Orders do not arrive one at a time in a vacuum. They sit inside an active queue, and mostly each open slot rolls a replacement order after it is cleared.

Based on the Gossip Harbor research, the standard order queue does not appear to look at the live board and then generate custom counter-orders. But the system obviously has to know which generators the player has unlocked, because new players cannot be served impossible requests.

So the clean claim is not "fully random" and not "fully personalized." The clean claim is that merge games use progression-segmented global order tables. As meta progression advances, new generator levels unlock and the relevant order-weight table might change with them.

[latex display='true']\text{order progress already sitting on the board} = \frac{\text{level-1 equivalent items on the board that satisfy the order}}{\text{level-1 equivalent items required by the order}}[/latex]

[latex display='true']\text{order progress remaining} = 1 - \text{order progress already sitting on the board}[/latex]

This is the simpler way to think about the queue: it changes board efficiency through order progress already sitting on the board. If the queue lines up with what the player already has on hand or is already producing, more of the board's output counts as useful progress and less progress remains. This also means efficiency increases as the number of order queue items increases. If the queue item requests become more exotic, board efficiency drops.

That is why a wider queue can accelerate progression. More active slots means more chances that one open order naturally matches current inventory or current work-in-progress.

It is also the obvious future design space. In match-3, adaptive drop changed the economy by reading the board state and steering supply. Merge games could do the same thing by reading the board before rolling replacement orders. If Gossip Harbor Reddit threads are correct, the standard queue does not seem to be doing that yet, but it is an obvious place for the genre to go.

Even without board reading, the queue still reshapes session economics.

The Exploration Space

This leaves a design space that feels underexplored. If some merge games are not already doing it, they should probably play more with per-item or per-generator energy boosters than with blunt board-wide ones. Those are more precise because they let the player buy down the exact generator where exponential demand is binding. Why can't I boost my Coffee production for a set amount of time?

The more advanced generator-side questions sit one layer deeper than this post. I am saving drop-table upgrades, power progression, and multiplier clipping for a follow-on piece.

Putting It Together

The combined daily model is:

[latex display='true']\text{meta progress on day } t = \frac{\text{expected coins per order on day } t}{\text{coins required for the next meta step on day } t}\cdot \text{board efficiency on day } t\cdot \min_{\text{across required generators}} \left( \frac{\text{level-1 equivalent supply on that generator}}{\text{level-1 equivalent demand on that generator}} \right)[/latex]

Progress accelerates when orders pay more coins, when the next meta step is cheaper, when the board's items progress the current order item request, when the drop rate of item levels inside a generator increases, or when the weakest required cooldown generator increases production.

The genre monetizes the spread between exponential order demand and the player's ability to compress that demand through generator power, board space, cooldown management, and time. As a fundamental departure from match-3, things like Super Light Ball and Toon Blast Disco Ball have yet to be discovered in the genre.

• Settling with Google + Apple resolution (mobile needs to win: see Roblox)
• Adding MTX to islands
• V-Bucks price increases (regional, inflation adjustment, but uncommon)

If Tim is taking this company public (Disney acquisition is the failsafe), he needs a better balance sheet. IPOs are about showing explosive growth, not shrinking margins. The press release also suggests they've been running at a loss, which would be something else if true.

Remember, this also follows layoffs in September 2023, which amounted to 870 jobs, or around 16% of its workforce. We can infer from that announcement that this is around 20% of the current workforce, or a fifth of the company, potentially even a fourth.

The press release mentions other savings that will amount to $500m+ on the P&L, which is a number specifically mentioned, so investors take note. Even if half of that $500m came from the 1,000-employee headcount reduction, that would imply $250k per employee. Not cheap!

Especially confusing, since so much of the 2018 headcount explosion was about setting up and finding talent outside Cary, North Carolina, with a strong European operation. Like Scopely's building out of Barcelona to prepare for an IPO that never happened, I wonder if Epic would have been better off building a European-based HQ2.

At the center of all this is something I've talked about frequently: it took Roblox over a decade to build what it has today. The lack of allowing islands to monetize means the UGC flywheel, in which creator revenue attracts more creators, which attracts content, which attracts players, which attracts wallets, which attracts more revenue, has never been able to launch. Only just this January have they even started on this monumental task.

For every $1 spent on user acquisition, how many dollars of gross revenue does a game need to generate to break even?

The recent Twitter brouhaha over paid acquisition has not been kind to venture capitalists, but it also reveals a misunderstanding of the LTV model. Bill Gurley's excellent piece, "The Dangerous Seduction of the Lifetime Value (LTV) Formula", is really an attack on sloppy usage of the model, not on the model itself.

All businesses run on LTV > UA; the question is how literal and useful it is. Gaming is one of the cleanest cases because firms literally calculate it.

The Three Horsemen

The three horsemen of game LTV are platform take, payback window, and the discount rate. The first cuts your revenue immediately, while the second and third cut it over time.

1. Platform Take

Let the app-platform take rate be 30% and the web-platform take rate be 5%. Developers therefore keep 70% of app revenue and 95% of web revenue.

[latex display='true']\begin{aligned}\text{Platform revenue share rate}&= (1-\text{App platform take rate}) \times (1-\text{Web revenue share}) \\&\quad + (1-\text{Web platform take rate}) \times \text{Web revenue share}\end{aligned}[/latex]

If we plug in a 30% app-store take and a 5% web-store take, that collapses to:

[latex display='true']\text{Platform revenue share rate} = 0.70 + 0.25 \times \text{Web revenue share}[/latex]

The first horseman tells us how much of gross revenue the game actually keeps before we even start talking about time. Even before discounting enters the picture, app-only economics already force a gross-up: if the developer keeps only $0.70 of every gross $1, it needs about $1.43 of gross revenue just to recover $1 of UA spend.

2. Payback Window

Even if two games generate the same gross revenue, the faster one is worth more. Revenue collected quickly can be reinvested quickly.

To keep the first pass literal, assume revenue arrives evenly over the payback window. That is a simplifying assumption, not a claim for every title:

[latex display='true']\text{Time discount}(\text{payback days}) = \frac{1}{\text{payback days}}\sum_{day=1}^{\text{payback days}}\frac{1}{(1+\text{Treasury rate})^{day/365}}[/latex]

The longer the payback window gets, the larger the penalty from delay. Holding the Treasury rate fixed at 4.4%, the time-discount factor is about .98 over a 365-day window and about .96 over a 730-day window. Put differently, a dollar collected evenly over a year is worth about 98 cents today; over two years, it is worth about 96 cents.

Hold the Treasury rate fixed at 4.39% and vary only the payback window, and the break-even curve looks like this:

3. Treasury Rate

The third horseman is the discount rate itself. A clean public benchmark is the U.S. Treasury market.

On March 20, 2026, nominal Treasury yields were 3.88% at 1 year, 4.01% at 5 years, 4.39% at 10 years, 4.97% at 20 years, and 4.96% at 30 years. The same date's 10-year real yield was about 2%. The nominal series is available in the official Treasury table, and the real series is in the Treasury real-yield table.

Nominal Treasury yields already embed expected inflation, so we do not need to add inflation separately inside the formula.

Hold platform mix fixed and vary only the nominal Treasury yield, and the sensitivity looks like this:

Putting the Model Together

Together, the three horsemen feed into the UA multiplier, or how much gross revenue a dollar of UA spend needs to generate at a minimum.

[latex display='true']\text{Revenue kept today rate} = \text{Platform revenue share rate} \times \text{Time discount}[/latex]

[latex display='true']\text{UA multiplier} = \frac{1}{\text{Revenue kept today rate}}[/latex]

[latex display='true']\text{Break-even gross revenue} = \text{UA spend} \times \text{UA multiplier}[/latex]

Take the simplest app-only case with a 365-day payback window and $1 of UA spend. The platform revenue share rate is .70. The time discount is ~.98. That means the revenue kept today rate is ~.69. So the cohort does not break even at $1 of gross revenue. It breaks even at about $1.46.

No empirical dark arts needed! The game pays for acquisition up front, gives a platform a cut, and collects revenue over time. The model's job is simply to translate those frictions into the UA multiplier the cohort must clear.

What the UA Multiplier Looks Like Right Now

Using the nominal 10-year Treasury at 4.4%, the scenarios look like this:

App-only at a one-year payback window lands around 1.46x. Stretch the payback window to two years and it rises to about 1.49x. Move 20% of revenue to the web and the hurdle drops to roughly 1.36x to 1.39x. Whether those platform-fee savings later get competed away into higher UA bids is a separate question.

Platform take is doing most of the taking, so web mix and margin matters. Moving just 20% of revenue to the web lowers the multiplier by roughly ten cents per UA dollar, which is a bigger swing here than extending payback from one year to two.

Listen on Spotify | Watch on YouTube

Extraction shooters are finally confronting their identity crisis.

Is Marathon the answer, or just another branch of the same lineage? Christopher Anjos, Chris Sides, and Phillip Black break down pivotal moments for shooters: Highguard's collapse, Marathon's high-stakes launch, and Overwatch's unexpected resurgence. The debate rages on: are we watching the evolution of battle royale into something more persistent, or just another iteration of Tarkov's legacy?

We discuss:

• Highguard's shutdown and the limits of modern marketing
  • Hate-playing, sentiment collapse, and losing control of narrative post-launch
• Marathon's design thesis
  • Persistent battle royale vs. traditional extraction shooter lineage
  • Lower-friction loops, faster re-entry, and softened risk systems
• The genre split
  • Arc Raiders as first-principles PvE extraction
  • Marathon as PvP-first with extraction layered on top
• The missing innovation question
  • What is Marathon's true "Bungie leap" beyond sandbox combat + loot?
• Overwatch's resurgence
  • Systems-driven live ops vs. content arms race with Marvel Rivals
• The future of shooter ecosystems
  • Extraction, hero shooters, and the erosion of the "black hole" game

Read More:

• Marathon Is the Next Evolution of Battle Royale
• Marathon, Extraction Shooters, & The Big Shooter Mistake

Phillip Black, Eric Guan, and Christopher Kaczmarczyk-Smith unpack what actually mattered at GDC 2026, and what didn't.

We discuss:

• A sharper critique of industry thinking
• Too many taxonomy talks, not enough opinions
• Why game talks should behave more like economics seminars
• AI's role on the show floor and conference
• Shift from generative art hype to code generation and workflows
• Why survey data understates actual usage and masks revealed preferences
• AI present but muted, Web3 effectively gone
• The collapse of production costs and what replaces them
• Near-zero fixed costs leading to infinite content supply
• Discovery, marketing, and CAC as the new binding constraints
• Why incumbents may strengthen, not weaken
• Ad spend and distribution advantages widening the moat

Links:

• Listen (Castos)
• Watch on YouTube

1. AI Won, and We're Finally Starting to Get It

The reality is that AI is our next and biggest wave of opportunity. There is nothing else on the horizon, and even Roblox looks margin-shrinking. It's AI or bust. It's taken the game industry a very long time to wake up. We've moved far too close to entertainment rather than technology.

From what I saw at the conference, the engineers are starting to play. This is incredibly important because the biggest imagination executives have is making more widgets at a lower cost.

2. Indies Will Lead the AI Wave, They Just Don't Know It

A couple of AI-native games like AiZU & hideout, made by young developers, gave me hope that the industry's creative lion is starting to awaken. It's by young teams with strong engineering backgrounds and close university connections. A segment of the population that should have been making games for the last 30 years, and yet somehow we've missed out on flipping into game devs.

I don't expect the likes of Tencent or Scopely to understand anything beyond "does AI slash this cosmetic content cost in half?" Still, in the innovator's dilemma, it will be the indies who rise and take advantage of the new technology. The Steam bulwarks and internet bandwagoning play a significant role in this movement, but it's a small battle in the face of a very clear tide of war.

3. No One Knows How to Take Roblox Seriously

Roblox didn't have a significant presence at the conference, but some Roblox developers did. While everyone is curious about the platform, the margins are crushing, and it's hard to scale a business. Adam Telfer is right: what to do about Roblox? With such a different audience and content paradigm, it feels like a growing parallel world, and in some ways, maybe becomes the third estate, with HD and mobile as the other two worlds.

4. American Supply-Side is Hurting, but China Demand-Side is Too

I've started to see Chinese expansion less as a form of manifest destiny and more as a desperation for growing markets. In the category of 4xx, for example, while Chinese firms dominate the supply side, the demand side continues to be, and not only continues to be but is increasingly driven by the American consumer. The Chinese presence continues to grow at the conference, at all seniority levels, and it's certainly welcome. The game developer conference is not segmented by region, but the purpose still feels strategic for the nation.

Ultimately, American irrelevance will only occur if growth slows. Sorry, Europeans. It's a truly astounding puzzle how the United States continues to grow despite so many bad policy choices. AI has again placed it at the center of global economic development;

Read these earlier GDC pieces too:

• Four Things from GDC (2024)
• Four Things from GDC (2025)

Despite the protests of Eric Kress, cosmetic economies continue to generate billions in revenue. For something that doesn't affect the gameplay loop, it's awfully peculiar how successful and widespread they've become. But this wasn't by accident: a successful execution in the cosmetic economy requires a deep and thorough understanding of the model and what tugs at each of its parts. We need a fundamental economic playbook for a cosmetic economy.

This is a talk I've wanted to give for a long time, one that culminates in many years of working on cosmetic economies. It is "The Economics of a $1 Billion Cosmetic Economy." Come join me on Wednesday 10:30-11 at GDC in room 2016 West Hall.

The talk breaks down how cosmetics sustain revenue through rarity structures, mirrors and stages, strategic bundling, and expanded identity layers like passports, titles, and player expression systems.

The goal is practical. Attendees will leave with a clearer framework for designing cosmetic economies that drive both engagement and monetization, along with a model they can use to evaluate their own systems.

Extraction shooters inherited Tarkov's pacing: "vendor"-based progression, and 3D social menus with uneven match insertions where players trickle into a map rather than insert all at once. Given that death means losing all gear until insurance slots are earned, softening death's blow means being able to enter the loop again. Extraction shooters are exciting in theory because the stakes are high, but those stakes only feel good when players experience the cycle frequently. If the loop is slow, the risk becomes frustrating. In Marathon, death to a new game spawn might only take 15 seconds, making the re-climb much faster.

The speed extends the biggest genre problem: drama. Battle Royale popularized a sophisticated curve: synchronized entry, initial tension, mid-game lull, and end-game frenzy. The shrinking circle helps keep player density rates constant, as deaths mount. Marathon borrows the Battle Royale philosophy. Players enter the world in synchronized waves, creating early conflict and establishing a shared tempo. Timed extraction points become the crescendo rather than quiet as a few spots emerge, then teams are allowed.

Traditional extraction games rarely follow this structure. Players spawn at uneven times, wander large maps with inconsistent player density, and often extract without ever experiencing a clear climax. The tension exists, but it is uneven and frequently anticlimactic.

Marathon also softens some of the extraction's harsher edges. Respawns, sponsor kits, and clear objectives reduce the punishment of failure and the endless wondering of prior extraction titles.

If the design holds, the market implications are enormous. Extraction shooters have long been treated as a niche genre because their friction limits audience scale. But if the friction comes primarily from tempo rather than stakes, the ceiling may be far higher than expected.

All the meta elements that made Extraction Shooter popular to begin with are also served with out-of-round progression. If the game can also finally master a live-ops curve in the same way Path of Exile did, it's a huge change to the nature of first-person shooters.

Wishlist orders, however, are far behind Arc Raiders, so it's hard to imagine this game doing north of $200m at launch. Which might seem like a tall order, but so much has gone into development. It feels like the goal may have slipped out of the studio.

While Marathon may not be Tarkov's successor, it may be Battle Royale's.

The team, LA-based and largely Call of Duty leadership, ran the Call of Duty playbook to a T when it comes to monetization, selling blueprints, main interface design with a tile UI, and gun attachments. We even see elements of Call of Duty's mastery camo challenges appear. There are a bunch of new elements, too, that made it into the title. Like in-round progression, Field Upgrades, that move players up a linear ability path in a single life. So far, the game has been struggling to find "productive" content, and I'm concerned it may not be able to cover its variable costs each month. If it can't fulfill this equation, it can't sustain a service. Beyond fulfilling some implicit box-service expectations, it's time to pack it up and move on to the next Battlefield. The curious question is why Battlefield Portal hasn't lent a stronger hand.

Battlefield Portal, first introduced in Battlefield 2042, is an answer to one of the long-standing challenges it and Call of Duty grapple with: how to ensure content produced buys enough retention or stacks up in a way that lasts until the next one drops. On a first-principles approach, we can think of the base game as providing a certain number of retentive units. But as the bulk of players hit the end of that curve or repeatedly replay certain loops, active users are whittled down if the ongoing install base and the retentive units of fresh cohorts are unable to replace them at a sufficient rate.

Portal unlocks the ability for players to create their own units of retention, which should be a massive supply-side to combat the content problem. It is the dark horse answer to Western cost profiles; we're outsourcing to the largest possible TAM. This started in Battlefield 2042 with an impressive web-based logic editor, harkening back to the days of Battlefield's Uprise UI. And now, Battlefield 6's Portal includes much deeper engine integration, allowing the movement of actual objects, enabling the recreation of old maps like Call of Duty's Shipment and prior Battlefield maps. It's an old idea that games like Halo's The Forge first started to really play with in shooters, but it doesn't feel like it has been a growing paradigm.

But for all this seems to solve on paper, Battlefield 6's DAU is still dropping, and Portal hasn't been enough to stem it. The question is why? Part of the answer is straightforward: they've had to limit certain XP values. The fear is that players would find easy ways to farm high XP per minute areas. Of course, this has always been a problem, harking back to why players like small maps like Metro with few defined choke points: kills increase, which increases score, which then translates to more XP. It's like going to the gym.

On the other hand, I'm not so sure I've ever believed this is a problem, and perhaps gaming XP is part of the mode's appeal. They've tried to solve this with "Verified Experiences," a curated list of full XP community settings, but who knows what share of Portal DAU that comprises. Instead, a simple set of hard back-end caps can limit the truly crazy stuff, and use that as your wage ceiling, and let players "game" the system. The difference between the hard cap and what players normally earn in the made modes becomes a wage premium. And that's okay, too, for driving players to the mode. As the UI placement is so poor, I actually didn't even know the mode existed after playing the game for a month. This should at least raise the mode K-factor.

Beyond XP fixes, there are still fundamental problems in Battlefield's core loop. For instance, a few players soak up all Conquest kills, leading to highly unequal K/D between players. It turns out players like shooting things and not getting shot.

Additionally, its content is "unproductive" at the margin. For instance, maps are extremely expensive and yet only provide a couple of percentage points in week-over-week retention boosts. Which, when translating to an uplift in marginal revenue, is pathetic. With guns, players only date them for a fling before returning to their main boo. And don't get me started on vehicles that only a very small share of players play. This has also been a recurring problem with class-based content. Distinct classes also effectively reduce your content's total addressable market by 25% if there are four classes.

The addition of those Mastery Challenges and gun attachment progressions also doesn't seem to have done enough to get the game to a steadier player equilibrium quicker. And that's okay, given how much progress was made on this Battlefield. But for the next iteration, Battlefield Studios needs to start going after some of the problems with the franchise.

The game needs to be re-baptized around the team structure and the importance of a single life. Chris Anjos has called this, and what Call of Duty represents is the ability to be an action hero. I think Battlefield provides its own version of this, but so far, that spectacle has brought in a heaping load of box sales but little retentive units. Elements like body dragging, which we were never able to execute in earlier versions of Battlefield, are done here. It's one of those key moments that feels just right in the game, and epic to drag a downteam teammate out of fire and revive them.

The game struggles to build a cohesive squad player, and despite the introduction of in-round progression for individual soldiers, spawns still feel empty. I'm still seeing issues like players not saving one another or Medics ignoring you. The element of reinforcements that were starting Battlefield 5 provides an interesting view, where collective Squad Points can summon powerful flamethrower tanks, similar to what it was. The game still feels best when a squad can make deliberate, obvious progress across a session's ruleset.

Simple in-round orders could also go a long way, which, when a squad leader uses Battlefield's original Comma Rose ping system, actually directs player action. If there's anything I've learned over the last five years of game design, it's that simple goals direct player action. There's far too much going on in the UI to expect players to make map reads every spawn, and instead, the game needs to choreograph to help coordinate different teammates.

While the game struggles to do matchmaking at all, given the large server sizes required, simple squad sorting rules and pairing players of similar playstyles together are about enhancing teamwork and achieving group objectives.

From what I understand, the internal forecasts were wild, and RedSec (again, going back to the Warzone playbook) was a repeat of Battlefield V's free-to-play attempt, Firestorm. There's just not enough there to justify its existence yet.

The game's promise of all-out warfare remains the franchise's defining feature, but an addendum to the Battlefield constitution, with a focus on the meaning of the squad, is needed. Directed play will help build and expand the audience, just in time for Battlefield 7, which this team will start exploring.

The key is to use a quasi-experimental method called diff-and-diff, in which a particular country is chosen as the region in which to vary the price. In contrast, another country serves as the control group. The interaction term gives you the causal estimate of the price move, subject to one critical assumption: parallel trends.

Parallel trends mean that, absent the price change, the treated and control regions would have moved together. You check this by plotting the pre-period. If, for example, Germany and France track each other for six weeks, and then diverge precisely when you move price in Germany, you have something close to identification. If they were already drifting apart, you do not.

The opportunity is not limited to base price! Deluxe editions are often priced on vibes, too. Is the correct spread $5 or $15? Does a $10 gap anchor players up, or does it cannibalize the base SKU? Gee, if only we could experiment with finding the answer rather than a 2-hour philosophy session every quarter.

I understand the argument from brand marketers that k-factor growth stems from lower prices, and that's valid, but it doesn't undermine the ability to experiment with price increases. Rather than undercut the argument for experimentation, it's actually an argument for it. And maybe you priced it too high! Maybe these games would have been more profitable with lower prices, too. I don't think that's the case, but that's the whole point of experimentation and science. Plus, even if you do not touch launch price, post-launch optimization remains largely unexplored territory.

By far the most important part of Matthew Ball's 2026 State of the Gaming report, and the one that seems to be more overlooked than it should be, is that gaming is losing ground in the attention war. Time is the only truly scarce resource, and as an entertainment product, our fundamental value proposition is tied to our ability to capture and monetize time. There are two key battlegrounds where there's prima facie evidence we're being pushed back:

1. On the time front, in the living room.
2. Secondarily, in the wallet, for gambling (for another day).

In the bid for Warner Bros., Netflix has repeatedly raised that YouTube, strictly considered as TV viewing, now dominates Netflix. According to Nielsen's Media Distributor Gauge (Jan 26'), YouTube accounts for approximately 12.5% of total TV viewing, compared to 8.8% for Netflix, measured as a share of television screen time. Notice that gaming is never mentioned because we've been immaterial to the living room.

As Eric Kress has pointed out, true console growth has been flat for nearly 20 years. While we celebrate Switch's all-time best platform sales, the challenge is that no one actually plays the Switch; they simply purchase it. Nintendo reports ~139M lifetime Switch units sold, but in its corporate financial briefings, ~128 million Annual Playing Users, defined as accounts that booted a Switch game at least once in a 12-month period. Meanwhile, Sony reports ~123M MAU on PlayStation Network, a 30-day activity metric rather than a 12-month one. Hmmm.

The root of all these problems is the controller. This is where Nintendo made some progress with the Wii and, to a lesser extent, with the Switch. If we want massive TAM, we need a better and more intuitive interface mechanic. The Apple TV controller wasn't it. Obviously. And the TV remote playing C-grade flash games probably isn't, either.

I have fading hope for Netflix here with some of their recent party game announcements. It's a start, but there's also an element in which the defining feature of games, bi-directional entertainment, may be harming us simply by increasing the cognitive complexity relative to omnidirectional content and linear media. There's an answer that's been growing inside Steam fairly rapidly, which is the simulation genre, and to a lesser extent, the expansion of Cozy Games we're about to see in 2026.

These are potential mainstream hits that have so far been constrained to crusty German Steam players and still need to find a way to drive ease of access on TVs. The growth of smart TVs has been a godsend that gaming hasn't taken advantage of, with game makers till stuck in the paradigm of trying to slap the old media onto the new.

While there's an array of startups taking a shot at this, the attempts feel half-hearted and small, not the big, bold bets we need to mount a true counter thesis to YouTube's growing share.

The ultimate reality is that the growth of smart TVs has been a godsend that gaming hasn't taken advantage of.

Related: Gaming’s Best Days Are Behind It

Holy shit, it actually happened. Last week, Gossip Harbor generated more money than Candy Crush Saga. This is a pretty monumental moment in puzzle, and while everyone's focused on Pixel Flow, this has been a story that's been building over the last couple of years. Incredible, explosive growth.

While Match hasn't necessarily shrunk, it certainly hit equilibrium. The Innovator's Dilemma has hit all of the Match firms, as I've previously written about; they've been caught flat-footed, and the Chinese have found their opening.

The inability of these firms to even build out extension products in adjacent genres like Blast or 3D is truly puzzling (pun intended). So much time wasted by King on things like Rebel Riders and the 50 millionth attempt at a mid-core title, rather than focusing on their core competencies.

We've seen this with Krafton, which has finally retreated from throwing investment darts at the wall back to expanding the PUBG franchise. Or Zynga, which went back to Poker and Words with Friends, or Ubisoft, which had some adventures and misadventures in the tropics with Hyper Scape and blockchain. And there's CCP, which keeps thinking that EVE is also a Shooter and VR franchise.

Over and over again, we find firms that, in the words of the great Drake Maye, "Get bored being efficient." I saw this at DICE as well, where people get frustrated with working on the mainline franchises. Suddenly, the new hot game in development is a chance to escape from the corporate bureaucracy and design mandates of the last franchise.

Leaders have not done a good job of managing this. Instead, they need to sell the vision that the mainline franchise is where top-tier designers and product managers can see their vision realized, and that they can remove the barriers to making it a reality. Efficiency is not the enemy of creativity, as we saw with innovations like Super Light Ball and Lava Quest.

For about 10 years, we were doing very stupid things across games: we didn't allow players to re-queue after a session ended. One of the secrets to Battlefield's success is that it puts you on a server that keeps playing the next match, leading to incredibly long session times I haven't seen in other titles.

For reasons beyond me, only in the last couple of years have we seen other games start to adopt this. First it was Apex Legends, then Overwatch. If you are a developer reading this right now and you don't have this in your game, you should implement it immediately. I'd love to show you an A/B test validating its value, but first, you need an A/B testing engine. One day, HD games. One day.

The place where this has the most impact isn't shooters, but games where there's a direct connection between engagement and monetization. In many ways, this has been the secret of slot success: they're able to tune velocity extremely high. It's not just that bet size is variable; the sink speed is high as well. This lesson especially applies to Match 3: less time in the meta, more time matching. When you finish a game, stop returning players to the home screen; move them directly into the next match.

Any game, even those with energy-based monetization, should be using this well. That includes 4X games with consumable troops, where the model should focus on getting players into battles ASAP, burning troops, and getting them to re-queue. Even in games like Magic: The Gathering, one of the smartest moves was introducing Quick Play Draft instead of real-time drafting. Clash Royale is missing on this, with Mega Draft, which takes far too long. They need to set a time limit for choosing a card using a chess-like clock. The same is true for real-money or ad-based games, where the margin increases with the number of games played. Velocity, velocity, velocity.

It is the end of Xbox after a painful decade. With Phil Spencer and Sarah Bond leaving or being fired, it confirms the obvious: Xbox is now mostly a tool for Microsoft's other ambitions. Microsoft's old vision of a Trojan horse in the living room is no longer interesting. Instead, Xbox's only useful function is to advance an AI agenda. Short of that, it will be folded and shut down.

Phil was one of us, and in an industry dogmatic about in-group versus out-group signaling, having someone who was "one of us" felt right. He publicly revealed his Gamertag (P3) and had a gamer shelf of titles and paraphernalia behind him in video calls. But the unfortunate reality is that executives are measured as stewards of capital, and Phil's track record was awful.

The defense against Nadella's AI ambitions would be for the segment to generate compelling profits. Instead of rivaling PlayStation, Xbox was crushed in Europe and never materialized in Asia. Spencer's 12-year reign largely began with Xbox One, which also coincided with the fumbling of some of its biggest IPs.

Gears of War, Halo, and Fable are Xbox's response to God of War, Spider-Man, and The Last of Us, and all three franchises are in the worst place they have ever been, relegated to supporting cast. Game Pass is a strategic miss on many fronts, a model that is fundamentally incompatible with games. While much of this comes from Redmond mandates, Phil's title is Xbox CEO, and at some point the buck has to stop with him.

We should not underestimate Spencer's role in saving us from the fate of industry fool, former Xbox Head Don Mattrick, who seriously attempted to bundle Kinect and force always-online requirements. The fallout of that conference is felt to this day. Backtracking those decisions delayed Xbox in Europe, allowing PlayStation to gain an early foothold. That alone deserves his face etched into gaming Mount Rushmore.

My hope for Phil & Sarah is to do something new and exciting, bringing gaming what it desperately needs: new platforms that bring in new opportunities to grow the market.

Asha Sharma, the new Xbox head, a former Instacart executive and Microsoft AI leader, is not a serious choice if you believe in games as an end, not a means. In a rare communication miss, Lulu Cheng Meservey, former Activision Blizzard chief communications officer, celebrated Asha's initial PR release, which appealed to the gamer base: "No soulless AI," "Games are art," and so on. But in Lulu's own playbook, these statements are corporate slop. The authentic message would be to show her Gamertag.

Short of that, the better move is to lay cards on the table, declare that she is not a gamer, but that she is excited about this world and diving in with a photo of herself playing the latest Halo game. Unfortunately, we know that will not be the case. While it is certainly not a mandate for executives to be gamers themselves, it is a powerful way to win hearts and minds with gamers and the game industry.

Genres like match pair looked like a promising emerging subgenre, but in the end, it became dominated by Match Factory. While that game has sustained an impressive run rate, growth stalled and the genre reached cruising altitude.

Hybridcasual has been experimenting with puzzle templates like sort and screw, but they have not achieved the level of retention we see in match. I pin this on cognitive complexity being too high, which slows velocity, reduces attempts, reduces the number of fail screens players see, and ultimately reduces sink. In many ways, match has become slots, with most early level ranges hitting around 1.0 to 1.1 attempts per success. The secret to match balancing over the last couple of years has been to largely punch down on APS and plan streak formation with hard-level difficulty labeling.

Merge 2 continues to be a smashing success. Teams have been able to pair stronger narrative elements this core demo responds to with new economy mechanics that actually expand spend velocity, something match has struggled with for a long time.

Stay tuned for Puzzle Monthly #1 with Aylin Yazici, David Nelson, Tom Storr, and Laura Taranto, plus an Avengers-style guest roster to come.

If you want to read more about where these subgenres are moving right now, start with:

• Sort Puzzles: How a New Subgenre Is Born
• Pixel Flow: The Publisher's Dream

The recent creator flare-up around Clash Royale was not about a blog post, and like the Highguard snafu, it reflects a playbook that developers have yet to write in managing the Internet. Ilkka's annual letter asked a senior Clash Royale team member to explain their success, and he did not mention creators or streamers. One of the more prominent streamers caught wind of this and thus began an internet tantrum. A classic Reddit-style boycott ensued, and an amendment to the letter acknowledging streamers was posted.

The industry talks about creators as if their role were obvious. It is not. Clash Royale is looking north of 40 MILLION DAU, while peak streaming for Royale sat in the low hundreds of thousands. Even if one generously multiplies that by downstream effects, clips, YouTube edits, and TikTok fragments, we are left with an attribution hole. There is, of course, the causal-chain problem too. Did streaming amplify the popularity of the game, or was the game streamed because it was popular? We behave as though this distinction is settled when it is neither properly modeled nor settled.

What makes the situation unstable is that creators occupy a hybrid economic role. They are not employees, but they are revenue participants through creator codes. By publicly correcting the omission and naming a creator, Supercell implicitly elevated creators from internet anons to the new gaming fourth estate. Responding this way means feeding a fire. We have ended up building a system in which creators and streamers function as quasi-stakeholders while retaining the volatility of independent operators.

An apology was the worst thing they could have done. The internet rewards leverage, and shitstorms (see Highguard). If public agitation yields public acknowledgment, agitation becomes a rational tactic, and it has been for over 20 years with gamers.

Instead, a small separate event, unconnected as a direct response, was the appropriate move. Maybe give a creator a small 2D player card or wrap them in a fictional leaderboard interstitial.

It really does not take much, and as Plato tells us, politics is the struggle for recognition.

Both Supercell's claim in Ilkka Paananen's letter (Supercell did not match their highest years!) and Matthew Ball's The State of Video Gaming in 2026 report completely change when we adjust for conservative global inflation estimates from the IMF. Don't get me started on Newzoo.

In the United States, where much of Western spending is concentrated, inflation has risen 25% since 2020: $1 today buys what $0.79 bought in 2020. Every game has been running a virtual discount if they haven't increased price!

There's an interesting side effect: 4X and games like Monopoly Go are somewhat inflation-protected, since their monetization systems operate like auctions focused on leaderboards. The real price for a top spot is more resilient to inflation than in PvE-dominant games like Candy Crush.

However, the inflation adjustment to currency prices is also far more modest. Epic is one of the few players to raise prices, but in real terms, the increases are nearly half as large as in nominal terms.

Clash Royale ran a Lil Wayne "halftime show" around the Super Bowl, and it should have nailed the butt crack. It's a fun brand activation, and going back to a talk I haven't stopped thinking about @Teut Weidemann's the rhythms of free-to-play, the CCU butt crack occurs at halftime.

CCU is strong before football kickoffs, collapses during play, and then at halftime, a clean vertical spike appears: the butt crack. A perfectly predictable surge as players pick up their phones for a few minutes before the game resumes, and then drop again.

Clash Royale clearly recognized the window: they built a halftime activation. This is the right instinct, but they didn't engineer around the spike. Something that's really perplexed me: the event went live on Friday, not during halftime. There was no moment-specific nudge tied to the surge.

This is where most teams misunderstand LiveOps. It sometimes gets defined as updates, UA bursts, events, offers, yada yada yada. It's taken me a while to realize this since listening to Teut's talk and thinking about mobile user acquisition, but really, LiveOps is about relevancy.

The whole point about "live" is that it captures fleeting moments. This is why so much of LiveOps is about holidays. This is how you immediately become relevant to the player. Lunar New Year, St. Patrick's Day, and Easter: we build events around these because they have a time-limited meaning in a player's life.

The butt crack talk reveals that we haven't considered relevance on a micro LiveOps basis. In mobile, a lot of this is because teams rarely check CCU. It's something that's been reserved for HD games, mostly to ensure it's high enough to sustain matchmaking, while nearly anything on mobile is reviewed on a daily cadence.

With March Madness and the World Cup coming up, it's a good time for mobile teams to plan how they'll structure their LiveOps to cover dead periods and stay relevant to players during those gaps, which is central to the mobile game thesis.

In this episode, Phillip Black, Christopher Kaczmarczyk-Smith, and Eric Guan unpack what this signal means for interactive entertainment, from Riot's 2XKO downsizing to Google's Genie 3 and the future of engines.

We discuss:

• The 2XKO reset and the economics of niche-within-niche genres
• Team size, burn rate, and why a 160-person fighting game team changes the break-even math
• Free-to-play cosmetics versus box-price DLC in a capped DAU genre
• Why betting apps can out-monetize most games on ARPDAU
• How appetite suppression might reallocate time, spending, and loop sensitivity
• Genie 3 and the cost curve of game production
• Engines as rule-governance layers in a probabilistic content world
• Cosmetic economies as foundational theory
• Scarcity, signaling, and equilibrium pricing in digital status markets
• Price discovery, private information, and turning trade into tabletop play

Links:

• Listen on Spotify
• Watch on YouTube

There's a peculiar strain of PM theater that rebrands ideas to make them avant nouveau. MVP became minimum lovable product. I'm sure someone is workshopping "minimum legendary product" as I type. Allow me to contribute my own economist-flavor: the Minimum Falsifiable Product (MFP).

So many of the teams I work with failed to accept or act on the experiment's conclusions. I've watched this play out repeatedly: a team runs a test, but the results don't "work" right. There's a parade of caveats in analyzing the result; a lot of "ifs" and "what's." And the proponents never accept the experiment's clear actionability result. It gets bookmarked as something interesting, and then the team moves on.

This is a failed experiment: no one learned anything, as everyone hunkered down in their turtle shells, too scared to act. It's a colossal waste of everyone's time, and if you're not willing to accept the results of the experiment, don't run it. The way to prep for this beforehand is to only run minimum falsifiable experiments.

What is the minimum set of process design timings, whatever it may be, that falsifies the idea under experiment? Whoever is the proponent of the experiment needs to define this, which backs them into a corner of accepting the results and acting on them once they come through.

Who cares about viable or lovable if the internal stakeholders won't actually action on the damn thing? If you are unwilling to define the conditions under which your idea fails, you are not running an experiment; you are buying optionality for your ego.

The idea is simple: the combination of ARPDAU and DAU produces an annual run rate. Consider each of these dimensions: what is the expected DAU you think you can get based on marketing, market conditions, etc., and what is the ARPDAU you think you can get based on your monetization model and genre norms?

It wasn’t rocket science to see that 2KXO was going to be dead on arrival. The top DAU for fighting games across all platforms is Brawlhalla (platform fighter!), and while that’s a low benchmark, it’s a starting place to think about “hey, what would it mean to 4x or 5x this in terms of DAU?”

Of course, the other dimension is ARPDAU. And it becomes a clearer strike against this game. Even at crazy 3x–4x DAU numbers, ARPDAU still needs to be embarrassingly high to get to the type of results that Riot would expect from such a product. A project doesn't need that sort of entry velocity, but it should at least be benchmarked against average variable cost (opportunity cost is for another day). The burn must have been high, especially in Santa Monica!

The people I’ve spoken with who were at Riot during the game’s development had the same concerns, and the warning flags were ignored. I’ve developed a strong distaste for Riot, and arrogance seems to have bitten them once again.

Whether it was ignoring Tencent’s call to port League of Legends to mobile because their IP was too “esteemed”, or blowing $250 million on Arcane, which I’m now hearing is a number that doesn’t include all costs. Similar to early Blizzard, these firms have become lifestyle companies where there isn’t just an office; there’s a “campus”, and socialization past the end of the workday is encouraged and subsidized. There’s nothing wrong with it, but it fosters a cult mentality.

High Guard’s first problem was sustaining a player base. Still, there needed to be reasonable expectations about when DAU would hit equilibrium, and ARPDAU needed to support that equilibrium given a constant team size. It’s again possible to cover average variable costs, but you need sufficient ARPDAU to support it. With strong design and internal choices, I still believe it's possible to run a high-quality, free-to-play HD game with fewer than 100 developers.

No one I’ve seen has talked about High Guard’s monetization, and as a new IP, this is a perfect opportunity to go back to loot boxes in the same way Apex did, or to move in the Warbond direction of Helldivers. Teams should start choosing monetization models that allow them to operate at lower platform rankings, rather than assuming they can hit the top 10.

The theme of '25 heading into '26 has been the absorption of squad RPG into 4X. Remember, on mobile, those who win the auction win the market, and 4X's endless spend depth and social dynamics grow LTV, outbidding squad RPG. The move to 3D in games like Genshin Impact or Solo Leveling is a counter-cyclical thesis to this. What if RPG just said, "fuck it, we can't win on mobile"? Put the game on HD, and then anything mobile is icing on the cake. Arknights Endfield is in this vein but also solves some of the more fundamental problems squad RPGs have had in sustaining the long run.

The original Arknights is the most important squad RPG, perhaps bar none. Unlike other titles, Arknights maintains a linear revenue curve since its launch in '19, crossing the billion-dollar revenue mark. It never suffered the six-month hangover that Lilith talked about in their famous Chinese post. Studying that is essential if we're going to reverse the squad RPG's decline.

Arknights has started pulling months where it outgrosses Genshin, and by the middle of '26, I expect it to beat it every single month.

Unlike other squad RPGs, Arknights isn't just a power-scaling game. It implements tower defense, in which characters defeat waves of enemies, allowing for more intricate puzzles than simply selecting the right squad. Placement and timing - and the squad itself - are part of the puzzle design. The breadth is vast, enabling far more horizontal meta-collection.

While the game also has a strong narrative apparatus, its standout feature is a base-building system reminiscent of Fallout Shelter. It's not only a character sink; it's also a place for additional puzzles. Configuring the base and determining the optimal resource chain is essential for long-term progression. And while other games like The Walking Dead: RtS played with base-building, Arknights found the right blend.

The rise of 3D has been an attempt to solve the RPG problem by amortizing costs and expanding the audience across more platforms. And this is the same thing we see in Endfield: the game is on mobile and HD, seemingly finding success across both platforms. But it's not enough, and we still see declines in many of these 3D RPGs, like Solo Leveling and Zenless Zone Zero.

The most important element they've carried forward is the base-building. The factory feels far more sophisticated - and harder to execute - in 3D than in 2D, which is why we see fewer 3D simulation games that take a character perspective rather than a godlike one. The tutorials are laborious, but they've made some nice concessions, such as zip lines for quick movement.

It should build a solid long-run economy loop that grows over time while maintaining a more reasonable budget, without an open-world design. So far, the numbers have been good, especially coming off pre-registration highs, with DAU starting to level off across mobile and HD platforms.

He doesn't pussyfoot around Squadbusters, unlike in all his previous posts. What went wrong? What are the lessons? They are clear and direct and actually reflect so much of what we've written about at @Deconstructor of Fun. There's real data too! And my god, who wouldn't kill for ~30% D7.

So much of his post reads like it's DIRECTLY lifted from comments on TWiG: new platforms and new audiences are opportunities to grow, and we've failed to do that in the industry, entering a period of stagflation. PMs' A/B testing of button colors on their way to innovation, or reskinning Royal Match, won't solve this. It's a function of bold new bets, and say what you will about Supercell, I'm glad that they've been able to accept more risk appetite. Yeah, Mo.co's probably a failure, and so was Squad Busters, but they tried something new, and learned new lessons, written in blood no less.

The challenge with Supercell is its inability to move on from bad bets. As they continue to scale as an organization, they become vampires sucking on resources and distracting the company. ¿Dónde estás NORTH AMERICA? Was the Supercell engine a success? And when does the Mo.co timeline end?

The bravest part of the letter and one that actually had real teeth is where Ilkka Paananen lays out what new game creation will take, specifically: profit-sharing, and tight constraints. "I've rarely seen great products emerge from teams with unlimited budgets and endless runways", he writes. I just wish he applied that to internal initiatives too, not just games.

Still, the missing piece is about how Ilkka will innovate on the Super "cell" itself. Initiatives like Sparks Labs or AI Labs feel like they're running away from the problem, trying to have innovation happen outside the Finnish HQ rather than dealing with the very messy problems of "corporate entrepreneurship." I poked at it for fun, but none of this letter felt like someone who's striving to become the world's least powerful CEO.

The financial results were impressive and should assuage some concerns that they were UA bursting Clash Royale. There's also an honest acknowledgment that Clash Royale is now on the decline again, and they need to find their next big beat. Looks like IPO is back on the menu boys?

All newly added to the website! It's also covered in the weekly newsletter, looking at the most prominent companies by market cap, including the Asian Tigers.

I was always getting up to speed on the macro environment, with so many different metrics. It was hard to keep my head on straight and understand the real players. Market cap helps cut through the noise and size who's who

As well as something that doesn't get nearly enough attention: how prediction markets are thinking about key video gaming questions. There isn't much liquidity in these questions, but stay tuned on that one 😈. But it's still curious to see things like GTA's $100 standard skew price plummet to a 1% probability, and Ubisoft's acquisition rate remains 50/50.

More to come.

Markets

Listen on Spotify

The question nobody in the West wants to answer honestly: the biggest shooters in the world are not the ones you argue about on Twitter.

Shooter Monthly #4 is here. Christopher Anjos (DICE), Chris Sides, Feras Musmar, and I talk Highguard and Spectre Divide as case studies for why "ex-Apex devs" is marketing copy, not a strategy, and why in 2026 a shooter does not get a long runway to "fix it live."

We discuss:

• Highguard as a MOBA-raid shooter hybrid, and why its best moments arrive only after five minutes of downtime.
• Defense phase, resource phase, then the game finally becomes itself: mounts, CTF energy, base raiding.
• Shadow drops vs. tech tests: why the Apex Legends playbook does not port cleanly into 2026.
• "Early access" is just launch, and Steam labels do not save you from retention.
• The 3v3 trap: high-risk formats where one leaver or one death warps the whole match.
• Why 3-person squads feel like a couple plus a third wheel, and when 3v3v3 or 5v5 might actually work.
• The "two genres" thesis: Battle Royale and tactical shooters as the market's default equilibrium.
• The creator-credit illusion: "ex-Red Dead lead designer" as a brand signal that often means nothing.

Links:

• Watch on YouTube
• Listen on Spotify

There's a funny bit where during the dawn of Battle Royale pundits asked if it was a genre or a mode. Nearly ten years after PUBG launched, it has not only become a genre but also the dominant force in the shooter market. But it's also clear that, as with Match 3, the market has reached equilibrium: PUBG, Apex Legends, and Warzone have solidified the genre, and Battlefield's repeated failure to gain market share is instructive beyond the corpses of titles like Hyperscape.

When castle walls have been fortified, the answer isn't to build a stronger siege weapon. It's to build a more innovative weapon, and that's what Extraction shooters have represented. Highguard, by contrast, represents an alternative path.

Whereas extraction shooters completely pivot to out-of-round progression - players earn items and gear that persist between runs - Highguard emphasizes in-round progression, wherein progress ramps during in-round, but does not persist between rounds. Battleborn (RIP), played with these elements, and Overwatch 2 has brought them back in Stadium.

It's a complete counter-thesis to the direction Extraction wants to take the shooter genre and instead doubles down on what many are beginning to uncover about battle royale: it's a roguelike packed with RNG.

Practically, Highguard's core loop:

• Select and secure based on the map; each base has +/-
• Fortify the location (Cannot move outside your base) [R6:Siege]
• Walls drop and scavenge phase begins [BR]
• Summon mounts and fight for Vesper [BR]
• Use Vesper to purchase in-round upgrades [MOBA]
• Converge on the Shieldbreaker
• Team who captures it then tries to siege the other player's base, in which the Shieldbreaker reduces the base's energy shields [Capture the Flag]
• Kill the anchor stone, which is located deep within the base (if fail, fight for Shieldbreaker again) [MOBA]

It's a valiant attempt to blend several elements into something fresh, but it needs another year of development. The spacing of the maps leads to periods of downtime, and PvE elements, which should give players a sense of satisfaction and mitigate zero-sum PvP combat, are empty. The power race, which MOBAs are infamous for, feels hollowed out. The power scaling and pacing aren't strong enough, and the missing key out-of-round horizontal progression is a damper on retention.

Highguard is dead. While every small developer hangs on to Warframe and R6 Siege as games that were able to grow over time and ultimately reverse the U-curve, those stories are the exception, not the rule. On Steam, it's already catered to 4,000 CCU, and with Western developers' stubbornness about adopting not just any sort of gameplay-affecting monetization, but RNG associated with it too. Its ARP DAOs will not be strong enough to persist.

There's a whole lot more to talk about this game, including its launch. In Shooter Monthly #4 next week, we will!

Not quite sure what the grandstanding achieves after it's delivered on U.S.-based Linkedin/Twitter/Discord/Bluesky/Substack typed on iOS/Android, after shipping games made with Unity/Unreal to Steam.

The concerns about cost are valid, but this is a recycled complaint every year. And this year, GDC actually tried to bring a lower-priced pass. San Francisco's safety record is embarrassing, but there's rarely an issue in the main business centre, and it's far better than a decade ago.

The most astonishing part about my experience living in Europe is the obsession with U.S. politics, even during the Biden admin. While it is the world's funniest reality TV show, miniseries from countries like France, Germany, and Italy rival some of the best U.S. seasons.

I'm a fan of the Touch Grass Movement and literally the whole point of GDC is to network with industry peers and build community. If anything, attending and embracing the extremely international community of game developers is a protest against the world's trending events.

Seriously, it's one of the industry's most marvelous revelations: it's somehow developed global hotspots. Normally, we'd expect to see more agglomeration effects and knowledge spillovers grow over time, but something about the diversity of games has given it a polycentric structure.

As game developers, we have the unique opportunity to take advantage of this and explore places we would otherwise never have the chance to experience. I'm living proof of this, having lived in Los Angeles, Sweden, and now Cyprus, becoming a dual US and EU citizen.

Amanda Cesario (analytics lead turned product leader) joins Phillip Black, Eric Guan, and Christopher Kaczmarczyk-Smith to argue for embedded analytics, sharper language, and game systems that produce cooperation instead of a cosplay community.

We talk about:

• The missing vocabulary for economy design in live service, and how it has harmed the industry
• Why office ball pits beat startup ping pong tables
• The analyst's real job: explaining "why," then realizing the only way to fix it is to own the lever
• Embedded analytics vs centralized service orgs: who wins?
• Roblox as a laboratory: aspirational visibility, server "neighborhoods," and system norms that communicate more than art
• Iterated Prisoner's Dilemma, Axelrod's tournaments, and why tit-for-tat is a design principle
• Monopoly Go partner events as rare, genuine cooperation-through-repeated-interaction design
• Why Discovery Zone died, but Chuck E. Cheese prints money anyway

Links:

• Watch on YouTube
• Listen (Castos)

What people are trying to gesture at are two distinct questions, and collapsing them into a single number is a mistake.

First: where does the power curve end, and what does it cost to get there? In games without infinite consumables (read: Peace shield, not XP boosters), spend depth is really shorthand for the price of convergence. How much money does it take to reach the end of meaningful progression, or close enough that marginal gains flatten? More importantly, what does the shape of that payment-to-power curve look like?

A healthy curve is convex early and concave late. Early spend feels impactful, mid-game spend accelerates options, and late-game spend buys diminishing returns. When people say "this game needs $10k of spend depth," what they usually mean is that the curve ends too early, so whales run out of things to buy that matter. That's an issue with the shape of the curve rather than the price of items.

Second: is some arbitrary spend constant actually worth it? Even if a player can spend $10k or $20k, the more important question is whether doing so produces outcomes where the costs justify the benefits. Creating a system in which that level of spend feels rewarding is literally the point of much of economy design.

𝗛𝗘𝗥𝗘 𝗖𝗢𝗠𝗘 𝗧𝗛𝗘 𝗖𝗢𝗭𝗬 𝗚𝗔𝗠𝗘𝗦, 𝗦𝗢 𝗪𝗛𝗔𝗧 𝗡𝗢𝗪? As if there was a sudden awakening, firms have realized that Animal Crossing can become a live service revenue juggernaut, and the race is on to figure out who can build it first. There are a handful of these coming out this year (including from Hoyoverse), and Heartopia is the first. In one biggest compliments to be paid to the West, the game lifts 80% of its loops from Singularity 6's Palia. Yet it struggles in the same way, and ultimately, this will limit the genre's ability to grow.

Heartopia is a useful case study precisely because it gets so much right. It launched aggressively on mobile, hit no.1 downloaded charts, and paired a Steam launch. More importantly, it advances the cozy genre where it actually matters: social structure. Neighborhoods with other players are a constant presence! Group quests exist! One player can purchase a group quest in which all players benefit. This echoes the pro-social logic that made systems like Pokémon Go’s lures so powerful. This is a tight, thoughtful execution of ideas that cozy has been struggling to get right.

And yet the economic outcome is familiar. Monetization remains soft and cosmetic-based. Execution is again strong, with tight theming and gacha to govern price, but at the end of the day, it's still non-gameplay-affecting. That matters because ARPU and conversion are not abstract KPIs in free-to-play; they determine how competitively a game can bid in ad auctions. The golden law of mobile remains "those who win the auction, win the market." If you cannot bid, you cannot scale. If you cannot scale, the genre stays aesthetically beloved but economically constrained.

Cozy games are deliberately non-punitive, non-violent, and low-pressure. That also means monetization systems need to do more work inside the game’s social paradigm. One credible direction, not a prescription, is competition. Not PvP in the traditional sense, but Design Home–style logic: asynchronous contests, themed prompts, entry requirements, voting, and rewards. Decoration ceases to be pure self-expression and becomes a matter of eligibility. In these loops, housing items unlock participation, status, and repeatable stakes without turning cozy into a DPS slug fest. Importantly, this preserves the fantasy while giving spending a mechanical consequence.

Heartopia hints toward this future through its neighborhoods and group monetization hooks. The missing step is making those purchases matter more consistently to participation and progression.

The next chapter for cozy is not about becoming harsher or louder - it is about making coziness do economic work. Cozy games do not have a demand problem, but they have a monetization ceiling that quietly limits how big the genre can become.

Western PC/console (HD) developers have been circling the same problem for more than a decade: how to monetize power without Reddit tantrums imploding into infinity (Eric Kress talks fondly about this). Some franchises are grandfathered in, or have grown seeds of no expectations into full-fledged monetization products like FC Ultimate Team and GTA Shark Cards. Meanwhile, Eastern developers face no such expectation, and games like Genshin monetize pay-for-power on HD. Yet, for reasons beyond me, one system gets little study: Call of Duty's blueprint system.

Now in Battlefield 6, the blueprint system "pre-progresses" weapons so players can use certain attachments, while it also bundles a cosmetic. Together, they represent a complete fantasy. Players normally earn these attachments, specific to a particular gun, by simply engaging the gun, but by purchasing a blueprint, they are earned ahead of time. Players are still awarded attachments as they level up, but this differs from a boost system, which applies a subsidy on time. Here, players can directly pay to skip, with no additional labor required.

This falls under the category of horizontal progression, but in a rarely scalable way. They come with trade-offs, not straight stat inflation, even if the aggregate effect of more options is usually more power. And unlike other Western horizontal monetization schemas, such as selling characters, blueprints are far more sustainable, as attachments generally have a low marginal cost to create and are designed for core progression reasons anyway. In many ways, Activision has figured out how to bundle existing content to monetization systems. That is the rare case where monetization scales without distorting the design pipeline. It's brilliant.

This has expanded to another of Call of Duty's understudied systems: its pro-tuning system. While this did not survive to the latter parts of the franchise, adding endgame progression systems to first-person shooters has remained an unsolved problem, and this was a gallant attempt. In this system, players could essentially tune individual stats in a spider-chart-like way, which also opened up the option to use preconfigured pro-tune blueprints themselves, creating infinite combinations. I do hope this comes back in some form or another.

In many ways, what the blueprint system ended up solving was simple choice complexity. Since Call of Duty 4: Modern Warfare, Call of Duty has steadily drifted toward RPG territory and is now firmly planted in it. With an individual loadout consisting of 40 to 50 choices across both gameplay and non-gameplay vectors, it foreshadows some of the RPG-like elements we've seen in Destiny and Extraction Shooters. Blueprints act as a cognitive shortcut. They reduce choice paralysis, offer a baseline of combat viability, and present a coherent theme players can attach onto (pun intended).

The PM's view of metrics tells a similar story. Teams obsess over FTUE completion rate, an endogenous variable! Define "app open" as FTUE, and congratulations, you've achieved 100% completion. If you want to know whether your FTUE works, D1 retention is the actual test. Did players come back after the on-rails stopped holding their hands?

This fixation on completion has warped mobile. Midcore and 4X have become a slugfest, with the first 40 to 50 taps preventing players from exploring the experience. It's an obsession with teaching players what to press rather than how to play. When I talked about PhD psychologists chained to surveys in games, it's exactly this that they should be doing instead: figuring out the best design to teach players how to play, rather than what buttons to press.

The good news is that Match3 has gotten to the point where they simply drop players in for level and ask them to start matching. The "learning by doing" approach is something PC/Console advanced far more. Kim Swift's legendary talk "Our Journey From Narbacular Drop To Portal", and Asher Vollmer's "Prime, Teach, Observe: Tutorializing Innovative Mechanics" echo the same ethos.

Economically, we need to model tutorials as a cost. Learning requires cognitive effort, and while there are payoffs, we make sure those payoffs come. Explicit rewards are not enough. It's really intrinsic rewards that should be front and center of a successful tutorial completion. At game start, players have no idea what extrinsic rewards are worth. They might have some priors based on other games, but it can be tough to rely on them outside of gacha-style 4X characters.

Instead of trying to do this, Mid-Core and 4X have created entirely fake funnels of casual gameplay that actually achieve the learning by doing approach. There's a lot of money out there for a team that can reinvent what the early part of the 4X and mid-core funnel looks like.

The defining characteristics of social in gaming are coordination and cooperation.

The clearest example is 4X. To take the seat of power on a world map, players need to cooperate by donating resources and strengthening the clan. Machine Zone founder Gabe Leydon once said that a $1,000 dinner bill looks expensive until you realize you're paying for ten other people. Players willingly subsidize others because the system makes progress contingent on synchronized group action. Without it, the group fails.

It's not simply spreading the resource love that grabs the seat of power. It's also coordination. Players need to time attacks and marches against rival clans, and to synchronize to achieve the goal. It's the combination of these two things that has unlocked what is quickly becoming mobile's biggest genre.

Soccer is another great example. There is one ball, and everyone wants the same outcome. No individual, no matter how talented, can score consistently without modeling others' intentions, positions, and timing. Success is inseparable from coordination and cooperation.

Monopoly Go's partner is a smaller-scale example of this. Coming out of economic ultimatum games, Partner events set up four donation stations at the center of the board. At each donation center, two players are paired. They share a progress bar that fills up once each player donates an event currency. Given the four stations and four different partners, players donate to a station that another player has also donated to, creating a tit-for-tat strategy game and essentially gamifying traditional economic games in a way that economists have failed to advance and economic designers have failed to study. But this, of course, requires coordination as well, and it requires feelings of reciprocity that, again, are separate from players acting individually and then cumulatively adding numbers together. The big irony here is how we can design games that are social, not as feature addenda.

The biggest social whiff has undoubtedly been in serving female gamers. The assumption that this audience's needs and wants boil down to match3 is lazy and wrong. No one has been able to design a female casual game that puts social described in this way at its core.

We know how to design true social, and keep choosing not to.

Even short of Skinner boxes or randomization, core UI and UX design still sits far from psychologists. They get relegated to surveys rather than onboarding, interface hierarchy, and progression pacing. Flow state is not just a designer's instinct; it is a cognitive problem that should be right up a psychologist's alley.

We have a buffet of levers where PhD psychologists could shape outcomes: the hedonic treadmill, spaced repetition, working memory limits, and the way players form expectations and reference points. These are not niche curiosities. They are central to retention, satisfaction, and monetization.

In a career first for me, I am not dunking on psychologists. They did not choose the survey corner. This is on senior executives who do not understand how to employ them, given the centripetal force of how they have been deployed before.

I pray for the day of PhD game psychologist liberation.

The "behavioral" critique is more subtle: utility is shaped by where outcomes land relative to an internal benchmark, not by their absolute size. Two identical end states can feel very different depending on whether the player experienced them as a gain or a fall from expectation. Consider match-3 under two scenarios:

A: Player never builds a streak. They have 11 attempts, winning 10 in a row and losing the 11th. B: Player builds a streak. They have 11 attempts, winning 10 in a row (activating Super Light Ball) and losing the 11th.

Nothing about the final state differs! Yet the second path feels worse. Not because the player lost more in absolute terms, but because the session included a drop below a reference point the game itself had just established. The streak redefined what "normal" progress felt like, and breaking it made later, otherwise-neutral outcomes feel like perceived losses. Identical outcomes can generate different utility once expectations are formed and then violated.

Returning to Squad Busters, remember their win streaks were functionally capped (after 10 wins, players received +3 "taps" or chances to upgrade the loot box), but the number on the badge continued to increment (some had thousands of wins!). The feature was removed, with Supercell citing that players felt anxiety about losing their win streak and stopped playing altogether, as noted in this Squad Busters update.

From a neoclassical perspective, nothing was being lost when a streak broke beyond the foregone future taps, which were capped anyway. From a reference-dependent perspective, a lot changed. Breaking the streak meant falling below an internal benchmark the game itself had encouraged players to form. Play after the break was not evaluated against a neutral baseline; it was evaluated against a recently experienced, higher standard of normal.

This is the same mechanism as the match-3 example: the path includes a rise to a reference point followed by a drop. Once that drop exists, ordinary outcomes are experienced as losses.

Seen this way, Supercell's explanation that players felt "anxiety" about losing streaks is not evidence of irrationality or fear-based loss aversion. Stopping play was not avoidance of loss; it was avoidance of reference violation.

While there could be an element of executional risk, overspending on UA and maybe launching before the KPIs justified it, there's a simplier explanation: 4X is eating the world.

Habby's monetization template is still clever. It leans heavily on annuity-style spend, but demand stimulation follows the familiar squad RPG playbook, wrapped in differentiated cores. Archero 2, Capybara Go, and Wittle Defender all sit squarely in the Evo-Fusion-Evolve meta-collection paradigms. Certainly, minigames and tournament live ops have improved with each iteration, but there's no getting around their power-score core.

4X is the most powerful monetization engine mobile has ever produced outside of Match. While power score remains the driving factor, it's wrapped in a socio-competitive economic core. In 4X, competition for power becomes a distributed auction. Kabam once ran a survey of spenders that asked a simple question: Why do you spend? The answer was always the same - for the clan. That dynamic creates effectively infinite sinks. There is, quite literally, a million dollars' worth to spend. (And indeed, an infamous Game of War case of a man embezzling company funds to spend over a million dollars on that game proves it.)

As Gabe Leydon once put it: "A $1,000 dinner bill looks expensive until you realize you're paying for ten people." squad RPG never solved that problem.

The real dagger wasn't just 4X itself, but its absorption of squad RPG. Modern 4X quietly incorporated squad RPG and economy builders into its core faster than squad RPG could ever absorb 4X-level socio-competitive economics. Reed Hastings once asked whether Netflix could become HBO faster than HBO could become Netflix. The same dynamic applies here, and with a clear winner in each case.

We've already seen the wake of the damage with Lilith and Moonton, while Habby appears to be the latest casualty.

As long as 4X can push higher LTVs, it can bid higher on UA. In mobile, whoever wins the auction wins the market.

The most effective product orgs I've seen came out of Scopely circa 2020: a two-product director model split between Roadmap and Performance. The evidence of its success has been its survival and its spread. It was copied by people who left, and I've advocated it to multiple studios. Beyond sticking with Monopoly Go for many, many years, and retaining key leadership, I consider this @Javier Ferreira's biggest win.

The insight wasn't just having two directors; it was the predicate that anyone could be one. These roles were not reserved for career PMs or recovering investment bankers. A game designer could own Roadmap PM. An analyst could own Performance PM. And indeed, many did. People were expected to bring their native craft with them rather than sanding it down into generic PM language.

Roadmap Directors owned the three-to-six-month vision of the game. This meant fewer direct reports, fewer fires to manage, more thinking. The Performance Director ran the game day-to-day: live ops, metrics, tuning, with a larger staff. The structure let one person focus relentlessly on where the product was going, while the other focused just as relentlessly on how it was behaving right now.

Under a GM, the two roles reinforced each other instead of competing. Strategy didn't get drowned in Slack emergencies, and live performance didn't get subordinated to abstract future promises. Together, they were meaningfully stronger than a single overextended head of product.

It also created real career pathways. People could start in live operations and grow into broader product leadership. Designers didn't have to abandon design to gain influence, but they could step into product roles if they wanted to. That permeability matters more than most org charts admit.

While I'm sure things have changed since I was first exposed to the model, and the image here doesn't reflect the structure exactly, and certainly not for all in development products, the broad strokes have survived.

1. Staking

In Web3, staking locks hard currency into a wallet in exchange for ongoing benefits. The explicit goal is price support (the quantity supplied of the token decreases, which in turn increases the price). But abstracted from crypto, staking is simply a way for players to set aside hard currency without destroying it while still deriving value.

Staking is not a sink in the literal sense: players retain ownership and liquidity in the future but gain benefits in the present. Since hard currency can't be extracted outside the system (unlike tokens), it effectively becomes spent and depreciated on a different time horizon, but still increases money flowing into the system.

As a design primitive, staking is closer to a savings account than a monetization trick, which opens space for VIP systems that reward balance maintenance rather than sunk costs. Mid-core games have been sleeping on this mechanic. There are some versions of it, like Vault and 4X games, but I wish everyone would push it harder, certainly Chinese games.

2. Uniqueness

This is not a Web3 invention. Physical collectibles have serial numbers, and trading cards have editions. Yet cosmetic economies have largely ignored true uniqueness, opting instead for infinite supply skins differentiated only by timing and (sometimes) price. It's such a small ad that can drive more price heterogeneity, and missing it seems nonsensical, especially for something like FIFA. Look no further than the recent social media trail of the "Kabuto King."

More recently, Garena Free Fire has gone further by attaching serial numbers to cosmetics and selling them through auctions. This does two things at once: it differentiates cosmetics within their own supply, and it lets the market, not the designer, discover value.

3. Auction-based markets

These never disappeared, but they are sneakily resurgent: Warthunder, CS:GO 2, Rainbow Six Siege, Delta Force, Tarkov (and Roblox?!). They cluster around PC shooters and around players who value mastery, collectability, and competition.

Markets create new player segments, not just function as producer-side price discovery. Consumer price discovery is actually a game that participants play, deepening the connection to the game, regardless of real-world value fluctuations. Even markets using only "in-circuit" money still open up gameplay.

A year ago, I wrote that Call of Duty Mobile posed a puzzle no one else had solved. One franchise made mobile shooters work in the West, while every other serious attempt stalled or failed. At the time, I thought the answer was controls. Call of Duty made it far easier with auto-fire to play with both fingers or thumbs on the dual analog sticks, rather than the "claw" position that many Eastern players take, which requires the phone to rest on a surface. It turns out the Western hold-up in building a monster mobile shooter franchise is labor-based.

China's console ban and firewall-walled Steam Access ended up creating Regional Dr. Perky to our Dr. Pepper. Crossfire is a juggernaut franchise that appeared in the wake of China's lack of CS:GO. But mobile always had stronger roots in China, partly because of the console ban and partly because computer cost relative to wages. As late as 2010, Chinese household disposable income averaged $2,000 per year. Not exactly GTX money! Internet cafes help defray the cost, but are still a significant friction.

In the West, shooter and mobile DNA evolved apart. FPS intuition lived in PC/console studios that treated mobile as a compromise. Mobile expertise lived elsewhere, optimized for UA, but often detached from real shooters. In a real "gamer" sense, FPS players are the core bulwark of Steam. You can bridge that gap with ports or partners (nice try, Netease), but not having a Bungie PM sit on Destiny: Rising doomed the product in the West.

Similarly, Call of Duty Mobile worked because it forced a synthesis: Western shooter product leadership paired with Chinese mobile production discipline. That combination is rare, not because talent is scarce, but because the overlap is. Apex Legends Mobile was on the success track, leveraging the same Western leadership that had subverted itself to mobile, alongside Chinese development prowess that meets content expectations for the platform.

Gen Z will change this equation, but less because of player attitudes than because of labor. Designers who grew up with shooters, mobile, live services, and cross-platform identity don't experience mobile as a downgrade (nor is it!). It's a different service area and a different medium that has different constraints.

EA's VP of Mobile Shooters' real challenge will be assembling the right team of Western product managers and designers and ensuring that they can adapt to mobile expectations while integrating Chinese-style content and progression systems.

The saddest part of all of this is that Vince Zampella would have been the perfect person to learn from and to achieve this mission. His loss is already felt.

Listen on Spotify

Shooter Monthly #3 is here. Chris Sides, Feras Musmar, and Christopher Anjos (DICE) join me to settle the only question that matters: is Die Hard a Christmas shooter movie? We also break down the latest shooter signals across release strategy, mobile traction, and hero-shooter retention.

We discuss:

• Call of Duty's rare public admission of failure, and what the end of back-to-back sub-franchise releases actually signals
• Marathon's delay, the $40 price point, and why fiscal calendars still distort launch strategy
• Why Valorant Mobile exploded in China, and why Western mobile shooters keep failing to break through
• EA hiring a VP of Shooters for mobile, and whether Battlefield or Apex can realistically return to phones
• Highguard's confused reveal, the danger of losing narrative control, and why innovation still matters
• Marvel Rivals as a ceiling, not a failure, and what the hero-shooter arms race gets wrong about retention

Links:

• Watch on YouTube
• Listen on Spotify

Talk of the Year: Roblox’s continued growth.

Roblox is the most dynamic sector of gaming, bar none.

DAU growth at this scale, with T1 growth, and flat APRU is breathtaking. Likely 1 in 5 U.S. and Canadian kids play every day (26M DAU, ~15M U13 DAU, 80M U13 kids).

The success of Steal a Brainrot, Rivals, and Grow a Garden proves there's a ton of movement at the top of the charts, and you can build real, sustainable businesses on a platform. Roblox's share price reflects this, and there's still room to grow. In fact, growth is accelerating.

The aging-out question deserves a lot of scrutiny, and the new verified age will appear in the disclosed metrics, giving us a more accurate sense of whether players are growing out of the platform and graduating to platforms like UEFN.

Correction of the Year: InZOI.

I wanted this to work. I wanted a credible challenger to The Sims that proved there was room for a new life-Sims thesis at scale. I knew this would face many challenges, as EA has built a very special distribution network with its players, usually directly. The game lacked the localization and player empathy needed to win in such a specialized market.

I still believe in the thesis, and I hope a Western studio will take a shot at it. I think they misread the very particular female demo that supported the Sims in the West.

Most Damage Received: Lilith.

I covered this in a prior post, but ultimately, Lilith has fallen tremendously, even with accounting for webstores. 4x is at an all-time high, and Lilith hasn't been able to keep focus, instead wasting time on Fartlight 84, which got another relaunch, and AFK Journey, which was launched into a squad RPG market that's collapsed to the rise of 4x.

The most unfortunate part is that they have all the tools in their portfolio to refocus on their core 4x properties and innovate those instead. Start focusing on Palmon!

Most Disappointing Game: Rainbow Six Siege X.

A relaunch is supposed to reset the franchise, but Siege X seems to welcome a historic decline. It failed to meaningfully expand the audience, failed to re-energize the core, and ultimately felt like an expensive acknowledgement that the franchise is boxed in by its own legacy constraints. That’s not what a relaunch is for.

Mice Nuts: CloverPit.

CloverPit is certified Mice Nuts, pulling in under $10M. But ultimately, it is the roguelike take on Bellatro that slots should have gotten a decade ago. This is ripe for Habby or a more ambitious social casino company to continue pushing the genre forward by poaching another from Steam.

Skeptical Move: Royal Kingdom.

Dream is an innovative organization, and that’s precisely why this launch puzzled me. The sequel thesis assumed compounding benefits that the genre no longer guarantees, while the positioning felt misaligned with where real growth in puzzle was actually happening. This wasn’t a bad execution problem, and was instead a structural bet I don’t think the market wanted.

Brilliant Move: Scopely acquiring Niantic.

Scopely is now the Go company, and ultimately, Monopoly Go and Pokémon Go have helped cover up many weaknesses in the broader portfolio. I covered this earlier, but Stumbleguys hasn't materialized in a lot of the acquisitions. The flurry of them before Savvy's acquisition seemed largely to have stagnated.

The great part is that none of this matters anymore. With these two franchises composing more than 80% of revenue, it's all Scopely needs to succeed. Meaning that games like Star Trek: Fleet Command and Marvel Strike Force are okay to be pushed to the side.

If the rumors that an ambitious cross-platform game from Scopely is true, it feels like a distraction from what should be continuing to double down on these key franchises and cloning them instead.

You Crushed It: Grow a Garden.

Grow a Garden proved that new breakout hits can still emerge on UGC platforms, not just survive on legacy momentum. It's undoubtedly true that the game has declined quite a bit, but it's still remained on the top-grossing charts for a long time and has very few developers. It reached ceilings we didn't even know were possible on Roblox. And Steal a Brainrot's one-two punch solidifies Grow a Garden's success as more than just a flash in the pan.

Limping Donkey: Dreamhaven.

We can be direct and unrelenting on This Week in Games because we want to see great game companies succeed. When Blizzard spinoffs started occurring, I actually got more excited, as I would have loved to see the creativity trapped within the studios go wild. Second Dinner absolutely delivered on this, but every other studio has been a letdown. Dreamhaven seemed to be the biggest one directly managed by Mike Morhaime. But like Belichick at UNC, it feels like the game has passed them by, reminding everyone how hard this is.

Now it’s your turn. Vote, disagree loudly, and leave comments!

There’s a far easier solution to “trading games” that avoids the design baggage of item-level liquidity (and the forever-same problems crypto keeps re-discovering): allow players to trade accounts.

Diablo III’s Auction House remains the cleanest case study. I went deeper on this in Sense & Nonsense in Blockchain Gaming: Autarky No More!.

Blizzard was explicit about what it was trying to solve from Diablo II:

1. Reduce fraud and third-party RMT
2. Create safer price discovery for players
3. Smooth progression by helping players find upgrades they could not reasonably farm themselves

All three goals were rational, and they actually succeeded. 50% of players regularly engaged with the auction house (!), with some players bringing in enough to supplement full-time incomes. Jay Wilson (Diablo’s game director) explained that the auction house performed as intended to combat account fraud in “Shout at the Devil: The Making of Diablo III”.

Given all this, why kill the auction houses? Blizzard didn’t want to sell progression, among other issues.

Crypto has been running the same experiment, loudly and expensively. When assets are individually liquid, players behave like labor. Bots arrive, and low-wage grinding dominates because countries like the Philippines can offer grinding wages that exceed local employment opportunities.

Account trading achieves all three original Diablo goals with none of the design baggage:

1. Fraud drops because transactions are rare and high-friction, while still being possible
2. Fairness improves because value is bundled with time invested
3. Progression remains intact because power cannot be surgically extracted from its context. The “journey” remains intact, rather than value being parceled out.

It’s simple to implement and effective at preventing unintended consequences while still achieving many of the objectives that trading was invented to achieve in the first place.

More games should consider it.

Chris Smith (PhD economist at Star Atlas) and Eric Guan (data scientist at Second Dinner) join me for Episode 46 of the Game Economist Cast.

We talk about:

• Sweepstakes, social casino, velocity, and why most players never cash out
• Why Wordle feels flat to some designers, and why elegance is not the same as progression
• Surveys as UX (not truth machines), and how to extract signal without lying to yourself
• Compensating differentials, handicaps, and why 50% win rates kill progression
• Bots, deception, and whether games are magic shows or fraud

Links:

• Watch on YouTube
• Listen (Castos)

sensortowerR

• Submitted for a wider CRAN release, but the GitHub version is available now.
• st_publisher_portfolio() — one-line portfolio analysis: Downloads, Active Users, and MAU as a year-to-date time series
• st_app_info() — publisher search support: search regional SKUs (useful when publishers ship under different regional names)
• st_parse_web_url() — convert Sensor Tower web URLs into API parameters

Get it here: sensortowerR on GitHub

Economy Flow Builder

• Live in Figma
• Rebuilt the collision engine and fixed layer ordering (lines now stay behind action boxes)
• Fixed all templated JSON to render correctly, so you can see how the package works
• A bunch of other bug and error fixes
• Still a work in progress, but it’s already been helpful for jump-starting my economy research

Cowen judges games as if they were films or novels, "...it's too context specific to the person playing it and doing it. It doesn't quite manage to generate narrative and vision in a way that is intersubjectively understandable", he says. But games are not omnidirectional art; they are bi-directional. In many ways, this is the defining feature of the medium. They complete themselves through the player(s). The right comparison isn't to art in the way he understands it, nor is it the only comparison.

They are the most comprehensive artistic medium we've ever built. Music, architecture, choreography, writing... converge within a system that responds to human interaction. Games are the first broadly accessible art form with over 1B MAU (?), in which the art changes when you touch it. That alone should place it on its own creative plane.

The medium has pushed commercial innovation further than most other industries. F2P, live-ops, MTX, experiments, virtual economies, and marketplaces... Every other industrial sector is a goddamn dinosaur relative to the last 10 years of games.

And then there's the science. I'm literally part of The Experimentation Group and only because David Nelson was crazy enough to work with Steve Levitt on crazy Candy Crush experiments. That became taking King's dead games, the ones they were no longer developing, and running crazy stuff on them. This is the stuff economists fantasize about, and games actually do. More practically, let's not forget gamers subsidized NVIDIA for decades and are largely responsible for America's lead in AI development.

One of the medium's most distinctive features is its pliability. There is no surface games haven't colonized: phones, TVs, watches, browsers, social feeds, even enterprise software (have you played Confluence Race yet?)... Any space that can host interaction becomes a colony. No other medium travels this well!

Cowen's critique isn't unusual and reflects a broader academic blind spot. Too many economists still treat games as toys, even the younger generation. These positions are outdated by at least a decade.

If our best thinkers still misunderstand games, it isn't because games are immature; it's a willful refusal to engage with them. Games are not behind other art forms. They have surpassed them. They are the clearest expression of human faculties because they encompass all of them and will continue to evolve alongside us.

It was easy to imagine Marvel Rivals as the next great moment when China seizes a Western crown jewel and shows the West how it should be done. NetEase telegraphed ambition early, promising new heroes every six weeks, dropping the collective Western jaws to the floor. Yet the engagement response has been meh. It calls into question the Eastern content arms race altogether.

Rivals has settled ~1.7M DAU and ~6.5M MAUs. At some reasonable mARPU ranges for a cosmetics economy without RNG-based monetization, the run rates are shockingly sober, ranging from anywhere to $100M-$300M. After licensing fees, active marketing (!), and what I'm sure is an army of developers that rivals small countries, it's hard to imagine this game is anywhere close to profitable.

The interesting comparison sits with Overwatch. Although far from its peak, it has not collapsed in the wake of Marvel Rivals. Instead, Blizzard leaned into something the Chinese approach rarely prioritizes: productivity. Perks, the lightweight, in-round progression system, and Stadium, the roguelike economy layer, deliver meaningful retention through cheap stat changes (announcement; video). These features are not "content" in the traditional sense. They are multipliers on designer time, creating more retention per developer hour than any new hero could. Blizzard chose craft; NetEase chose scale.

Both decisions now run into the harder truth that the hero shooter resembles Match3 and the MOBA before it. The category has reached equilibrium. There is loyalty at the top, churn that recycles within the genre, and no new surface area left to capture. The next ten million players are not hiding behind a new character or a new map. They may not exist at all.

When a genre stops absorbing new ideas, it drifts into crisis. Designers keep pushing, but the gains are marginal, while publishers spend more, but the audience does not grow. That crisis is often the precursor to a paradigm shift. This is why the rush to build extraction shooters was and is so essential: new sub-genres are new surface areas and new opportunities to compete.

It's not fair to say that Marvel Rivals is a failure, but it is in an awkward middle age, and it's hard to imagine active users growing from here. Instead, it needs to face the struggle that nearly every Western shooter faces: how to monetize. And here is where NetEase made the greatest mistake, which is firmly planting the game's paradigm into the West. Meanwhile, I look across to Delta Force, which could give fewer shits, and it will be rewarded with obscene amounts of revenue.

It's remarkable how long Blizzard has avoided the obvious. Cross-play has matured across genres. Diablo Immortal is looking north of $250m lifetime between China and US iOS alone, and even bolted on a PC client. Meanwhile, World of Warcraft - arguably the most durable live-service in history - still lives a lonely PC life. No mobile SKU, no cross platform, and no F2P (WoW tokens are for another day). Giving this franchise to NetEase to launch a Chinese-specific mobile SKU is the most obvious play in history, besides maybe porting League of Legends to mobile. Yet arrogance always seems to get in the way of player empathy.

A decade ago, you could argue that the technology wasn't ready. Maybe. Westward Journey Mobile launched in 2015, while the initial PC version launched in 2001. There's no technical barrier to putting World of Warcraft on mobile. Diablo Immortal proved the other bits: outsource the mobile build, regionalize the monetization, and preserve the "sanctity" of the Western-facing product by pointing critics toward selective outrage. It's the cleanest form of plausible deniability this industry has ever invented, just as Amazon did the opposite for Lost Ark.

Which makes Blizzard's refusal even stranger. The studio still behaves as if the IP is too sacred to risk translation, the same mistake Riot made when it convinced itself League of Legends couldn't work on mobile. Tencent solved that argument by creating Honor of Kings and birthing one of the largest entertainment properties on earth. When a partner fills the void you refuse to enter, that's not stewardship; it's self-inflicted irrelevance.

By sitting out, Blizzard has allowed every Eastern MMO to blossom uncontested. Black Desert, Lineage, Moonlight Blade, and now Where Winds Meet; pick your example! All of them ate the lunch WoW left on the table. This is a kind of IP narcissism. A belief that the brand is too refined to mingle with the devices and monetization frameworks the rest of the world uses daily.

The market doesn't care about that insecurity. Millions (billions?) of Eastern players have already voted for mobile-first MMORPGs, cross-platform, and monetization calibrated to regional norms. Blizzard could have owned this space. Instead, it's repeating the same pattern: protecting their idea of World of Warcraft, rather than delivering it to the people who still want it.

What's needed right now is an inverse Diablo red-shirt guy - the Eastern player who stands up at BlizzCon and says, "Where the fuck is my MMO?"

If Diablo Immortal proved anything, it's that Blizzard knows how to do this: ship it to NetEase, and move on. The cost of failing to do so is that it has lost much of its autonomy and has had to contend with the dearth of product managers. Just self-inflicted wound after self-inflicted wound, and yet at any point, Blizzard can choose to opt out of this. All it takes is the will to serve players.

Listen on Apple Podcasts · Spotify (show feed) · Episode page (Podscan)

Call of Duty's worst launch in a decade collides with the rise of extraction shooters, and the balance of power inside the genre shifts. We dig into why Black Ops 7 cratered, how Battlefield 6 stole the spotlight, and why Arc Raiders is suddenly the most important new IP in shooters. Chris Sides and Feras return to map the fault lines, spar over SPMM, and weigh whether Tarkov's disastrous Steam debut marks the ceiling for the genre's original heavyweight.

We discuss:

• Black Ops 7's collapse and why marketing, futurism, and CPI pressure all converged
• Battlefield 6's surge, the RedSec flop, and why innovation is still missing in BR design
• The extraction boom: Arc Raiders as the "PUBG moment" and its long-term retention risks
• PvE as the emerging driver of mainstream extraction demand
• Skill-based matchmaking as an economic problem, not a matchmaking one
• Tarkov's failed 1.0 launch and what its country mix reveals about stagnant reach
• Whether Embark should kill The Finals and move every resource into Arc Raiders

Moonton's RPG bet rests on four games: Mobile Legends: Adventure, Watcher of Realms, One Punch Man, and the seemingly (?!) abandoned Acecraft experiment of turning a Western Steam hit (Cuphead) into a live service. Acecraft never cleared the runway! Watcher of Realms has seen monthly revenue sag since 2024, while Mobile Legends: Adventure (MLA) went from $27M in 2023 to $16M in 2024 to $6M this year. Y-I-K-E-S.

I say this as someone who has spent $5,000 in MLA. Few games have built a more progressive runway than MLA, but it's also a clear demonstration of why squad RPG is out of new ideas. A SINGLE hero has four different progression verticals and two shared ones: Soul Vessels (2 slots), Orlay Cards (9 slots), Glory Gems (8 slots), the Wheel of Time Nexus (four currencies), shared Artifacts and Companions (one each per team).

In practice, it all collapses into a single scalar: power score, but only as a function of attack, health, defense, penetration, or damage reduction. MLA shows how far you can stretch a stat economy without ever solving the underlying design question: how do squad RPGs create fun problems for players to solve instead of the next thingamajig that adds +100 attack.

The current answer has been to build 4x mechanics, which add much stronger coop-PvP gameplay than squad RPG ever has. We saw early streaks of this in, say, Marvel Strike Force's Alliance War, Star Wars: Galaxy of Heroes' Territory Wars, or Walking Dead RTS's Faction Wars. Of course, the answer here is much more literal, which is to shove a squad RPG inside of a 4X game, which is precisely what White Out Survival and Tile Survive have done.

My great hope was that DC Dark Legion would take off, which is a much more moderate integration of 4x mechanics into a squad RPG core. Unfortunately, the game has fallen off a cliff and looks like it'll be shut down within the next six months.

There's an emerging second answer that has been around for over a decade. With the release and success of Where the Wind Meets, there might be a real opportunity for power resets to maintain squad RPG without it becoming 4x.

In seasonal reset economies like Diablo, FIFA Ultimate Team, and Tarkov, gear is lost, and players earn it back over the course of a season. There are numerous downsides to this model, the most obvious one being that time invested now does not provide long-run benefits. But the design flexibility and the never-ending chase also offer players a new opportunity to enter the funnel without facing a daunting progression system built over ten years.

Yet, the industry keeps treating it like an engineering puzzle: tune the constraints, loosen the tiers, hide the MMR, expose the MMR... The only thing less stable than the industry's approach to it is the experience you get without SBMM. While it's painfully obvious SBMM is retention boosting, it's economy design that will remove its rough edges, not more mathematical tuning.

The Call of Duty matchmaking white paper makes this painfully clear. Using skill as a factor lifted end-of-match rank and improved retention for ~80 to 90% of players. Of course, SBMM increases retention versus random matching; there is no serious counterpoint left on that axis. A small minority of players benefit without SBMM because they can stomp the casuals. Another case of Reddit's vocal minorities shouting down the silent majorities!

The tension sits elsewhere. As Feras Musmar points out in Shooter Monthly #2, fairness is not fun. A system that perfectly equalizes outcomes and quietly pushes everyone toward a 50% win rate drains emotional payoff from play. The problem is not the algorithm; it is the absence of any economic structure that shapes what a win or loss means under different conditions.

The complexity rises quickly once the sophistication of modern multiplayer games is taken into account. Battlefield is an excellent illustration of this. Battlefield's kill distribution has a very high Gini coefficient of inequality, meaning a thin slice of players absorbs most of the kills, while the majority absorb most of the dying. This is not a failure of SBMM; rather, it is a geometric constraint of needing to fill 64 or 128-player servers in a reasonable time.

The first emerging answer is found in Clash Royale and Marvel Snap. Bots break the zero-sum structure and give players controlled, confidence-restoring wins. While unlabeled bots are unethical, it's also readily apparent that this solution dominates an A/B test. The risk is that it reduces trust in the system.

The second solution is system's design: build an economic model that absorbs unevenness and celebrates rather than pretending the matchmaker can eliminate it. Command and Conquer Rivals is a good example of what this looks like: challenge battles protected lower-level players from losing medals and paid out double if they upset a stronger opponent. The matching system stayed the same, but the payout table changed the emotional meaning of the match.

If we want SBMM to feel better, we cannot tune our way there. Matchmaking math already performs its core task: preventing the ecosystem from collapsing into a shark tank. The real work now belongs to economy designers. They have to build systems that compensate players for risk, reward outlier success, and give uneven matches a story worth telling. This is an economics problem, not a math problem.

Recent reports suggest Saudi Arabia is feeling a capital pinch after buying Scopely and EA (~$61B). Now, with the closing of Sandsoft it's become apparent the country is world-class at buying capability, not building it. No state has spent more on assembling a global portfolio while producing so little local talent. And despite all this interest in games (ha, eSports), it's baffling that Saudi's Vision 2030, a plan to reduce oil dependence, doesn't have a games plank. Welcome to GAME VISION 2030.

Michail Katkoff told a story on TWIG that captures the underlying confusion. After a day of investor meetings in Saudi Arabia, the group was invited to a private residence. Inside, the hosts had arranged rows of high-end PCs, coaches waiting at the stations, and a full professional League of Legends setup ready for play. This wasn't some industrial strategy; instead, it was a glimpse into the preferences of the elite. The royal family likes games, and personal taste rather than state capability has driven billions in acquisitions.

The reality is that you cannot buy an industry, but you can develop an ecosystem. Like all economic clusters, the Games industry depends on agglomeration effects, knowledge spillovers, and dense local networks of artists, engineers, designers, and producers. Vision 2030 has... no vision here. Without a talent base, capital doesn’t multiply; it evaporates into one-off deals.

GAME VISION 2030 corrects this with three planks.

First, build a top-tier game university. Not a regional school, but a global institution the likes of which the world has never seen. Recruit leading designers, engineers, and researchers. Make it the world’s most concentrated hub of game craft. This is how you create durable spillovers.

Second, require every Savvy-adjacent investment to establish a local support studio. If you put a billion dollars into Niantic, Scopely, or EA, the country should receive something more tangible than a stake. Each partner should operate a team in Saudi Arabia, contributing to real production: Monopoly Go live-ops, FIFA content, Pokémon GO event pipelines. This is how capability forms.

Third, seed a domestic fund for Saudi-based founders. Provide grants, early capital, publishing support, and a path from prototype to live service. No country builds a creative industry without local entrepreneurs at the center. I'm sure Ilkka Paananen would spin up another Supercell Spark program after Tokyo.

While the personal preference for games among Saudi capital allocators is flattering, actually contributing to the games ecosystem is far more ambitious. There are millions of citizens with whom we can share the joy of this sausage-making process, and GAME VISION 2030 is how Saudi Arabia gets there.

Nutrition labels work because calories and grams are chemistry; AI is not. The boundary between "AI-generated" and "AI-assisted" is a thicket of judgment calls. If an artist uses an AI upscaler, does that count? If an engineer accepts a Copilot autocomplete, which is already universal across software teams? What if game designers used AI to think through designs? Under Valve's logic, the answer is yes: any use triggers disclosure. This tells us only that modern tools exist.

A better analogy is not nutrition; it is fair-trade labels. They are voluntary, definition-fuzzy, and mostly marketing. If consumers truly craved "AI-free" games, publishers should race to advertise them the same way they trumpet "no loot boxes" or "no DRM." The reason the suggestion gets rejected is revealing. Information is not the goal. Punishment is.

Notably, none of this internet energy is aimed at AI code. No one is lobbying for "built with AI coding tools," despite the fact that half the industry already is. The outrage concentrates on art because this debate is not about consumer transparency; it is about guild politics. It is an attempt to cordon off certain tools by stigmatizing the people who use them.

Economically, this is fighting gravity. Since the invention of the pulley, human progress has been a story of labor-saving tools. Photoshop compressed months of manual work into hours, just as the pulley, plow, and combustion-engine innovations did for farming. The absurd idea of 3000 BC Mesopotamian "plow" disclosures on grain harvest should tell us something.

In the United States, farming went from employing nearly half the population to under 2% while output exploded (source). Productivity rose, prices fell, and living standards increased. For small teams, AI is the most literal labor-enhancing technology games have had... ever. It expands what a five-person studio can build and shrinks the cost of ambition.

This is why the panic will not last. Steam's label system will drift toward universal tagging as AI seeps into every corner of production, at which point the disclosure becomes white noise. Meanwhile, the teams that embrace AI will ship more inventive games with fewer people, and the rest will still be arguing over the fine print on a label that no longer means anything at all.

Fortnite’s newest season arrived with the newest roster of improbable IP mix: Power Rangers, Uma Thurman, The Simpsons, Ferrari, Darth Vader. The point isn’t the novelty of each deal; it’s the sheer gravitational pull required to make all of them exist under one banner. In many ways, instead of becoming a UEFN platform, it's become the 𝘢𝘯𝘵𝘪-𝘱𝘭𝘢𝘵𝘧𝘰𝘳𝘮.

Bill Gates once argued that a platform only deserves the label when the value created on top of it exceeds the value of the platform itself. Yet, the most valuable part of Fortnite isn't what sits on top of it; it's how it absorbs everything around it.

That's why UEFN’s stall doesn’t matter much. Developers don’t need to build on Fortnite when Epic can pull entire IP catalogs directly into the core game faster than any creator ecosystem could ever mature. That magnet recreates something gaming lost as the internet becomes increasingly balkanized. Fortnite remains one of the few places where monoculture briefly reappears. A teenager, a lapsed player, and a parent who hasn’t touched a controller in twenty years can all recognize the silhouettes in a Fortnite trailer. The game became a shared language, not because of its mechanics, but because of its capacity to ingest familiar icons at an industrial scale.

It’s a rare achievement. Riot did something similar when League of Legends became the foundation for an entire company, while King turned Candy Crush into the backbone of a 3,000-person company. Ultimately, Epic seized the opportunity and never let go for years on end. While I've heard horror stories of developers working overtime, I've also heard the size of their bonus checks. No one should be crying in Cary.

It's another reminder that gaming's secret weapon is its pliability. The medium adapts to any screen, input, or business model. Epic saw this earlier than anyone else and recognized the commercial opportunity in becoming a cultural aggregator. They built the systems, the licensing machinery, and the creative cadence to sustain it. Remember, the fabled story of them pivoting from Save the World to Battle Royale mode only after seeing the check PUBG wrote them on its own massive success.

I hope @Tim Sweeney and @Mark Rein walk around the office with Carpe Diem caps on—they've earned it.

Roblox is rolling out an age-verification system that uses a voluntary facial check to unlock chat features for users 13 and up. This is happening at the exact moment when political pressure around social media is rising (thanks Jonathan Haidt), and Roblox is drifting into the crosshairs.

Of course, reality is worthless. Roblox has built some of the most sophisticated trust-and-safety infrastructure in the industry. They've invested in detection tooling, moderation capacity, and AI-based monitoring. Do remember they host 100 MILLION DAU - bad actors exist everywhere, and at that scale, of course, some of them are going to be on the Roblox platform. The question is always on the margin: is Roblox relatively safer than alternatives? Unequivocally, the answer here is yes.

In a perfect world, this would be the moment when they translate all that capability into political credibility. Instead, that Hard Fork segment was the opposite. It was a Tour de France in what happens when you send a Silicon Valley-bubbled founder into the lion's den of journalists looking to skewer. The rest of the world doesn't speak that way, and ending every tweet with high-five gives me a level of cringe I haven't experienced since Katy Perry went to space. He tossed out lines like "it can help us make cool decisions" and threw sarcasm back at the interviewers ("Fun, let's keep going down this").

The press was always going to press. They aren't wrong to question Roblox, but they're doing it the way journalists approach every moral panic: creating a villain rather than context. Remember the Hot Coffee mod? Roblox is just the cartoon of the week. Yet State AGs, class actions, and a growing set of child-safety investigations are real, and the more heated the atmosphere gets, the more likely it becomes that games end up treated as the next version of social media.

Roblox's comms team should never have let that interview happen without serious prep. This was a stress test of how the company handles public scrutiny on its most politically sensitive issue. It didn't go well.

In politics, perception is reality. Regulators don't parse technical nuance or internal moderation investments. They look for a narrative, and right now the narrative is straightforward: kids, platforms, and predators. Once a story reaches that level of salience, the details may as well not exist.

Australia is already there! Starting Dec. 2025, platforms defined as "social media" must keep under-16s off their services or use age verification. The list includes TikTok, Instagram, YouTube, etc. The definitions aren't fixed, but gaming isn't on the list. But Roblox gets this wrong; we'll get swept into the next legislative cycle.

I've seen recruiter after recruiter pitch Stockholm to Americans on bike lanes, fika, and the "European" lifestyle. Yet, when it comes to compensation and taxes, they become notably quieter. For American Game Devs, this likely means a 50% salary cut with 2x the taxes. If you think you'll make it up in benefits or lower cost of living, boy, do I have a bridge to sell you. But Europeans keep selling culture when the real selling point is financial! And it's a secret they don't even realize they have.

Nearly every Western European country has expert tax regimes. Sweden's version lets you exclude 25% of your income from tax for five years. Denmark, Finland, Spain, and the Netherlands run similar schemes, some of which push the benefit closer to 30% or include tons of deductibles. Most game developers meet the criteria without much drama, since it's such a specialized and unique field. Yet, recruiters rarely mention it, and I've never seen it in a job ad.

Then there's the pension stack. Europeans think of pensions as background noise; Americans think of them as a precious gemstone. In countries like Sweden, you get both a state pension you can eventually bring home and an employer pension that can amount to something around 30% of gross salary. You even get to choose from a menu of funds, letting savings compound. Ironically, it mirrors what George W. Bush once proposed for Social Security, yet we passed on fervent protest and outrage.

This pension stays with you even if you move back to America - take it from Swedish CEOs. They move to Portugal (~5% pension tax), draw down their occupational pension at 55 over five years, and then move back to Sweden. The Florida of Europe, Ladies and Gentlemen. Look, "work-life balance" is great, but "tax-advantaged comp vehicle with optional early withdrawal" is better.

If European studios want American talent, especially senior talent, stop burying the lede! Americans don't uproot their lives for cardboard straws and GDPR cookies. They move for upside, and somehow, bizarrely, Europe already offers it. Recruiters should lead with the tax break and the pension contributions, because these are the things that actually close the gap.

Put it in the job posts and say it on the first call. It's the simplest fix for a decades-long hiring problem: show Americans the money that's already on the table.

The solution is simply a slider or input field where the player chooses exactly how much they want to spend, and a discount that scales with that choice. The classic bundles are still available for quick selection. If quantity discounting lifts spend and reduces decision friction, then a continuous option is simply the logical endpoint of a curve we already trust.

A cleaner way to think about the problem is an ATM or a foreign currency exchange. ATMs offer quick options of $20, $40, $60, etc. They also give you something more powerful: a custom amount! Wild that we haven't copied this.

Webstores are the ideal place to do this. They have no SKU ceilings, no price point rigidity, and no platform rules that force you into fixed bundles. Webstore buyers are already the highest spend segment: they show higher ARPPU, more repeat transactions, and lower price sensitivity. A continuous discount curve lets developers reward that behavior without cluttering the store or fragmenting the economy. You can even cap the curve at a rational point so the discount never becomes irresponsible.

At GDC, Valve revealed that 33.7 of Steam Users have Simplified Chinese set as their Primary Language in 2024, 0.2% above English. None of that is a typo. It's also a market that really has never had consoles. Remember, in 2000, they banned the import of consoles, and it was only lifted in 2015. Valve should be doing press with Chinese outlets and releasing press releases in Simplified Chinese. I'm not joking. Some Chinese distribution plan needs to be front and center for this to thrive. So far, Valve has announced that they'll ship the SteamCube to every region where the Steam Deck is sold today, and the Steam Deck isn't sold in mainland China.

The other key part is not actually hardware sales, but let's excuse that for a moment. The Steam Deck ships about 2 million units a year. While I thought that was low, it's actually shockingly close to Xbox, which will likely run around 5 million units this year (PS5 did 18.5m). It's entirely feasible that Valve will become the number 3 console after the device's release.

But the pressing issue here isn't hardware sales; it's whether it adds incremental activity and users to the platform. This is where I have a hard time believing it'll accomplish that, similar to the Switch. I suspect it will sell units, but it won't convert into incremental active users. Essentially collecting dead weight like the Switch does on my TV console.

The giant problem here is: what does this thing actually solve for players that the Steam Link or Steam Deck doesn't? If this lands at $700 plus, Valve will actually need to answer this.

Look no further than Love and Deepspace (L&D) and Nikke: Goddess of Victory (Nikke). In one of life's great ironies, they are somehow mirror images of one another. One sells tenderness, romantic routine, and soft intimacy, while the other sells well... you know. Sensor Tower reports the inverse US demographics: 84% female for L&D and 79% male for Nikke, with the same average age of 27. They both have cemented their place in the Top 75 grossing games IN THE U.S, and monetize primarily on gacha character collection. The same principle shapes both games: understand what players actually desire, then build products that honor those desires without embarrassment or hesitation.

Nikke fuses team collection, shooting, and a signature style that the West has never attempted at scale. Love and Deepspace blends ARPG combat with romance arcs, photography, wedding systems, and a sleep companion feature (!!!). These choices are deliberate acts of empathy, acting on revealed, not stated, preferences. The contrast, with say, @Big Run Studios Inc., is stark.

Big Run is one of the few American companies that tried to build for an underserved audience, yet they framed that audience through identity categories. It's a distinctly American instinct, and the intention was genuine. But the market is a harsh mistress, and developers need to begin with a clear understanding of what resonates with players, not what developers believe should be meaningful.

No Western focus group predicted that a husbando ARPG with a sleep system or a fan-service-heavy gacha shooter would become a global top performer. Yet, these are the products Western players now reward.

The East is outperforming the West not because of cost structure, or 9/9/6, but because it listens more closely to players' revealed preferences. And when it acts, it acts with more conviction. It is willing to take tastes seriously, even when those tastes do not fit Western assumptions about what games should be. The empathy gap is real, and the East is profiting from it.

YouTuber Mattias Krantz tried to achieve this over six months and nearly gave up... until he discovered the holy grail of System/Economy Design! Seriously, stop now and write down a strategy for solving this before reading on.

At first, Mattias designed custom keys since octopuses have no bones and use pulling/grasping motions rather than the "poking" motion humans use for piano keys. Instead of pushing down, the octopus has to wrap a tentacle around the lever and pull it down. Each lever connects to a waterproof switch that sends a signal to a computer outside the tank, which plays a MIDI piano sound. And because octopuses are effectively deaf to airborne sound, he installs underwater speakers so the animal can physically feel the notes it produces.

Next comes the light system. A specific key lights up, and if the octopus pulls the lit key, it gets food and a positive sound. If it pulls the wrong key, it gets nothing but an even buzzer or timeout. The UX is clean, and the rules are understandable. Something System/Economy Designers all too often ignore!

Then we get the "training phase" or three-month grind fest in classic Pavlovian fashion. Mattias starts by simply putting food near the keys. When the octopus accidentally touched a key while grabbing the food, it got a reward. Ironically, Mattis had to stop giving food away for free. The octopus had to learn that pulling the lever = food. Hmmm... sounds a lot like a hard-currency inflation problem...

Midway through the video, the octopus simply stops cooperating. It's bored or frustrated, and the loop offers no structure to pull it back in. This is exactly where UX collapses in game design, too: clarity without consequence has no teeth.

The key pivot comes when Mattias reaches out to an engineer (surprise, surprise) to brainstorm a solution. Naturally, it is a progression system.

They build a progressive Skinner box. A small crab sits inside a clear tube above the tank. When the octopus pulls the correct key, the crab drops slightly. Pull the next correct key in sequence, and it drops again. Play the full "song," and the crab descends step by step until the octopus can finally eat it. Yes, ladies and gentlemen, they built a progression bar.

It's a brilliant combination of economics, psychology, and game design. The progressive element is classic Systems/Economy design: it guides the correct sequence, rewards incremental success, and makes progress visible.

Sensor Tower reports Royal Kingdom (RK) spent $200M* digital ad spend in the United States alone (major *). Oh, yes, this doesn't count talent pay for LeMatch and aging '00s sitcom stars. While Royal Match (RM) took its time to scale, comparing the two games' revenue per download (RPD) reveals stark contrast: RK should be hitting acceleration, and instead, it's flatlining. This is the clearest indication that a genre reached equilibrium.

Match3 is facing a 2-year paybacks (!), and with the flat RPD, it suggests RK's facing razor-thin margins. Everything about this launch screams "burned earth strategy," and with the private equity sale, something feels rotten in the state of Turkey.

Yet, the real story is unfolding outside the walls of Match3. Puzzle has been reshaped by two subgenres: Merge and 3D | Tile match (someone call Blake Robbins and Mitch Lasky!). Together, they represent over 1/4th of puzzle revenue, up from 3% in 2022, while Match3 could close the year under 50% of puzzle revenue.

These genres are not fringe experiments; instead, they are the new center of gravity. Merge Mansion defined the space, stabilized, and stalled (classic Finnish Game industry). However, the category did not stall, as China and Israel took over, with Gossip Harbor, Toy Blast, and Travel Town posting explosive growth. Travel Town demonstrated that a puzzle can absorb higher spend velocity, and Gossip Harbor surpassed many Match3 incumbents and continues to climb. Meanwhile, Toy Blast proved Tile ("match pair") is a sustainable long-run business, unlike whatever happened to Spyke Games.

This is the part Match3 incumbents failed to see. Playrix is trapped in soft launch reskins (Familyscapes? Really?) while King abandoned its 3D Match experiment before it ever had a chance. Dream pushed RK toward a male skew that never materialized, then retreated into safe territory. All three behave like incumbents facing the innovator’s dilemma. They optimize the old model instead of building the new one, so trapped in their own paradigms, they can't see the next one around the corner.

Genres do not die because players leave; they die because incumbents stop exploring the edges. They die when companies believe that the last decade’s playbook can be recycled. RK did not stumble because Dream mismanaged a campaign; it stumbled because Dream followed Match3 logic too faithfully.

Not only were there no co-signers, but more than two weeks after its release, no major CEO has come out and supported it. You're telling me Ilkka couldn't call any of his European friends on speed dial and get them to sign something? Not even someone like Mark Pincus who while not in the EU, could at least lend a voice to this? And where is Tim Sweeney? He spends all day on Twitter bitching about monopolies, but can't be bothered to spend a single tweet on the issue? As recently as Oct 11., he tweeted, "Want freedom to compete? Do business in Europe." I get he wants EU regulatory action against Apple and Google, but leaving your own industry by the wayside is something else.

The content of the letter wasn't particularly inspiring either. Why in God's name are we getting into analogies with theme parks and families of four in a policy letter? The goal is short and sweet: “jobs and coming to the table to find reasonable solutions that encourage European growth, employment, and political power.” Cooperation, not defiance - this sets up watered-down regulations that produce a symbolic win, the highest political good. Instead, the letter positions the regulation as a referendum on game monetization.

As the wonderful, totally sane, and informed LinkedIn comments on his post will tell you, Reddit sees this as its next effigy to beat, which politicians may coop. In the post-Battlefront II world, games have avoided regulatory action outside of one cranky Dutch politician by reducing salience. By reopening that debate in the letter, Illka effectively raised the corpse. While a 𝘮𝘰𝘳𝘢𝘭 𝘥𝘦𝘧𝘦𝘯𝘴𝘦 of F2P is necessary, a policy letter isn't the right place for it, and politicians don't care.

The deeper issue is that everyone in the industry knows these regulations are being drafted by people with almost no understanding of how games actually work. That doesn’t make them harmless. Bad law written by ignorant hands is still law. The correct strategy isn’t moral outrage in the press. It’s organization, data, and lobbying; the European games industry has a right to defend itself!

Economist Arthur Laffer’s original curve describes the relationship between tax revenue and tax rate. A government makes no revenue when the tax rate is 100% because no one works, and they make no tax revenue when the tax rate is 0% because there’s no collection point. With some simple assumptions, we can show that there exists some maximum point in between the 0 and 100 percent. The question is, what is that tax rate which maximizes tax revenue? This creates the interesting conclusion that in some cases, reducing the tax rate actually will increase the tax revenue, making it something obvious everyone should do. W

While we normally think the tax rates at which countries apply suggest increasing rates would increase revenue, Roblox’s platform “tax” is so high (from 70% to 80%) that it may be possible that decreases here are revenue-maximizing, especially given the maturity of the platform.

I remain perplexed that no one seems to embrace Steam’s inverse marginal brackets, where developers that earn above a certain amount face a decreased revenue share for the marginal dollar. This creates a superstar effect and has also kept key publishers on the platform while maximizing Valve’s revenue.

At the very least, we can assume that there’s some sort of elasticity effect or response to the amount of content developers provide to Roblox based on the payout rates they get. The higher the payout, the more capital they’re willing to re-invest, thereby increasing Roblox’s GDP and potentially more for base for Roblox to capture at a lower rate.

The rev share mechanism, a new exchange rate for Robux cashouts, also has an unintended consequence of acting like an effective interest rate, rewarding users who didn’t cash out, potentially decreasing future Robux cash out propensity.

The econ team has also been slaying with regional price discrimination, helping all developers, and what I’m sure is a lot of ad tech cooking.

Roblox’s continued growth and retention of aging players will start to represent a serious threat to the game industry. It’s something we need to start coming to grips with.

Roblox’s per-user monetization is lower than traditional F2P and older demo titles. The platform’s massive scale (over 40M users online at the same time) cannot be overstated. At the RDC conference, you’d have to admire the platform’s expansionary gall, and also recognize its larger meaning.

Gaming’s timeshare dominance threat has been short-form video, which incorporates bi-directional effects that games are infamous for. Viewers have agency on these platforms, which reflects an experience to them, even if it’s implicit. Yet, Western short-form video growth has slowed and settled into its long-run equilibrium. Now, it’s in the trenches for timeshare.

Roblox Moments (TikTok-style short form) captures discovery, but engages in and of itself. In fact, that needs to be the primary imperative for it to succeed. Will players want to discover and engage with content like this? Does this unlock a new Twitch-esque audience? Interestingly, the video is vertical rather than horizontal (how most games play). Maybe one of the platform’s original sins!

Other UGC platforms stagnate – it’s becoming clear that the next wave of UGC embraces AI. @Ben Thompson has been formulating a thesis that ultimately platforms like Facebook exist simply to serve ads (not untrue!), but they are increasingly becoming means independent. The entire concept of the “feed” is merely a series of feeding troughs that can be filled with various legos. It’s utterly unglamorous, but even LinkedIn now has games built into it. Facebook is also reportedly testing AI video, with impressive results. Make no mistake, this puts gaming under the lens.

The problem is that if Roblox Moments succeeds, it becomes the aggregator, owning demand and thus distribution. Already, Roblox owns the engine on which games on its platforms are built. Like the App Store, it starts as an opportunity but ultimately becomes a nightmare.

The only hope is Fortnite. It’s not looking great. Epic’s stubborn refusal to allow creators to directly add MTX to their islands prevents the UGC flywheel from starting. Monetization attracts more creators, who bring content, which attracts players, who bring monetization. While Epic’s attempts at being developer-friendly, including an increased rev share and public platform, they won’t say the magic words that let people invest more capital in learning the platform and mastering it. Indeed, it seems Epic is doubling down on big brand partnerships. But if this is the case, Tim Sweeney needs to stop responding to the annual tweet in which a Twitter user asks, “Is Fortnite a platform?” and Tim Sweeney replies, “Ask me again in a year.” It’s becoming increasingly apparent that the answer might always be “sometime”.

There’s still no game for this. We’re missing out on one of the biggest cultural phenomena of all time, especially one that originated in the East and should have a Cookie Run Kingdom spinoff within months. In case anyone is not familiar, the CEO who made these dolls is now worth more than Peter Thiel (~$22B).

At the very least, we should be having IP collabs. These have timelines of days, not months or years. And where is Fortnite? Fortnite was supposed to be our cultural vacuum, sucking in anything that’s relevant outside of games and bringing it into our world. That’s the metaverse I want to live in – the one in which we aggregate all the cultural elements and provide new ways to interact with them, fundamentally different from the other forms. Bi-directional, not omnidirectional.

2. K-Pop Demon Hunters

K-Pop Demon Hunters is now the most-viewed Netflix movie of all time, accomplished in record-time. In case anyone is also unfamiliar, another product of the East, and perhaps the one with the strongest cultural voice. Perhaps in a bizarre way, Squid Games walked so K-pop demon hunters could run. And yet, despite dominating not only the most viewed charts, the movie’s songs have crept up into the top 10 listening charts. The silence from Netflix Games is deafening. Ironic for a music-based movie.

One of the key challenges Netflix Games faces is that it’s hard to know who the winners and losers will be. Good luck trying to build a 2+ year dev cycle around new IP. The fact that Eastern culture continues its westward drift is also a reminder that perhaps Netflix’s game strategy should too.

Our only counterattack right now is Roblox, which has fostered not only the phenomenon in Grow a Garden, which continues to sustain a healthy, active user base, but also Steal A Brain Rot. Repeated hit-fostering is a crucial bellwether symbol, showing that Roblox has a chance to capture the Zeitgeist in a way that only Fortnite, Call of Duty, and Halo have previously (sorry, Nintendo stans).

The worst part is that everyone is in on the game and knows this. Part of the game is pretending that these things don’t matter. You’ll hear this over and over again: “I care about the work, not the title.” Yet, as I’ve sat on interview panels for smaller firms, you’d be absolutely gobsmacked at how much of the negotiation comes down to title.

I recall a story fondly where an individual spent many months negotiating because they had “Chief” in their title at a smaller startup and felt that taking a “Vice” title would be a step backward, despite a substantial pay increase. There’s no guilt in this request, as when you invest a lot of your life into work, title becomes identity. Look no further than LinkedIn, where many people describe themselves as an “Executive”. The issue isn’t the request, it’s the hypocrisy.

Similarly, one of the things people identify as a difference between Juniors and Seniors is that Juniors have an obsession with particular ideas. At the same time, Seniors are more idea-neutral, focusing on outcomes instead. Again, this seems like morally clairvoyant positioning: we should all care about outcomes more than whose idea was whose. However, when you’re a junior, one way to advance is by attaching yourself to a big idea and then riding it to success. For many gaming executives, who already secured their positions, outcomes matter because that’s what their incentive structure is tied to.

Founders also struggle with this. “Why isn’t everyone working as hard as I am?!” Well, because they don’t have 80% equity in the company.

People responding to incentives is a far stronger explanation of human behavior than many of the  “experience gap” explanations for this phenomenon. As Munger said, “show me the incentives, and I’ll show you the outcome.”

While both China and the United States use each other’s cultural imports with great suspicion, the one pop media category where this doesn’t seem to arouse the same suspicion is games. Yet, for all the concern that companies like TikTok bring in the United States, no one seems to be sweating the infiltration of 5-star Genshin characters in the States. The only ticket items to have even raised an eyebrow, far worse than TikTok’s data scandal, are kernel-level anti-cheat common in nearly every Chinese game and even in American-based subsidiaries like Riot. The reason for gaming’s limp soft power is that it fails to effectively communicate social norms in the same way that film and TV do.

While China’s censorship level is on another scale, the fall of the Great Firewall from a Politburo that called games the “spiritual opium of the people” is remarkable. For over a couple of years, American firms have had a free backdoor into the Chinese game market – permissionless entry! While this is a boon for games economically, it’s also another reminder that we don’t have a seat at the big kids’ table. This doesn’t seem to be unjustified either.

@Chris Heatherly’s brilliant podcast with @Erich Schwartzel discusses these points with a fascinating conversation about how the oil states are also in competition with one another (hello @Savvy Games Group). The conversation also reinforces the growing role of anime in the West. I have frequently commented on the proliferation of ramen shops in Stockholm over the last decade, and this is a universal finding when I visit other major and non-major cities. Anime is a Trojan horse for a lot of soft power, and that actually may be where China finds a foothold, and where games are relevant. Anime has quietly built the kind of bridges that games like Black Myth: Wukong, a distinct Chinese export, failed to achieve in the West.

Another example is Dragon Ball’s inspiration from the Chinese classic, Journey to the West. Dragon Ball is the number one mobile RPG in the United States by lifetime revenue, and yet Journey to the West sounds like an Oasis album to most Americans.

Ironically, the best way to counter this cultural soft power threat is to embrace the strongest potential counterbalance to it: Japanese anime. It wasn’t a mistake when miHoYo adopted the English translation of their slogan, using the Japanese word “otaku” rather than its Chinese equivalent, and it should remain this way. Hollywood was America’s greatest soft power weapon of the 20th century. If the 21st belongs to anime, then games will become more politically relevant than ever.

Billions of annual dollars for a mechanic so simple you hate that you missed it. When Royal Match introduced the additional win streak tier in 2023, @Harshal Karvande wrote a year later that it “increased revenue baseline by more than 1.5 times.” On a net present value basis, the invention of Super Light Ball will return tens of billions in recurring casual puzzle GDP increases. Pretty good for a feels-so-obvious mechanic. And yet, it took over 31 years since match3’s debut in Shariki and at least eight years since the win streaks first appeared in Candy Crush for match games to discover the obvious. Its discovery, though, makes clear that industry stagflation is a choice, not a destiny.

The most remarkable thing about the discovery of Super Light Ball is that nothing about it is a new ingredient. It’s simply introducing and doubling down on mechanics that have been in match games for years upon years. And despite the fact that UA has been ferociously bid up with paybacks pushing two years, the discovery of the mechanic was still out of reach — even for corporations like King that spend millions of dollars a year on research.

Broadly, science has been grappling with existential questions about its own growth, not unlike those that gaming has been facing. At New Things Under the Sun, @Matthew Clancy has been documenting the empirical evidence, ultimately showing, in diverse and clever ways, that yes, new ideas are becoming harder to find.

Gaming mirrors many of the findings that he observes in science. For example, gaming is becoming increasingly specialized and more narrow. Not only are PC/Console hires skeptical of mobile and vice versa, but genre specificity is becoming increasingly important. Working on Counter-Strike, applying to a job at DICE is simply more persuasive than an application from Voodoo.

However, Clancy also recognizes that harder idea spotting doesn’t necessarily mean stagflation. Counteracting declines in productivity can be achieved by increasing investment in research tools (i.e., technology). One of the things we often forget, however, is that game design is also a technology, something Nobel Prize Winner Paul Romer would likely be supportive of.

Like Romer’s technology description, design is non-rivalrous, particularly in mobile free-to-play. Romer discusses the most valuable ideas being “recipes,” which help them scale. Super Light Ball is a near-perfect demonstration of this! Only three months later, it was copied in Toon Blast as Disco Ball.

While A.I. is a topic for another discussion, game design is not a new technology; it’s one of our oldest. However, when I examine the R&D budgets of these firms, it’s almost as if they refuse to double down. King’s Experimentation Group was disbanded despite well over $100M in annual value (hard leveling labeling!), while I search for the fourth win streak.

Clawee is the most ingenious game/app/experience I’ve played in the last five years. Players control a real claw machine, with the app streaming the video as players make claw moves in real-time. Everything outside of that execution is dogshit, and you know what? That’s actually okay. Not only is it okay, it’s exactly how it should be.

When Offer Yehudai was trouting the development method of FIRST putting up $500 paywalls last year, I couldn’t endorse it loud enough. In the explore-exploit framework of modern development, way, way too many devs spend too much time digging, when they don’t even know if they’re sitting on an oil well.
One of the earliest and strongest signals is demand, despite polish, not because of it.

Everything about Clawee’s UX and monetization is utterly a mess: I honestly thought it might be adware on first open. The menus respond poorly to touch, and the UI art appears to have fallen straight out of an early 2010s iPod Touch App Store featuring. However, the company’s ability to raise funds is now significantly easier. They’re operating downhill because they have real KPIs that validate their market insight. Downhill momentum is a powerful result of the dancing pedestal monkey, too!

If you’re teaching a monkey to juggle on a pedestal, the first problem to solve isn’t getting the monkey on the pedestal; it’s teaching the monkey to juggle. If you can’t solve the hard problems first, solving the easier problems is worthless. Counterintuitively, game founders should start by asking: what is the hardest part of my thesis that I need to validate first to make this whole thing worthwhile?

Many times at The Experimentation Group, we force brutal organization of ideas. The Experimentation Framework, even if not literally applied, is extremely useful: what is the thing or feature that, if I removed from your app, would cause KPIs to collapse? That’s where your real value is. That’s the thing you need to double down on or validate.

Squad RPG provides a clear example of this, where new content waterfalls across different vectors over time. A new character might have a collection event with a distinct set of rules, or a seasonal loot box with differing odds from the normal distribution. The system designer must balance the short-run price inelasticity of new content against the more difficult-to-analyze effects of limiting the marginal ability of labor to acquire the same good.

The neoclassical Cobb-Douglas framework remains useful here: a player’s production is a function of labor and capital, and designers control the relative payoffs to players by adding an additional unit of labor and capital. However, labor pays explicitly nothing to the developer, where instead labor value is some hand-wavy value add that ends up amounting to marketing k-factor no one ever bothers to measure.

A continuing innovation is that designers have carved out parts of the economy and progression model where event engagement is a core part of player income. Income, of course, is a coupon for more progression. Monopoly Go appears to be a prime suspect here: The normal board resolutions (locations, community cards, etc) source little rolls relative to event leaderboards and milestones.

The same is likely true for Royal Match: adding Lava Quest, Sky Race, etc., is a higher LTV way to source gold to players than the end-of-round bonus based on difficulty or moves remaining. Royal Matches’ “Bonus” Levels also become interesting under this framework: should all match-3 games eliminate end-of-level rewards entirely and source gold through these levels? What about exclusively from events? The system designer challenge remerges!

Every time I’m in the mall, I wander into a toy shop hoping that video game imperialism will have finally conquered a domain that should be a layup. Yet, despite nearly 80% of U.S. kids playing games, and almost 50% on Roblox, games occupy 5-10% of the kids’ toy shelves.

It’s worse, too, as we look at the Roblox shelf, perhaps the weakest of any game IP, holding almost no original value (they’re trying to fix this with their first live ops event), and it’s instead filled with Stumble Guy toys on deep discount. Instead, games have become a vessel for other IPs, with nearly all of the IPs on the shelf having integrations back in the games, including a notable one between NERF and Stumble Guys. In the same way that Netflix rewrote the rules for Hollywood by aggregating demand and then using that to control the supply chain, games need to do the same when it comes to toys.

Angry Birds has been obliterated, Brawl Stars merchandise is nowhere to be seen, having been pulled off shelves, and Sonic has run away despite movie domination. Fortnite skews a bit older, and sometimes the odd toy or two will appear again, including Nerf crossovers. However, like Roblox, Fortnite’s IP is ethereal; in fact, it’s almost built that way. Minecraft is far more successful, and the movie reinforces that there are real IP elements. Despite this, any Lego shelf (and only the kids-targeted Lego sets) occupies nearly 4 times the shelf space. Sadly, we’re completely absent from the girl toy section. Many of the most severe criticisms of free-to-play are reinforced by examining these images. There’s simply no brand identity attached to mobile F2P games, and so it demands little memorabilia.

The exception is the Anime Heroes shelf – a new sight in the West and anecdotal evidence of anime’s growing popularity. The East has intertwined games and anime to such an extent that games have become anime, and anime has become games. It’s not just from a story perspective: the business models are far closer, too.

Gaming’s greatest strength is its ability to adapt and mold itself to any platform or IP, but we seem to have lost our ability to bend other media to us.

Where are the multi-armed bandits in games? I’ve asked myself this question repeatedly because the use cases are vast, and other tech companies have done most of the analytical and product-heavy lifting. While they exist in certain quadrants, they’ve certainly failed to scale. If we consider, for instance, the proportion of active mobile gaming users in a multi-armed bandit, it’d be hard to imagine that it has grown to any significant portion over the last five to ten years.

This isn’t super secret sauce either; the Superscale guys were singing the models’ praises years ago.

The idea behind the bandits is to identify the high-return slot machine by testing different machine pulls and then adjusting future bets to double down on winners and cut losers. This becomes “experimentation as a service” since it’s easy to add new “slot machines” to the pool and see if they emerge as “winners.”

Modelers can also set the exposure arm, i.e., how much sample the new slot gets to see if it’s a winner. It’s always interesting to see what the “open-minded” individuals put down for value here, since this amounts to an R&D budget.

Practically, seeds are a great example from match3: test a shit of variants, and select winners. This is exactly what the prior, public experiment in Bejeweled did (Dynamic Difficulty Adjustment for Maximized Engagement in Digital Games). Simply randomizing the starting board state on a single level drove as much as an 80% difference in attempts per success. Without seed stability, level designers are balancing against potentially billions of combinations. Practically, imagine trying to balance a foot race where each participant starts from a random point on the starting line.

The REALLY cool thing about what Metica is building is the allowance of multiple equilibria rather than single-winner convergence. In the seed example above, multiple seeds could emerge as winners. For example, some players may prefer (as measured by retention) more difficult or easier level seeds across the entire player lifecycle. Remember, research from Apex Legends presented at GDC highlighted how becoming too easy also leads to churn.

In a match, churn from too easy a difficulty is harder to observe than churn from difficulty. Hard difficulty churn is obvious and literally sticks out on dashboards, while easier difficulty is a subtle poison. The best solution is to add new variants, where instead of levels being measured between one another, they are measured as sequences. Think of it as evaluating the entire roller coaster rather than a linear subsection.

Frankly, Squad RPG has been in the shitter. 4x has incorporated SquadRPG faster than Squad RPG has incorporated 4x, making it unclear where the genre goes from here. Umamusume: Pretty Derby might hold some answers.

While it’s still a character collector, Pretty Derby hones the process of creating rare and unique characters. The core gameplay isn’t endless auto-complete battles and energy recharges; it’s a series of strategic choices in training one character over ~30min, ultimately determining its final stats. That trained character can also influence its offspring, not dissimilar from actual Horse Breeding. (All of the 2025 Kentucky Derby horses were descendants of Secretariat, a dominant racehorse.)

Dragon Ball Z has been the most surprising one, which is also something we cover in the podcast. But who would have guessed that this has generated the most lifetime revenue in the United States? It aligns with Pretty Derby’s accent, Nikkie, and Honkai: Western audiences will accept Eastern RPGs (Summoner’s War is practically a U.S. at this point!). More than just general appeal, the game deserves further mechanical study, particularly around its endgame (I’m currently playing it!).

The continued dominance of Raid and Star Wars Galaxy of Heroes has somehow gone under-examined, which is something Adam Smart and I take a look at in a forthcoming podcast.

Mystic Mayhem and DC Worlds Collide may look like muted launches or dead on arrival, but I think it’s fair to say publishers are increasingly using slower and slower buildups on a global scale rather than region-locking players behind. Watcher of Realms is a Moonton relaunch, so I don’t expect a ton of growth after the initial wave.

Finally, there’s the curious case of Dark Legion, which appears to be holding on. Next month will be extremely important for it. It represents a new hybrid breed of Squad RPG with stronger, more explicit, and direct 4X elements than we might find in, say, Star Wars: Galaxy of Heroes, which uses 4X elements but not quite as literally.

Of course, there’s the true reflection of this trend in dual-SKU products like Genshin Impact or, for a brief period, Call of Duty: Warzone. It’s also worth noting that dual-SKU titles like Fortnite and Roblox are impressive data points, even though both have a single-platform dominant share.

To the extent that they’ve unified, it has been a result of the model shift, where long-term monetization via service-based mechanics maximizes revenue. Internally, we’ve also seen this reflected in labor profiles. It’s very common for PC and console developers to hire now Directors of Product: roles that didn’t exist 10-15 years ago. Specific monetization and progression designers are also now essential, even for single-player titles.

Mobile also started to become friendlier to strong creatives. If you don’t know the founding force behind Monopoly Go, he falls under this category, so certainly would Second Dinner founder Ben Brode.

All of these cases still seem more like proof of concepts than actual paradigm shifts, though, or any evidence of a convergence moment between the two platforms. The big wedge, the divide that doesn’t seem to be crossed, is user acquisition. The most compelling evidence for this is the difference in the budget shares between the two projects dedicated to acquiring users.

Downfunnel, this doesn’t just translate to budgets and user acquisition specialists. It also forces mobile studios to have their analytics house in order. Mistakenly, mobile gets referred to as a “margin game”. Of course, PC and console is a margin game too. However, in mobile, cohort-level cost-benefit analysis is not only possible, but it’s also the essential ingredient that scales marketing.

I’ve rarely seen PC and console developers primarily focus on metrics on a cohort basis. Instead, load the home page of most PC and console title analytic dashboards, and the first metric will be something like week-over-week retention.

Many startups are working to add cohort-level marketing attribution to PC & console. There’s no reason, for instance, that you couldn’t install the Adjust SDK into a PC or console game. Indeed, I’ve seen some developers do this! However, the results never seem reliable, and the CPIs remain inconsistent. Remember that the funnel challenges for PC/console are much greater than direct on-device links to complete a download on the trusted App Store. If this gets solved, singularity awaits.

Its share price is trading at near all-time highs despite flat quarterly earnings. And now, Call of Duty (CoD) looks weakest: not so much as a whisper for an annual franchise built on top of a marketing war machine. While heated early 2000s debates emerged between CoD and BF fans, CoD ultimately emerged victorious, for whatever that means. However, the one time BF won was when CoD was at its weakest (Advanced Warfare) and BF was at its strongest (Battlefield One). There’s a very real possibility that history will repeat itself here.

While it may seem like returning BF to its former glory is an eternal mission, not a recent one, EA has finally made the necessary investment in the studio ecosystem (see “Battlefield Studios” branding). Bobby Kotick’s most profitable improvement was to set up a series of cascading CoD studios to reach annual releases. Each studio would have a three-year development cycle, while still adhering to the annual release cycle. It was a brilliant strategy, and EA briefly flirted with a frail imitation in BF: Hardline. Again, nowhere near the resources of CoD. 

BF has struggled to adapt to the modern live service climate, which anyone close to the franchise, including someone who worked there myself, would be the first to admit. However, EA has admitted wearing the studio thin, internally acknowledging the burnout and tech debt. Over 3 years, DICE shipped:

• Star Wars Battlefront 1
• Mirror’s Edge Catalyst
• BF 1
• Star Wars Battlefront 2
• BF V

On a profit per headcount basis, DICE must have neared the AAA HD equivalent of Supercell (is that an oxymoron?). While it received support and studio help on nearly all these projects, it falls short in comparison to the sheer size of CoD’s operations. Simply compare the two start screens, and find eight studio logos in CoD: BO6, and four in Battlefield 2042.

While it may seem like pouring hundreds of millions of dollars of additional investment into a fading franchise is risky (rumors swirl the game’s cost exceeds $500M), the capital is a risk minimizer. This is what it takes to win in gaming’s number one genre, while the fruits of success outweigh the costs! Activision and King were built on singular franchises. EA must absorb the cost and minimize operational risk as much as possible. 

It pains my heart to see the brain trust of Battlefield move away from Sweden to California, especially since this is the franchise that helped build the Swedish games industry, but the decision is understandable. I vaguely recall a Swedish games industry report mentioning that something like 30% of all Swedish game development was concentrated in a couple of square blocks of DICE’s office. It’s certainly not over for DICE, but it’s increasingly looking like a supply chain member rather than an organizer.

Nearly the entire growth of gaming over the last decade has been a direct result of mobile and, subsequently, the free-to-play model. And while revenue may be equal between mobile and PC consoles, players are certainly not, with mobile F2P active players likely exceeding 60-65% global player platform share. In the U.S. alone, 70% of smartphone owners play games. The massive growth in gamers from the free-to-play model has been nothing short of astounding.

Of course, the East is an entirely different situation with the current Politburo alternating between calling gaming “spiritual opium” and allowing global versions of Steam games through the Great Firewall.

Gaming simply won on its own playbook: eXplore, eXpand, eXploit, eXterminate. The free-to-play model turned everyone’s mom into a match-3 lover (the modal match-3 player is a 37-year-old woman)! Not every 18-34 male was a gamer in the decades prior, something reserved for “geeks.” Now, every football player hops on Call of Duty after practice.

A decrease in global and U.S. violence levels, along with increased play during the pandemic, has also helped the video game entertainment industry.

Roblox’s relatively uncontroversial rise to nearly half of U.S. kids still remains a remarkable achievement. In many cases, the parents of these kids are what we might call gaming’s first parental generation. They understand the difference between risks from games, risks from the internet, and what’s novel versus new.

Cultural video game critics need to get their facts right if they want to be taken seriously. In this piece alone, the author mistakenly claims Fortnite monetizes gameplay, a crucial distinction (“Success often depends on purchasing weapon upgrades, skins, and battle passes, making ongoing spending feel necessary”). There are reasonable discussions to be had about the costs and benefits of modern gaming to adolescents; however, every time an editorial-type piece is published on this topic, it comes across as activists needing to gasp for breath between each shock & outrage example. There’s an overwhelming use of oddball survey data, and obscure examples instead of real and public sources. For example, Roblox publicly reports $20 bookings per monthly unique payer. There! Evil monetization schema resolved!on schema resolved!

The rise of Monopoly Go! was unperceived not only in games, but also in all of entertainment and nearly the tech industry. However, its stagnation is hindering its teenage years, and regression is looming. Broadly, all of Scopely is in the red, with Monopoly Go grasping onto revenue as downloads decline.

And yet, all of a sudden, when we re-examine its portfolio, the chess game becomes more apparent. A lion’s share of its growth has come through acquisition, and the impact of Pokémon Go on the company cannot be understated.  

This seems to have come together slowly, as Scopley was hesitant to marry any studio before a 2nd party date. Scopely made many of the right moves when it came to early thinking around IP in mobile, including its relationship to cost per install, the role of live operations, and what pains me to admit is what CEO Javier Ferreira called the “Scopely OS.” For all these right moves, however, it took until Monopoly Go to achieve breakout success, which they not only seized but also pushed to its marketing limit.

However, whenever Monopoly Go sneezes, the company will now catch a cold. So much of Scopley’s success hinges on Monopoly Go growing. Instead, it’s shrinking. Even just a 1% change here more than makes up for 1% across every other game in the portfolio combined. While web stores may help, the decline in both downloads and MAU is telling and suggests there’s real pain across the portfolio.

The second-order priority has to be applying the “Scopely OS” to Pokémon Go. It was one of the few titles to grow year over year, and it needs to grow even more to not only mitigate Monopoly Go’s potential decline but also counter the looming threat of Pokémon TCG Pocket, which may be driving up the cost of installs.

The portfolio also tells a fascinating story of how only getting even one idea right is enough to subsidize many mediocre results. Similar to the story of Candy Crush or even Clash of Clans, you can ride forever franchises for…forever? Seeming slam dunks like Scrabble Go or Dice with Ellen have disappointed, while co-CEO Walter’s white whale remains elusive: a top match-3 title. Stumble Guys can only be considered a disappointment even on substantial ad revenue, given declining downloads and MAU.

Scopely is now the custodian of two of the top-grossing mobile games of all time. It remains up to them to cement them as forever franchises that have yet to see their best years, just as games like Candy Crush or Clash of Clans continue to deliver all-time great results to this day.

When making a completely unfair comparison between income inequality in developed nations and the top mobile puzzle and RPGs in the U.S., we find a couple of interesting things.

1. Unsurprisingly, RPGs have far higher Gini coefficients relative to Puzzle. The Gini coefficient here is the daily revenue inequality, i.e., how unequal are some days in revenue contribution relative to others. A Gini of 0 means each day has the same revenue contribution over the last 360 days, while a score of one indicates a single day captures the entire contribution over the last 360 days.

2. Two-year match paybacks are built on remarkably stable performance. Toy Blast scored a Gini of .074, near 0! It’s a whopping 86% difference from the Fate / Grand Order. Gacha games are content-sensitive, and it shows here. However, the constant levels were low; the third most unequal game (Pokémon Go) only scored a 0.27 Gini.

3. Most countries have higher income inequality than the U.S. mobile games have daily income inequality. This means nothing, but hey, why not? I’ve measured this for spender amounts within a game, and depending on the denominator (revenue/{downloads}{active users}{active customers}), the Gini coefficient ranges from 90 to 99%.

4. Match’s multiple and attractiveness to private equity make more sense after looking at this. Raid: Shadow Legends has about 2x the daily revenue inequality of a Toy Blast.

Tonight, I’m lighting my annual memorial candle to honor the 8th anniversary of the passing of the Mario Lopez one-season Candy Crush show on CBS.

In nine glorious episodes, seemingly longer than most CMO tenures, contestants flailed at wall-sized smartphones in what Lopez dubbed “full contact Crush.”

What many big mobile gaming brands, such as Monopoly Go and Royal Kingdom, misunderstand is that TV ads featuring celebrities are not enough. Like Riot’s Arcane, the key is to integrate the product so profoundly into the ad that it’s practically invisible. And we can’t just count the download lift, proper attribution includes the real-time increase in spend during the show’s timeslot airing.

Look forward to next month, when we’ll be holding another memorial for the 9th anniversary of the passing of Dice with Ellen.

My new Figma plugin, Economy Flow Builder, is now publicly available on the Figma plugin store. This makes it easy to convert and order game economy flows via JSON, automatically converting them into charts. Finally, reproducibility and jumpstarting of economic flow diagrams.

The kicker is that it also syncs back to JSON. If you make changes, such as deleting some boxes, creating new ones, or reorganizing flows, you can sync them back to JSON with a single button click. Tons of nice features in the JSON too. For instance, the ability to create subsections. So if you’re creating a giant economy flow diagram and there’s a particular feature that has a sub-economy, it makes it easy to visually designate this without adding a ton of visual clutter.

I created some dummy examples for presets, but since the plug-in is open-source, I would highly encourage people to contribute back to the community and upload JSON file formats so they can serve as presets for game economies they’ve worked on or are passionate about. But oh boy, wouldn’t it be nice if you could use something like Gemini’s Deep Research to break down a game economy into this JSON format, and automatically pipeline it into the plugin?

Damn, I hope someone works on that next.

I also created an R package for accessing the Fortnite API, but currently, there is no way to retrieve the top islands. If you know what specific island you want to look at, you can get great metrics in one line of code. However, I’m still waiting for Epic to update this one to add a bit more substance! Nonetheless, it’s remarkable that Epic released a public API that includes KPIs likr as retention, unique players, and peak CCU. A huge step forward.

In this case, the primary way to become more powerful was to level up, which required XP, which required combat. Other mechanics, like world map bounties, kept the bloodbath going. Another combat-focused mechanic was the ability to respond to enemy marches with a reinforcement that could only be deployed in the window of the march. The game explored strong (but now standard) 4x price points with content that other games are still hesitant to try, such as ALL Legendary Gold and Army items at $700+ points (web store).

For its part, Kingdom Maker, which was published by Scopely and developed by Global Worldwide, reintroduced sex into 4X (also some humor). Why can’t you create dynasties? It was a fascinating game thesis that lent itself to a literal game title.

However, the mobile market runs on highly observable and reactive margins. With marketing now consuming 50% of budgets, the gains from optimization are literally halved, while surprisingly, content innovation is also less. Both of these games had some fascinating ideas, but ultimately, if you don’t have a way to connect that to genre economics, it’s hard to scale.

Gaming industry people often have an inferiority complex around “tech”, feeling never quite properly understood, and usually looked down upon. I understand the frustration!

Replit hitting $100M ARR, over two explosive growth quarters, is hailed as a remarkable achievement. However, when we compare it against Monopoly Go (MoPoGo), which achieved multiples of Replit’s actual quarterly revenue over the same acceleration period, it falls much flatter.

Much of this is actually a MoPoGo story, which has experienced an acceleration towards an apex and is now “reverting to the mean.”

The rapid rise of MoPoGo is unprecedented in mobile games; it’s hard to think of anything that has scaled so fast, reaching heights higher than Game of War or Clash of Clans in their heyday. It’s certainly been reported on, but somehow still underappreciated.

The clones and genre expansion are getting into full gear: even Moon Active is launching “Coin Master Board”, while Fish of Fortune is the most promising game I’ve followed in a while.

1. This has the best fundamental mecha core of any title, with a great integration of the Z-axis play. Lock-on makes the controller and cross-play easier, and it supports 3rd-person gameplay.
   
2. This obsession with extraction needs to stop, even if the monetization is interesting. Nearly all of the gameplay modifiers, including the vertical progression mods, exist only in the Extraction mode, ripped from the PvP after Reddit complaints. The extraction mode features paid consumables that allow for early gear extraction, thereby acting like insurance in case of death (also seen in Call of Duty’s DMZ).
   
3. The system’s design is poor and will likely sink the game. Unlocking three mechs over three months through a long grind with no gameplay upgrades or progression in the PvP mode is a poor response to pay-to-win concerns. Similar to Overwatch’s introduction of earn-only loot boxes, throwing out the mods system is the baby, if the hard currency is the bathwater.
   
4. Bandai failing to establish a live-service Gundam game boggles the mind. Robots are a layup for F2P design. WAR ROBOTS has done over a billion in lifetime franchise revenue. How is there no “definitive” mobile Gundam squad RPG? Only recently has the license been opened up to anyone, and the results have been disappointing, often resulting in poor Chinese clones.
   
5. Marketplaces have expanded, but this one could have a deflationary spiral. Marketplace items are priced in hard currency, so the only currency inflow is MTX payments. Players spending hard currency on 1st-party items and the 5% transaction tax will drain currency from the economy. If the drain rate is greater than the source over the long run, prices will start to rise, holding all else constant (the currency becomes more scarce). The expectation of future price increases incentivizes players to hold goods off the market…thereby setting the spiral in motion. They will need more stable sources to peg prices against!

Like many, I believe Marvel Snap is a magnum opus in CCG design, creating the first true mobile CCG. But the ultimate reality is that revenue collapsed. Thirty-three months after launch, it’s nearly the worst-performing Marvel game by revenue. The hypothesis of a sticky core hasn’t materialized as Sensor Tower estimates that active users have declined at a similar rate to revenue.

I’ve previously covered the struggles of their progression system, as mechanical band-aids after band-aids have arrived, but I’ve wondered if this is just a more fundamental mismatch. Despite a phony core, Pokémon TCG pocket, a loot box simulator, will pass Pokémon Go in cumulative revenue on a launch basis. I certainly didn’t expect this, and was wrong about Pocket! It’s a remarkable achievement driven by 5x per user monetization on 4x the user base, almost exclusively Japanese players alone.

CCGs function like lifestyle games, and inserting a new one has nearly always been met with failure, with 20-year-old bodies littering the genre graveyard while web3 arrives with new bodies daily. Like MOBAs, TCGs have a “centre of gravity” around players’ lives, and it’s no wonder that their physical versions still go strong. Disney Lorcana TCG’s eventual mobile launch may be anchored in a similar concept, while I imagine Yu-Gi-Oh! and One Piece can’t be far behind.

Apple’s new EU rules are receiving less press because the U.S. trial gave away the game. These are desperate measures to prolong a high-margin revenue stream long enough to dump the consequences onto another Apple executive. The EU will surely act; the only question is whether it’ll be able to protect itself… from itself.

The EU has hindered itself by focusing on destructive regulatory frameworks, prioritizing ex ante (before acting) regulation over ex post (after the fact) regulation, and positive rights (the obligation to provide a service) versus negative rights (the inability to create a restriction). There’s no better example than the Digital Markets Act (ok, maybe GDPR), which requires platforms to build new technology around interoperability API layers.

Part of the requirement is a new licensing schema where Digital Gatekeepers (another new legal status, that may change based on active users) need to justify every rejection for API access, in addition to providing certain KPIs. The rules require certain features for applicants, which, of course, the gatekeepers already possess, having happily complied with GDPR.

Not all regulation is created equal, a fact that often gets lost in modern political discourse. In the U.S., private actors (Epic) were able to press the case rather than a Commission Investigation (could there be a more bureaucratic term?). Instead of the DMA’s arbitrary 10% global revenue fine, which is based on a set of rules with no precedent clarification, there is also clear injunctive relief in the United States with modest pay-the-loser legal fees. Nothing could be worse than an outcome that incentivizes uncertainty and punishes with significant financial penalties.

This is becoming apparent as Apple has paused the iPhone mirroring from EU-based macOS devices. What is a “Core Platform Service”, which would require interoperability, and what isn’t? What Mac features now need to be ported to Windows? Another EU Commission Investigation will surely clarify.

Yat Siu, Co-founder and Executive Chairman of Animoca Brands, steps cast to defend Web3 against @Eric and Phillip Black vigorous skepticism. Christopher Kaczmarczyk-Smith just wants to know why gamers don’t get it. Is Web2 fundamentally incapable of grasping the promise of open markets? What is and should be promised to token holders?

We discuss:

• Laying down Web3’s steelman case
• Why the West still doesn’t get Web3 like the East
• Examining the original token sin, where did it all go wrong?
• Do digital property rights actually hold back economic growth?

Listen Here
Watch Here

1. Game Founders: Get In Front of the Camera

One team had success marketing to Gen Z by going direct. This is another strategy that Lulu Cheng (fmr. Chief Communication Officer ATVI) advocates, but for different aims. One for founders to remain in the driver’s seat of their own story, while another is about great marketing creative that drives authenticity.

I rarely see game founders take the lead on either of these dimensions. To the extent game founders “talk to customers”, it’s usually customers talking at them rather than proactively seeking to extract from random samples. UXR is useful here, but it’s one data point, just as talking to customers is another.

The great thing about marketing creative is that it’s cheap to try, so there’s not a lot to lose. At worst, you experience the entire marketing funnel hands-on.

2. AI is Still A Supply-Side Phenomenon

AI is justifiably everywhere, but the discussion remains on improving existing workflows. It’s telling we haven’t seen massive increases in quantity yet, however. What game is shipping 4x as many cosmetics, or even 2x as many match-3 levels, because of AI?

Early game developers were engineers, not game designers, who essentially tinkered with technology. I see few engineers doing the same to utilize AI to create unique experiences. Even the obvious stuff, like an AI-powered Nintendogs, remains a year away.

3. Crazy Rich Asians Nailed It

Singapore is what was promised. I’ve never seen a greener city, and possibly never better transit. Urbanists: there’s a functioning cable car system that fills crucial transportation cap. That includes things like a monorail. Gardens by the Bay is as green as the film portrayed, and includes daily night light shows.

The food is top-tier and includes an incredible mix of Korean, Japanese, Indian, and Chinese cuisine. Biryani was out of this world, Pork Rib soup was savory, and may we rendezvous Chicken Rice.

The most valuable economy diagram anchors in player action, where currencies play a secondary role. All diagrams start with either spending time or money: the two raw ingredients that all output derives from in games. Each further step is an action-driven refinement of the last stage. Along the way, usually deriving from the final production step, is a “final” good. This is ultimately what entertainment produces, or the “bedrock”, akin to appealing to a Bartle-type.

Goods with Green + or Red – are currencies that maintain debits and credits, while player XP is orange (a store of value), since players can’t “spend” XP. A simple tally of red and green boxes along a particular currency type usually makes it obvious where there might be currency imbalances, or worse, contagion effects from one activity centre to the reward centre of another. The core loop almost “pops” out of the process, as final inputs may materialize into raw input enhancers.

AI, specifically ChatGPT’s Deep Research product, has been a boon for creating more rules-based output. Economy design has been slower in adopting AI, as the integration of Excel and Sheets is laughable, and the vibe game economy is still more meme than reality. I’ve gotten lackluster design feedback from ChatGPT, which makes sense given how little design philosophy is codified relative to other fields.

However, given the model described above, Deep Research can repeat distilling information into a standardized format. We can create an AI Economy visualization pipeline with a Figma plug-in, which is still very much in Alpha (and linked here). It works by having Deep Research create a JSON file, which converts a custom Figma plug-in into the economy model visualization. 

Game economies are opaque, but I’ve become increasingly convinced that AI will play a larger role in mapping them out. For example, an automated agent could easily map the top 100 game economies along this ruleset.

1. Deck Loadout Everywhere

Due to short mobile session times, there is insufficient time to build depth through in-round progression. Marvel Snap is a goddamn gift from the CCG gods in this light, and remains the top tier benchmark. Players must be able to theory craft, which rarely happens in real-time. In auto chess, players have specific deployment phases, which is also true for CCGs, MOBAs, and Vampire Survivors. Players also enter with a known strategy.

Squad League gives players this choice, but it’s limited to one game mode and buried in the UX. The experience should focus on building lineups and encouraging a specific playstyle.

2. Alliances

Building on autochess mechanics, my blop of units creates unclear strategy readability for me and my opponent. Alliance abilities are granted only when a certain number of units share an alliance trait, improving  strategy readability.

3. 70% of squad abilities are conditional passives

The abilities are boring. Heroes and Squaddies need more if/then-type abilities. Players want a match runbook, and maximizing triggers based on specific real-time gameplay patterns adds an enormous layer of depth.

4. Squads survive encounters stronger, not weaker.

In Squad Busters, players face off against multiple enemy teams rather than a single enemy. Even after a successful encounter, other players attack as they lick their wounds. This typical play pattern in Apex Legends mirrors packs of wolves (three-man squads) roaming for weakened prey.

This difference is that Apex’s victory over prey means power increases through a leveling system or the dead enemy’s loot. Squad Busters features only a single chest key for defeated enemies and a mad dash for gems that attracts MORE enemy attention.

5. Drop Gems or Chests

In Squad Busters, some modes determine ranking by the number of gems players collect from PvE enemies and killed player units. The game also utilizes coins, which can be used to purchase chests that unlock powerful units. It’s too much to manage and too hard to cut one or the other from the herd. Maybe gems automatically level up the Squad (no more chests), or Last Man Standing becomes the default mode.

6. Single health bar and improved animation loops

Readability remains a pressing issue, with numerous health bars to monitor. A single bar that’s the sum of all my units ‘ health and my own gives me only two points of focus during combat: my health bar and the enemies. Blinking red units, the death sound effect, or other indicators can still provide more subtle feedback on particular unit status.

Squad Busters’ original thesis has been falsified. What once seemed like KPI paradise: hypercasual-esque core, aggregating Supercell’s beloved IPs à la Super Smash Bros., and infinite consumable-based monetization with zero marginal cost, is now unrecognizable. I was brought in, and called it the game with Supercell’s biggest ceiling! Yet, just a year later, the hypercasual core has vanished, giving way to multiple player-activated abilities and completely removing consumables. Squad Buster 2.0 attempts to simplify nagging launch criticisms, including combat readability, character UX, progression speed, agency out-of-round with unit selection, and agency in-round with abilities. It gets halfway there, but in solving some issues, surfaces others.

The original mechanics captured contemporary design zeitgeist: merging, star systems, in-round progression and economy, roguelike, consumables, hero collections, and mods. It’s a system designer’s dream, but it’s spun the game into a series of interlocking parts. It lacked the iterative clarity evident in SpaceX’s Raptor 3 engine (see below), illustrating the ideal design process of stripping complexity to reveal elegant simplicity. While 2.0 reaches for this goal, intertwined elements complicate straightforward refinement.

For example, stomping out an opposing player is easier: simply kill the hero unit. But this means opposing squads sit their squad atop yours, negating the reduction in total units to help combat readability. The two new units spawn with Hero units, helping speed up in-round progression, meaning faster combat, but also a more frenzied reminder that the game lacks any comeback mechanics. “Squaddie” UX is far simpler: opening an in-round chest emphasizes the single ability addition of the character to the squad. Yet, this reminds players how mundane abilities are, with few conditional passives, limiting how squad choice in-round finally changes strategic gameplay. Theory crafting remains a sore spot, particularly for a game about amassing roguelike abilities. 

Squad Busters languishes beneath Supercell’s smallest title, Hay Day, and the gap widens. Last week, Hay Day generated $4.4M, while Squad Busters came off as a rounding error at $.4M. Its survival hinges less on issues like monetization or retention, and more fundamentally on a core something worth saving.

Squad Busters was set to catapult Supercell into a record-breaking year and trailblaze the path toward an IPO. Instead, it’s fought for its life, becoming a litmus test for Supercell’s Next Chapter. Weakness hides in numbers, and as Supercell scales and vested interests around resourcing increase, killing live games with a funeral and champagne (actually) becomes more fraught. Ultimately, Squad Busters 2.0 finds a clearer identity but fails to prove it’s a compelling identity.

Overwatch’s new Stadium mode officially marks the launch* of Overwatch 2, building on the previous release’s perk system. Inside Stadium, there’s one content productivity hack, known previously only to the savviest Steam roguelike indie developers: conditional passives.

Retention is the sum of each piece of content multiplied by the level of retentive units that content adds. For example, a profile picture may be low-cost but will likely add only a few units of retention per profile picture. In a world where titles like Call of Duty employ thousands of developers, the counter-thesis can’t be to just produce smarter, but under an entirely different system instead. 

FPS developers quickly realized that map packs are poor retentive units producers. Packs typically resulted in a 2-3 week bump in week-over-week retention before resuming long-run equilibrium. The framework was busted too: the value of maps depended on having other players who owned them to play together. Guns didn’t work either: shooter players usually only equip two weapons, so it’s hard for them to utilize more during a match. Players are always solving for the optimal gun, so like maps, it’s a brief love affair before resuming their previous meta loadout.

Introducing characters in Shooters seemingly altered the equation: players must understand each character’s advantage and how other choices affected them. New characters create choice ripples! Balance became paramount, and cheap content to produce. Balance changes amount to an equilibrium reshuffle in PvP environments; the search for dominant strategies in an ever-shifting equilibrium is the game itself.

Stat changes are cheap with near-zero marginal cost: an item with +10 health comes for free after an item with +5 exists. However, marginal benefit exists as a system. If gameplay comes from theory crafting, building a Rube Goldberg machine is far more exciting for the brain. Inside one system, a +10 health item might produce a little retention, while another might produce a lot.

Stadium’s hack is “powers” or conditional passives: low‑cost stat bonuses that activate when a player does something they already do – jump, shoot from behind, etc. Because the trigger is embedded in normal play, the passives feel rewarding without adding mental load. Each passive is cheap to make yet interacts with dozens of existing abilities, maps, and team compositions, giving exponential strategic and gameplay variety. The result is a modular and scalable system that, most importantly, produces many retentive units.

Productive systems beat expensive assets. Marvel Rivals’ six-week character releases seem ungodly to Western developers, but it remains to be seen if this supply-side assault attracts retention. In the meantime, Overwatch appears to have set up a longer-retentive engine at a fraction of the marginal developer cost.

Pokémon’s patent of spherical objects throwing of cartoon creatures threatens Palword’s lifeblood, while Tim Sweeney has lifted, at least a percentage point, in total gaming GDP with its injunction success.

• How does Apple’s rent-seeking rate change in the face of this ruling?
• Should Apple lower its rate to 15%, like it did in subscriptions?
  • Remember, it faced competition primarily from “webstores” too.

We premier a new segment: SOLVE that for EQUILIBRIUM. We discuss the marginal monetization effects and debate the benefits of personalization opportunities (hint: there are none) with webstores. @Chris is intrigued by Joost’s piece on rising game costs, while AI’s effects on the industry are measured in the Solow model. @Phil insists rising game costs mean rising revenue and stable margins, while Eric has his own doubts.

Eric’s on IP Laws, Joost’s On Gaming Costs

@Blake Robbins and @Mitch Lasky argue in Gamecraft Ep.18, “The Mobile Gaming Duopoly,” that Apple and Google have squeezed innovation out of mobile gaming. Mobile innovation hasn’t vanished; instead, it has splintered. And while it didn’t vanish, it has slowed, but the culprit isn’t IAP or a top-grossing chart—it’s ATT.

Robbins and Latsky state, “There’s this race to the highest ROI category…casino, match3, etc. Because the UA spend required to bring those games to mass market scale forces you out of innovative content and into these tried-and-true play patterns.” Yet, this is how *all* markets work!

As capital expenditure rises, it seeks less risky returns. Venture capital is a perfect example; investors need more market traction at each successive investment stage to write bigger checks. “Tried-and-true” genres like match-3 (M3) run dry of innovation as they mature. Instead, innovation inherently evolves through specialization. M3 is the prime example! M3 naturally splintered into specialized subgenres over the past five years, like 3D match, match-2, and match-pair (e.g., Tile Busters). 

An evolution into specialization isn’t unique to mobile; it’s how all knowledge works. Philosophy first covered the entire range of knowledge before splitting into natural philosophy and political philosophy. Natural philosophy splintered into math, astronomy, chemistry, etc. Specialization is not stagnation; it is the expected end state when ideas diffuse and the easy wins are mined.

Robbins and Latsky go on to discuss the “dark side of F2P”: “whale-hunting.” Conversely, the hypertargeting revolution initiated by Facebook and enabled on mobile via IDFA fostered small, niche businesses, like Dollar Shave Club, and game genres like style décor. Without hypertargeting, these genres were deprived of the oxygen they needed to survive. The result has been a decline in the genre’s existing titles and the failure to launch of others (RIP Drest). 

Mobile innovation first splintered, like all knowledge, but it’s ATT, not whale hunting or top-grossing, that’s impeding further innovative growth. Reducing platform fees might help everyone’s margins, but won’t spark new innovation.

Developers: you should ship updates to your in-app webstore UX TODAY: side-by-side pricing, knocking down scare screens, and an all-out webstore offensive the likes of which Stash and Xsolla have never seen.

Tim Sweeney’s vendetta and suit against Apple’s alleged antitrust violations have dragged on for four years. After the 2021 ruling and failed appeals, the issue was largely closed: Epic lost on most accounts but won in reducing anti-steering provisions, specifically Apple’s refusal to allow developers to link to any URL containing a non-IAP payment method. These bystentine regulations forced experiences like Formula 1 to split into a standard app, which occasionally links to web content, and a TV app where users may IAP to buy a season subscription, and no web links. The issue, of course, is that the F1 website offers non-IAP subscriptions, which reduces Apple’s revenue.

Judge Gonzales Rogers’ original injunction ruling stipulated that Apple must allow apps to provide users with external links to pay developers directly for services. Just as necessary, developers could use the email addresses acquired in the app to communicate with users about webstore payment offerings. AFK Journey was one of the first major titles to take advantage of this, using WhatsApp numbers from sign-up to send users web offers via text.

Apple’s response to the injunction allowing webstore link outs was an “entitlement application” that forced apps with external links to webstores to pay 27% revenue share instead of the standard 30%—gee whiz. Oh, and it’s worse. Apple claimed it was entitled to ANY revenue a user makes on a webstore seven days after tapping the link. If a user taps the webstore in-game on Monday and then accesses the store via a desktop browser on Friday, Apple claims they’re entitled to 30% of any revenue the user generates. Yes, you read that right. Apple used the ruling to advance its position, not inhibit it.

The drama doesn’t stop there: as part of the entitlement program, Apple openly discussed amping up a “scare” screen so users would be deterred from proceeding to webstores. Court documents revealed how UX designers discussed design particulars: “if we want to ‘scare’ users a bit, I like the addition of ‘out’ because it raises questions and hesitancy haha. out? out where? omg what do i do?”

Apple executives also lied under oath about the intent of specific policies, which was backed up by internal Apple meeting notes and emails. The judge has recommended that the case be examined for criminal proceedings.

The court’s response to Apple’s shenanigans? “This Court will not play ‘whack-a-mole,’ nor will it tolerate further delay.” The new ruling, in the attached screenshot, is effective immediately.

Habby (Happy + Hobby) rewrote mobile’s growth playbook when Archero exploded in 2019, and the studio has remained a case study in scaling hybridcasual hits ever since. The game coined “hybridcasual,” surpassed $500M in revenue, and, more importantly, created a meta sturdy enough to be recycled in every project that followed. Despite a surge in hypercasual puzzle games over the last two years, casual developers have yet to adopt the Habby Meta.

The “Hobby Core” layers roguelike ability drafting, evo‑fusion gear, elemental Roshambo, and a try‑fail‑progress saga over whatever moment‑to‑moment gameplay is compelling. Capybara Go and Archero 2 are now scaling the same meta to nearly $200m+ a year games, when Android China and ad revenue kick in. Habby is poised to break into the global top‑twenty publishers on the strength of a single, extensible meta.

Since the meta is established, it’s allowed them to take more shots on core-gameplay-goal without breaking things. If a game succeeds, they know their meta is sustainable enough to avoid economic collapse – a painful lesson from the original Archero’s mess of a long-run meta.

Wittle Defender, in soft launch, is a “Habby‑ized” iteration of The Tower, while Capybara Go borrows many loops popularised by Monopoly Go (MoPoGo) but with each grafting on the usual meta, plugging long-run progression holes present in both The Tower and MoPoGo. Each experiment swaps in a new core mechanic but falls back on the same progression rails the moment players leave the FTUE.

Rivals have tried partial imitations. Dream Games’ Royal Match solidified the streak-meta, and hypercasual was quick to adopt the notion. Yet, no single hypercasual developer has pursued this meta with the same steadfastness and veracity of Habby. Rollic comes close, but Twisted Tangle and Screw Jam feel as though two different teams passed source code across the hallway on a Friday night. This is often the case with hypercasual games, where third-party publishing is common, and even first-party teams are decentralized. The unreal publishing speed doesn’t help with centralization either. A “meta SDK” feels like an opportunity to unite desperate games across a meta that can be consistently developed between game launches.

If this sounds obvious, remember how Habby built Archero with off‑the‑shelf Unity art to keep costs microscopic until the model proved out. ROI, not polish, was the gating KPI. That discipline, spending on the parts that repeat, and ruthlessly minimizing on the parts that do not, remains the firm’s true superpower. Given hyper’s future is hybrid, the first step is building a meta that survives and grows with each iteration.

Nothing is as misunderstood as loot boxes in gaming (virtual currency is a close second). Critics decry them as a normalization of gambling that preys on weak impulse control. Strip away the moral panic, and a sanguine explanation emerges: loot boxes solve two vexing economic problems.

Games create heterogeneous value among players, and second‑price auctions reveal it. Magic: The Gathering cards and Counter‑Strike weapon skins routinely fetch five- or even six‑figure sums. However, while auction houses help maximize price, they exclude large portions of the player base from participation. Loot boxes address this inequality problem by correlating reservation price (willingness-to-pay) with actual price.

Like any lottery, loot boxes fractionalize value by creating a chance-based “ticket”. Many games extend this to “shards” or token-like “pieces” of character that may be summoned when players collect a particular amount. This is a common economic value distribution system in many other facets of life, such as the NBA draft, charter school admissions, or the distribution of the Switch 2.

The second reason is simply the reduction in cost from a loot-box store (pun intended) versus a direct store. Selling a mystery box is operationally cheaper than running a forever‑changing shop. “Here’s a pack—take it or leave it” trumps curating endless bundles and flash sales. The contrast between Overwatch 1’s vending‑machine interface and Overwatch 2’s department store tells the story of one click versus cognitive overload. Players face a single decision: yes or no, while developers save on store logic.

Furthermore, nearly all popular loot box systems give players direct choice, slashing costs for top “menu items.” I.e., when a loot-box-based game like Apex offers crafting currency in loot boxes, players craft their most preferred item, then the second preferred item, etc., as they collect currency.

Loot box critics have struggled to acknowledge the functional intent, but understanding this purpose is crucial to any meaningful dialogue about responsible monetization.

Two troubling signs for gaming: First, Danielle Tran from Konvoy reports that AI accounts for only 10% of gaming investments, compared with 71% across all venture activity. Do VCs believe AI will transform every industry except gaming? AI is a growth mechanic, and the idea that gaming won’t benefit as much is a tough pill to swallow. But it’s not necessarily wrong, as our relationship to entertainment rather than tech may limit AI’s relative effectiveness in optimizing the supply chain.

Second, the game from the famed AI match-3 studio Cosmic Lounge is in soft launch, receiving another update this week, following its late December launch and initial appearance. Pets & Puzzles looks like a standard match‑3, albeit with a super‑cute Nintendogs‑style meta. This is the first public AI gaming bet stepping up to the plate, and it looks like they’ll have to win on the supply side, as the core matching doesn’t scream anything groundbreaking. This is one to watch!

Monopoly Go (MoPoGo) spent truckloads of capital quickly, but UA is simply maxed out. There’s no better evidence than the game’s shift to celebrity-driven TV advertising following a nearly linear decline in downloads since its August 2023 highs, dropping from 13m a month to 3m. Stocked in the F2P dev/pub medicine cabinet is F2P’s natural aspirin, a prescription known for its reliability and convenience: a game extension/re-skin.

Embraced by the warm arms of Machine Zone (Game of War/Mobile Strike), King (Candy, Soda Saga), Playtika (Grand Harvest, Disney Solitaire), it’s both a cost and design play. The re-skin “sweet spot” combines what’s proven, what’s improved, and new audiences—a bastardization of Sid Meier’s Rule of Thirds.

The Game of Life (GoLife) leverages what’s proven: wealth accumulation, a coop-competitive core, live ops tied to IP, and the board as the central action resolution centre. GoLife players flick a wheel, navigate a linear board, and encounter various life events like marriage, taxes, or even the lottery, all significantly impacting their virtual net worth. The game’s collections, loans, and investments are deeper economic mechanics than the original Monopoly, while maintaining similar core motivations.

GoLife directly addresses one of MoPoGo’s shortcomings: meaningful long-run progression. MoPoGo’s progression can feel like rearranging deck chairs on the Titanic—lots of movement without advancement. By contrast, GoLife offers distinct, repeatable loops of progression through multiple LIFE playthroughs, creating compelling sim-style arcs.

The Game of Life IP doesn’t match Monopoly’s fame nor its success in Vegas Slots licensing. Yet, it resonates closely with Monopoly’s foundational fantasies of accumulating wealth with a social core. It’s transformational design target differs too: just as MoPoGo casualized social casino, GoLife can do the same for simulation. Hit sim game, Bitlife, a text-based adventure, did 130m installs despite lackluster monetization at $1 per download. Herein lies Scopley’s opportunity!

Discord represented gaming’s best chance to be recognized among tech’s elite. A recognition that increasingly means political influence rather than only the validation the games industry always seeks. Jason Citron, founder and longtime Discord CEO, is stepping down at a pivotal moment, as the company gears up for an IPO. It’s a significant blow to what ought to be gaming’s broader ambition: claiming a seat at tech’s most influential table, dominated by founder-CEO figures like Zuckerberg (Meta), Chesky (Airbnb), the Collisons (Stripe), and Ek (Spotify). 

Founder CEOs craft mystical founding narratives, something gaming needs but rarely cultivates. Tech founders have dramatizations like The Social Network, The Playlist (Spotify), or Super Pumped (Uber). Gaming had a chance to elevate its mythos with the unironically named ‘Mythic Quest’, our version of ‘Silicon Valley’, yet we failed to show up: season 4 will be its last.

Our founder-mythos roster carries caveats. Activision’s Bobby Kotick became something of gaming’s Steve Ballmer, founding the company on improving relationships between publishers and developers. Roblox’s David Baszucki pivoted from educational software to interactive physics, but games are rarely mentioned as inspiration. Roblox later rebranded “games” to “experiences,” as if games weren’t enough. Valve’s Gabe Newell, undeniably our smartest mind, remains aloof, reportedly spending his time aboard yachts in New Zealand waters rather than championing gaming’s cause.

Citron had a chance to rewrite the narrative. He founded Discord to build a MOBA, but pivoted to unify gamers fragmented by Skype, TeamSpeak, and Steam’s unreliable voice chat. Discord publicly renewed its vows to gaming as the core of its strategy as recently as 2024, a parcel of good news among gaming industry hopefuls. Discord boasts 200M MAU, with Discord Quests potentially becoming an acquisition channel marketers only fantasize about, particularly in a world where platforms are hostile to their largest customer industry.

Gaming now brushes board policy spheres: tariff, AI, crypto, advertising…but we’ve historically failed to leverage our clout. Eric Seufert wrote in 2021’s “The App Store is The Games Store”, that games accounted for 70% of all App Store revenue (this has since shrunk). Yet, gaming suffered disproportionately from Apple’s rules, countered only by Tim Sweeney’s seemingly personal vendetta. Gaming needs a broader array of influential figures to continue growing.

Everyone I’ve spoken with has praised Discord’s new CEO, Humam Sakhnini, but his backround is consulting, not a deep desire to improve the medium. Regardless, Discord remains gaming’s most crucial IPO since Roblox’s 2021 offering. It’s a chance for the market to reevaluate the growth prospects of gaming as a whole, and its S-1 is likely to highlight new political and regulatory risks.

The answer to extraction shooters lies in system design. By creating a model, we can derive implications that help guide key design decisions and frame challenges in a quantifiable way. A model also lays plain some of the variables that shape the experience of extractions. There’s a reason Marathon’s design director mentioned a target survival rate of 50% (the percent of players who successfully extract). This single variable has enormous implications for how players experience progression. Hint: it’s the casino dummy.

Similar to social casino, the genre creates “runs” where players can extract wealth or resources. Each successful run means more wealth and greater stakes. And like social casino, the genre has developed player “hedges” but turned them into progression vectors. For instance, players can ‘stash’ income or utilize ‘insurance’ slots, which enable them to protect specific resources in the event of a failed extraction. Players “buy” down cooldown time by earning cash in-game, opening the possibility of strong out-of-game economy activities.

These “persistence” elements boost survival rates through their relative effects, and setting targets against these is the most crucial decision an extraction game makes. The higher the power score ceiling, the higher the long-run win rates for a given player throughout a season. Critically, PvE elements provide predictable encounters and stable resource gains, making player survival rates tunable.

Player-to-player trade is critical. Like Zynga Poker, players almost inadvertently trade with one another when they die and drop gear. Gear degradation, such as weapons losing effectiveness over multiple runs, helps control in-game economic inflation. Escape from Tarkov implemented a similar system, and the model is further discussed in “How to Solve ‘THE’ Web3 Problem: A Staircase Tax.“

Similar to the Monopoly Go model, to understand the effects, I created a simplified version of Call of Duty’s DMZ mode. I wanted my model to:

• Simulate the first 100 player runs
  • Track net wealth and player level
  • Show how vertical progression influences the survival rate
  • Show how “insurance” and stash influence player progression
  • Show how many successful consecutive runs a player will have throughout a season or timespan
• And take less than 3 hours to build!

The result is here and available for download. Eventually, we’ll produce this:

Markov Chains & the Extractor’s Dilemma

To understand the “reset” effect, I utilize simple Markov Chains, which model transitional states. To start, I create a simple matrix and start a list of runs to model.

Here, I suggest that the chance of surviving if the player last survived is 35%, while there’s a 65% chance of dying. If the player last died, it’s far more likely (80%) that the player will die again. This is an assumption, but it simply reflects a simplifying aspect of the “resetting” aspect of extraction-economies. The massive inequality between the Survive and Death conditional probabilities could be called the Extractors’ Dilemma.

Vertical Progression

However, vertical progression improves survival odds as players complete contracts and gain stat boosts or new weapons. I create a lookup table and bound the power score gain relative to a wild stab at what FPS players expect. Survival odds of consecutive survivals increase by 15 percentage points by level 20, to 50%, and the probability of surviving after a death jumps 10 percentage points.

I calculate XP values for Survival and Death, and based on the survival rate, we determine the expected XP per run. We then use a lookup table to find the player’s level. The new player level (if applicable) will influence the survival chance, which we vary based on whether the last player’s result was death or survival at each level.

There are two additional survival odd modifiers, a slot modifier and another one for streaks. The slot bonus is an additional survival chance assumption based on the particular insured slot a player enters the match with, while the streak bonus is a cheeky modeling assumption for consecutive survivals.¹Together with level, they shape a long-run upward survival rate.

Once we have the survival rate, I simulate a single player by generating a random number, and if the survival rate exceeds it, the player survives. This is also how streaks are created: by counting consecutive wins.

Cooldown, Insurance, and Weapon Slots

The final, and most considerable part of the build, is the assumptions around cooldown, insurance, and weapon slots. Here, we assume a player unlocks an insured weapon slot at particular levels. Each slot has a specific cooldown, in minutes, during which a player cannot use the weapon; however, earning cash in a match reduces the cooldown. We also assume a player takes one weapon into the match, and the slots are ordered from best to worst (since cooldown increases for the n+1 slot, a player always wants their best weapon in slot 1). If a player does not have a weapon off cooldown, they spend cash on a weapon.²

Getting the cooldown logic right was a step-by-step process of layering on additional rules. First, there’s the rule only to use the slot if the slot ahead is on cooldown, then there’s the cash cooldown reduction, and finally, a day “reset.” This was achieved by simply setting a runs/day assumption.

Together, this finally determines the slot in which a player will enter the match equipped.

P2P, MM, and More

What’s not here is player-to-player trade or matchmaking effects – introducing separate maps, or SBMM, or an Apex-style ranked point system are worth building extensions too. At the very least, our model build helps us understand the genre and consider how design inventions affect the player experience.

Chinese developers are releasing AAA content at a historically high pace and quality, yet the outcomes remain…complicated. Genhsin, Honkai, and Zenless Zone Zero (ZZZ) are massive capital endeavors with limping revenue tails. Genshin Impact’s mobile revenue alone dropped from over $100m monthly last year to around $30m this March. Genshin’s annual budget is reported to be ~$200m on 700+ headcount, so while the game remains a profit powerhouse, its margins are significantly shrinking, a trend consistent across the entire portfolio. Honkai and ZZZ are following a similar but quicker decline from their launches, and it’s unclear if those projects remain in the red.

The launch “love bomb” blitz from marketing and content appears to converge on a game’s long-run engagement equilibrium within months, not years. Is the long-run equilibrium higher given the initial push? I’m struggling to see it. But in all the conversation around engagement, we’ve neglected monetization. Rivals will likely struggle here too, and it remains a more pressing concern than sustainable engagement. Embark’s The Finals was written off soon after its launch, as it attempted to revive part of Battlefield’s content productivity thesis: destructibility invites emergent gameplay. Despite this, it’s likely settled around 400k-500k DAU when considering console and PC. That’s a substantial player base, and certainly enough to build a sustainable business, but the familiar and broken cosmetic-only monetization model continues to hold back the expansion of HD F2P.

At 500k DAU, ARPDAU must exceed $0.40 to generate a $100m annual run rate, but this is an extreme benchmark for 99% of cosmetic economies, where $0.15 is considered a “great” outcome. At its peak, Fortnite reached ~$0.40 ARPDAU, where even mobile titles like Royal Match reach 20-30 cents ARPDAU at scale. The answer from most titles and developers is to go premium, and it’s the right one. However, if the HD F2P model is to expand, it must support lower DAUs, and that means new models that can reach mobile-esque monetization levels.

In the talk, @Teut Weidemann documents the rhythms of F2P engagement and monetization curves like an archaeologist uncovering a secret code. He illustrates what happens when you zoom in. Like, really in. For instance, concurrent users during football matches (the boring European kind) dip during game time and spike during halftime. Yes, that creates a butt shape (wait until you see his next shape). This kind of minute observation makes Ultra Local Live Ops™ compelling, finally providing a meaningful avenue for the “personalization” agenda.

“Personalization” has become an ugly stepchild to price discrimination, mainly focused on triggering successive live ops offers. CRM efforts might localize notifications or reference local holidays, but rarely capture the immediacy of “what’s happening now” or what’s physically around players. Although some AI companies are undoubtedly tackling this problem, I know game design can do something more straightforward and immediately feasible to build Ultra Local Live Ops™.

There’s a second, broader point when zooming out. Games like Path of Exile (PoE) have an exotic PSU curve. While most seasonal HD games display the dreaded stegosaurus pattern (credit @Laura Taranto), PoE’s is greatly exaggerated, resembling a gacha F2P game rather than a dungeon crawler. The reason is straightforward: resets provide entry points for new players. This approach mirrors the anthology revival in linear media (White Lotus, True Detective), similarly driven by audience retention economics. Vertical progression or serialized storytelling naturally disperses players across a broad spectrum of power scores or episodes.

Games like FIFA Online and Tarkov likely experience similar season-start bumps, but it’s curious that this design hasn’t broadly expanded to other titles, particularly on mobile where power scores are common. While 4x players are spinning up new accounts on new servers, it differs significantly from fully resetting collections, as seen with Genshin Impact.

Dream Games is opening the UA floodgates for Royal Kingdom, deploying significant marketing spend and celebrity-driven creative. This move reflects a familiar pattern: Royal Match itself required about two years to scale effectively. But now, the Dream Games UA team faces a new paradigm: how to optimally scale spending across two titles and manage new franchise-level LTV dynamics.

In the short run, allocation optimization is obvious as UA managers frequently “rotate” spend across channels within a day as CPIs spike with additional channel capital. The marginal game should function as a new channel of sorts to balance spend against ROAS. However, franchises also bring new modeling considerations.
The expected value of players engaging with both games simultaneously can be modeled explicitly as:

Here, the probability of simultaneous engagement represents the remaining LTV in the existing game for the ith player, and is the new game’s LTV for the same player. To justify dual-game investment, this combined LTV must surpass the potential LTV had the player remained solely in the original game. It is possible to decrease portfolio LTV by shifting players to a lower LTV game (my poor Squadbusters)!

Maximizing the probability of a combined LTV is a function that considers the genre margin or how similar two franchise games are. The closer they are, the more likely players are to play more than one simultaneously. While similar genres maximize the DualPlay probability, it comes at the seeming cost of portfolio diversification, which introduces more risk.

Companies are hedged only when selecting uncorrelated assets, and game genre fits the bill along some dimension. The relative strength of casual and the 4x genre reinvigoration in the post-ATT world is a good reminder of this. The reality, however, is that the standard deviation on return to a new genre far outweighs relatively uncorrelated genre portfolio assets. Dream Games building a 4x title is riskier than another match title.

A clear trade-off exists between maximizing genre margin to avoid cannibalization, maximizing DualPlay LTV, and managing variance to ensure stable returns.

Seeds are one of the most powerful variables in game design, and one of the least acknowledged. EA’s Data Science team published a paper showing a 5x difference in win rate—from 15% to 75%—based only on the seed used to initialize a board. The level design and other match parameters remained the same!

Computers can’t generate genuine randomness, so they “anchor” on seeds: a fixed number generates a pseudo-random sequence. The initial board state is the same if the seed stays the same. Without fixed seeds, every level must be balanced across the entire range of possible outcomes.

When match designers look at attempts per success, seeds hide so much of the variability in experience, creating uneven ground to balance against. Battle Royales do this too, wherein loot tables are often seeded, either consciously selected or not.

The hierarchical rules of seeds unlock incredible experimentation opportunities. What if the seed changes based on the player’s win streak? What if an easier seed is selected on a player’s (x+1) attempt? What if the seed changes after a continue purchase? What if the engine automatically determines the highest retention seed? Some of the most fun we’ve had at The Experimentation Group is answering these questions.

Seeds demand analytical relevance more than anything. I can count on one hand the number of times I’ve seen seeds on a dashboard, either on mobile or HD. Overlooking seeds leaves optimization gains untapped!

Full paper below or here:

Extraction shooters were supposed to be the next big leap, and Escape from Tarkov was the movement’s exemplar. Titles like Jager’s The Cycle: Frontier and Call of Duty’s DMZ failed to generate meaningful traction. Many mistakes stem from a misunderstanding of Royale’s place in shooter evolution, creating an inability to conceptualize what’s next. Bungie’s Marathon is the genre’s last great hope.

Battle Royale radically transformed the FPS shooter landscape. Since PUBG’s 2019 launch, Battle Royale formats went from 0% to nearly 50% of the shooter market, capturing 1/5 of all game playtime. Nearly every AAA shooter franchise was forced to bow: Call of Duty, Battlefield, CS:GO, and GTA all experimented with Royale formats. BR was a revolution that every major franchise felt, and yet, since Warzone in 2020, the subgenre has been in innovation limbo. Studios burned VC funding at scale, but genre-shifting innovation hasn’t followed. Meanwhile, a separate Eastern crop of mobile Royales has propagated, with its most notable feature being traditional Eastern live ops cadence.

Extraction shooters felt like the next logical evolution, a subgenre advancement. Their core thesis appeared tied to the return to the stakes Royale first introduced, but was watered down with more frequent respawns in each iteration. However, this thesis misunderstands shooter history.

Extraction shooters represent DayZ’s true lineage, while Battle Royale represents a branching alternative. Escape from Tarkov originated in 2016, three years after DayZ’s release, and was developed nearly simultaneously with H1Z1, DayZ’s battle royale spin-off mod. The original DayZ contained elements familiar to players of survival games like Ark and now even Palworld: hunting, crafting, and base building. Extraction is the shooter’s interpretation of these elements.

The Battle Royale genre succeeded by embracing roguelike in-round progression: scavenge randomized loot, grow stronger during the match, and fresh strategic variance every round. Brendan “PlayerUnknown” Greene’s original design was on this basis: randomized weapon spawns create different conflict spots across the map, and fresh strategy mastery. Successive attempts have embraced collection but reduced stakes, but there’s been little innovation since Call of Duty’s Warzone. Publishers seemed to believe they had picked all the low-hanging Royale fruit.

Instead, AAA shooters tried incorporating Extraction, but Battlefield and Call of Duty whiffed. Ironically, My.Games’ Hawked came closest to a design breakthrough: simplified objectives and clear win/loss states. The ceiling for extraction will remain low unless someone convincingly rewrites the meta.

So where does that leave Marathon? With a bold art style, the ambition is there. However, Marathon needs to move past Tarkov’s influence and reimagine what the next battle royale looks like, “extraction” influence or not.

Recurring events are a key source of how the game creates “runs” by stringing together level climbing in different, interconnected progression centres. While Monopoly Go certainly wasn’t the first to invent “move around board, trigger events”, it’s undoubtedly popularized it. It’s starting to appear as its own genre, and even morphing into Archero 2 mini-games.

The events “tickle” the mind by creating cascading effects where multiple reward centres trigger, in some time interconnected ways. The economy is of particular importance in casino games, since the main currency functions as energy – without it, player engagement stops, but make it too plentiful and purchases also stop. With more games implementing these loops, I want to setup a basic model to understand how they work. Games in prototyping should be creating economic models to guide design making. The process of making the model is as valuable as the model itself. Creating time-boxed rough model sketches helps vet key design decisions.

I wanted my model to:

• Simulate the first Player 100 rolls
  • Track Dice Balance, and Net Activity Per Move
• Create two “reward” centres to reward additional dice beyond board rewards
  • Players earn progress toward milestones in each centre
  • Monopoly Go features a two main and reoccuring collection events.
    • Beyond login rewards, sticker collections, board completion, Chance cards (that’s a lot!).
• Regenerate rolls on a schedule.
• Take less than 3 hours to build.

The result is here, and available for download. Eventually, we’ll produce this:

Process

I started with the Archero 2 minigame template, with some small changes, to build the board, and first affirmed the expected dice roll reward per roll was less than one (i.e., the player won’t play for infinity).

Next, I want to spin up the player moves. I created 100 and then generated a random roll, incrementing from the player’s last movement of the board. Next, the resulting tile is wired to the tile reward: these trigger the reward centres. I’ll revisit the multiplier later.

There are two: Otta Treasure Level and “Cash.” Both reward Dice rolls for collecting a certain amount of Treasure or Cash. We create look-up tables for all these values, putting in random number to play with and scale. The point of the timebox is to keep moving; good economy designers don’t get weighed down by certain assumptions, and polish later.

Rewards Look Up

The next part of my build, and the messiest part, is the rewards look-up after each roll. We initialize the player at level 0, and if the player earned Treasure or Dice, we calculate the levels they moved up and give the player the corresponding reward.

The rewards from the reward centres and the rewards from the simple board tile are added to the player’s wallet.

Multipler

Now that we know wallets, we can build out the multiplier function. I assume:

• The player always bets up until the allowed maximum
• The maximum is a ratio between the multiplier and the player’s wallet

We can use this rule set to calculate the player maximum multiplier that they could play, and assume the player uses it. The ratio we set significantly impacts the number of rolls a player needs to play! Here, the player had 50+ rolls, but had to churn through them at a slow pace, and the 100x multiplier significantly bottlenecked pace too.

The next feature is Dice regeneration. First, we create a ruleset that triggers a regen period if the player’s balance reaches one dice. The period refills them to their wallet maximum, and we consider that a “session” (the player only returns once their wallet maximum is regenerated).

We’ll add a bunch of other derivative metrics that the model sets up, too. Since we have regen time, we can apply a “slack” rate or the percentage of missed Dice rolls players fail to expend. For example, if Dice regenerate at 100 per hour, a player could log in 24 times a day, perfectly one hour apart, to maximize the reward. The slack rate discounts this amount and implies the number of sessions a player will have per day.

And since we know a regen event triggers a new session, we give every roll a cohort day. Our model predicts the roll will complete the 100th roll by day 2.08, with an RTP or total return to player of 90% (the amount of dice sourced versus sink).

Result

Our net activity chart is the most interesting; we can see that we triggered some of the early reward centres, and two events were strung together. We also see that consistent waves without the reward centres are boring and predictable. The “Go” effect is about setting up these triggers in interesting ways to connect and cascade.

I wouldn’t use Google Sheets for this project in the long run. We only set up a single-player simulation, but Sheets makes it annoying to deal with randomness beyond this (and don’t get me started on seeds). The goal was to spin up a model of understanding, and honestly, there’s enough here to run Archero 2’s mini-game. In the long-run, I would convert this to a Python or R project, and it’s needed at most level of social casino.

Teams should create these models, even for real-time match pacing or combat, and certainly early in production. Modeling building is an exercise in playtesting the design!

While game director and sim genre maniac Director Hyungjun “Kjun” Kim seems to have delivered on the game at this stage in development, Krafton needs to step up to deliver on the title’s business promise. Krafton needs to flex its newly formed publishing arm quickly, or a franchise could slip away.

InZoi’s initial arrival has been good but not great so far. Initial estimates point to Steam revenue in the range of $35+ million, with Krafton stating it sold than one million units. However, retention is slipping, and active users are converging on The Sims’ Steam figures. So far, there’s been zero Sims PSU impact, unlike something like the meator Marvel Rivals left on Overwatch 2’s numbers.

However, InZoi successfully introduces long-requested Sims features like drivable cars and third-person character control. While players give it credit for this, the game remains plagued with bugs. It’s an “Early Access” title, but that’s an increasingly meaningless term. Early Access is eligible for nearly every Steam benefit, and subsequent public rollouts rarely exceed Early Access launch figures. It’s another challenge for publishers to properly calibrate, as the Early Access launch “sucks the air out of the room.” Who, for instance, seriously thinks Palworld’s global launch will exceed its Early Access sales figures? 

The game’s distinctly Korean localization is strange, considering South Korea comprises 3% of the player base. For instance,  one playable city is modeled after Seoul, yet merely 3% of players originate from Korea. Accounting for regional differences is part of player empathy, just as understanding their technical and platform choices. The Sims is a unique audience, and EA has built special marketing pipelines and relationships with these players. Overwhelmingly female, the entire franchise rakes in an obscene amount of revenue, over $3-400m+ per year on a DLC model. However, InZoi misses the fact that these players have lower-end machines, making the recommended system specifications (a GTX 3070 graphics card) inaccessible. Similarly, the lack of a Mac version doesn’t help.

Still, outlook remains cautiously optimistic. A detailed roadmap through 2025 includes a “Cat-city” update, and coop remains strangely missing in the genre. With Paradox cancelling its earlier life sim, Life by You, the runway is clear for InZoi to take a bigger bite of a genre previously comprised of one title.

Phillip & Eric navigate the strangely subdued landscape of GDC 2025, pondering if there really is such a thing as a free lunch. Christopher Kaczmarczyk-Smith dials in, wondering if his absence is secretly the key to Eric’s roundtable success.

They dissect the talks, the conference economics, the rise of mobile’s respectability, and a guest in economy designer, Charlie Hsu.

In this episode:

– Is the game industry actually shrinking, or just taking a nap? And if Web3 isn’t the savior, what’s left besides… sweeps?
– Is GDC just a cleverly disguised wealth transfer from sponsors to developers?
– What’s the latest “reasonable” pitch for Web3 in games?
– What’s the economic model behind those San Francisco walk-up shops overflowing with candy bars right next to the register? High margins? A tourist trap? Something… else

Watch: https://lnkd.in/e66H5DR6
Listen: https://lnkd.in/ekQUjrKX

Another wave of AI advances, another chorus of artists crying foul. Each wave has been unoccupied by serious argumentation; pro-Copyrightists need to advance something beyond an assumed conclusion if they want to win minds, not just hearts. Copyrightists routinely declare AI-generated art and “training” as copyright infringement and insist artists are owed royalties. This presumption is anything but obvious.

 If every artist owes royalties to those who came before, should every “pop” artist who creates in the style of Eduardo Paolozzi owe his estate royalties? Film director Tarantino routinely mirrors Spaghetti-western shots; his royalty bill must be absurd. Art and science have always evolved by imitation and influence; the very idea of copyright is a recent invention. No one knows this better than games!

The universal outcry against WB Games’ patenting of the Shadow of War’s Nemesis system, widely condemned for stifling design creativity, starkly contrasts enthusiastic support for generative AI art royalties. Mobile gaming, in particular, is an industry built on iterative copying and rapid innovation. (How exactly again would these AI royalties  function?)

Mobile gaming, infamous for clones and slight variations, thrives precisely because imitation accelerates innovation; it’s one of the great examples of a near-perfect competition economic model. All this cloning and imitation haven’t dampened creativity or investment; instead, they’ve grown it. In fact, mobile gaming is the strongest case against state-enforced copyright altogether. 

One of AI’s biggest mistakes was painting a target on its back. Calling it “artificial intelligence” misrepresents the human ingenuity embedded in it; it’s AI is anything but artificial. Instead, it’s the best of humanity’s startling achievements. It learns from all great masters across history, aggregating their brilliance into something, and yes, distinctly human. AI actually stands for “aggregated intelligence”. Claims that AI lacks a “soul” collapse when blind art submission contests consistently rank AI-generated works at the top. New art tools like the musical synthesizer or Photoshop have always faced moral lobbying before deep acceptance and eventual adoption.

Is there room for royalties or specific protections? Perhaps. But this case must be articulated, not assumed.

Netflix Games finally has a coherent strategy—hallelujah! Yet, despite this newfound clarity, it’s missing perhaps the most obvious acquisition opportunity since Sony scooped up Insomniac Games: Jackbox Games.

Netflix pushes games to subscribers via a dedicated row within its mobile app, directing users to separate App Store pages for downloads. The strategy was flawed from the start, as 70%+ of Netflix consumption happens on the TV. The Games group was desperate for awareness, so they turned to physical billboards that read: “Wait, Netflix has Games?” This is a modest improvement, but it still fails to meet subscribers where they consume content—the TV.

While Netflix did have an AAA FPS studio, it was dubious that it could ever solve for distribution or, more importantly, input. Netflix’s sprawling ecosystem of devices and lack of a dedicated controller made it suspect from day 1—unless, perhaps, another ill-fated game streaming service was being planned (RIP OnLive, Stadia, GameFly, and soon-to-be xCloud).

Enter Jackbox Games: a family-friendly (depending on who you play with), party game experience with minimal hardware requirements, scaling across multiple platforms on a single code base, and using the phone as the primary input. It’s difficult to overstate how perfectly Jackbox aligns with Netflix Games’ renewed strategy and its platform challenges. It could seamlessly scale across Netflix’s existing platforms, leveraging technology from Netflix’s shuttered interactive TV division. It’s easy to imagine individual Jackbox games sitting in its own content row on Netflix’s TV app. Beyond technical feasibility, integrating Jackbox would help Netflix revive a broader social vision.

While game makers tout social as driving long-run retention, Netflix has moved away from social. Friend’s list, queues, and group watch have been dead for nearly a decade; I suspect the killers are tech PMs who A/B tested features into oblivion. However, the underlying issue is a lack of visionary ambition, not social as a concept. Netflix Games could—and should—lead the charge in crafting a richer, more socially interactive entertainment platform, especially with Disney+ nipping at Netflix’s heels.

Netflix Games is brimming with exceptional talent; however, from what I understand, corporate politics has hampered the division’s ability to form its own identity. Netflix leadership seems determined to follow in the footsteps of Google, Facebook, and Apple in failing to respect the medium of games. While Amazon Games isn’t a smash success, its relative success over the last five years is a function of Amazon significantly loosening the cultural reins and creating an internal corporate annex: games are to be run by game people. 

Instead of fostering innovation, Netflix Games opted for the safe, uninspiring route of accumulating vast content libraries. This strategy fundamentally misunderstands how players engage with games. It’s a strategy consistent with the passive content consumption model Netflix was built on—not the interactive, social-driven medium games inherently are. For Netflix Games to thrive, leadership needs to embrace gaming’s unique dynamics rather than treating it as merely another category in its streaming catalog.

1. We have to want to live

This revelation insight from former Machine Zone CEO Gabe Leydon spun my head throughout the conference. He's right: our passion isn't dead, but it's undoubtedly been crowded out. Mobile obsesses over acquisition, while HD fixates on fundraising. New platform dreams--web3, AR/VR, HTML5, UGC--have largely fizzled. Yet these platforms persist, surviving but not thriving. Mock AR/VR as the "Dippin' Dots of Games" if you like, but at least those developers burn with genuine passion.

The rest of the industry must reignite the will to survive--the kind we see from firms like Supercell, Lessmore, and Hypehype.

2. AI start supply-side; innovation to follow

Michail Katkoff made an excellent observation about the absence of the "world-changing" bravado heard from SF founders intoxicated by Paul Graham essays. Instead, the talks were pragmatic--topics like "'Angry Birds' and AI in Practice: Finding Our Own Way." Games' AI progress has been slower and more modest than I expected, but the supply-side evolution is well underway.

The indie game section demonstrated this shift, overflowing with polished 3D worlds instead of the usual charming 2D sprites.

There's no shortage of seed AI ventures either, restoring my confidence that we'll soon master this tool to advance the game-making craft. Realistically, we're ~2-3 years from seeing a major mainstream AI-driven hit.

3. Mobile has finally earn industry peer respect

It's about time. While a decade late, HD developers and indies no longer dismiss "mobile" as trivial. I counted over 30 mobile-focused talks! Seeing companies like Scopely actively participate was an encouraging sign that mobile also wants to be a peer of the game industry, not the tech industry.

GDC is a conference of peers, and it feels like that at each talk. It talks about actual developers sharing lessons and perspectives that advance the craft of game-making. Talks like "Tabletop Summit: Timeless Design Lessons from 25 Years of 'Duel Masters' TCG".

While many opt for the Expo-only pass (myself included in past years), I intend to get the full pass annually. I read a lot of copium from non-attendees, and they're wrong; if you're a developer, this is the place to be every year--if you're serious about advancing the game craft, GDC is essential.

4. Sweeps, sweeps baby

One of the few EXPLOSIVE growth sections has been in sweeps, or games where players earn virtual currency they can redeem for real-world prizes. While this is a regulatory timebomb, the results so far are breathtaking. I understand the muted enthusiasm from the developers, but the reality is that we need to approach everything with an open mind, including one of the new growth sectors. Like hypercasual, sweeps may look messy now, but their innovation inevitably trickles down.

If you’re European, you need the U.S. economy to grow. If you’re Canadian, you need the U.S. economy to grow. This sobering reality underscores nearly every Western mobile development hub today: for mobile gaming to grow, the U.S. needs to grow. Although optimistic narratives from developing markets like India and Brazil are abundant, their growth rates still pale compared to the U.S.’s real spending power growth. For the foreseeable future, it’s U.S. or bust in mobile gaming, a reality suggesting mobile gaming doesn’t control its destiny.

Political literature in the early 2000s enthusiastically promoted the BRIC nations (Brazil, Russia, India, and China) as future economic powerhouses. Since then, however, these countries have largely underperformed expectations. Rather than converging with the U.S., the gap has widened, placing America firmly at the center of mobile gaming’s economic stability.

In 2024, the United States accounted for a plurality of global mobile gaming spend, increasing its share by a significant six percentage points to ~35% compared to 2023. While countries like India might boast higher real GDP growth rates—6.5% in 2024—their absolute spending gains remain modest. For instance, India’s per capita purchasing power year-over-year increase was just $400-600, compared to the U.S., adding double that in real purchasing power with less than half the growth rate. While low-income countries have faster growth rates, rich countries have larger real and absolute growth.

Another disparity arises from mobile gaming’s classification as a luxury good (?!): individuals increasingly allocate marginal dollars towards gaming as incomes rise. India’s explosive growth in smartphone penetration—from 30% to 70% in under a decade—hasn’t translated to significant IAP (ads revenue, correlated yet distinct, paints a slightly different picture). Brazil has managed somewhat better, though its monthly IAP spend of $30M remains negligible compared to even markets like Taiwan, which average $150M monthly IAP despite Brazil having 10x the population.

As global population growth slows and smartphone penetration reaches saturation, income growth emerges as the only reliable macroeconomic driver for mobile gaming expansion. Consequently, the United States remains the indispensable anchor for sustained mobile gaming growth until other economies manage a meaningful catch-up.

Tripledot’s filings reveal an astounding 98% ad revenue share, which brought in over $300M last year. Yet margins languished around a frustrating 15%, underscoring why the rush toward IAP and hybridcasual feels more like a strategic necessity—aimed at gaining leverage over spiraling acquisition costs and boosting future valuations. Again, gaming economics resist easy analogies and maintain diverging economics from traditional software or subscription models like Netflix (35% margins, pricing power, fixed costs!). 

Eric Seufert’s recent “Economics of Mobile Gaming” underscores how essential ads remain for value distribution within the mobile gaming ecosystem. Titles employing a “fakeout” strategy, such as Last War, Frozen City, and Top Girl, edge toward vertical integration, directly integrating the value routing system into the game. It’s a tangible opportunity for Tripledot, with surprisingly few Western attempts. Where are the Star Trek Fleet Command fakeout funnels? Machine Zone, with its expensive-to-develop 4x engine, offers intriguing potential if Tripledot can duct tape something hypercasual on top.

Applovin’s robust social casino segment (25-30% revenue) presents fewer exciting opportunities. Project Makeover, contributing a substantial 15-20% to Applovin’s studio revenue from IAP, has already leveraged aggressive ad creative. While Applovin’s genre diversity may appear hedged, it’s more of a management headache, as its genres maintain domain-specific knowledge. Other suitors like Scopely likely passed on acquiring these studios due to similar management complexities, a factor evident in their selling price – a sub 1x premium over the studio’s annual revenue.

Typically, acquirers streamline headcount aggressively; however, in this scenario, cuts risk losing critical knowledge in nuanced genres like 4x, social casino, and survival. As Zynga demonstrates, structured performance payouts or golden handcuff arrangements are prudent for revenue stability.

With $300M in cash assets, Tripledot’s leveraged acquisition strategy involves assimilating studios with higher revenues and larger headcounts alongside significantly different business models. Despite impressive revenue, Tripledot’s growth trajectory showed signs of plateauing, and looming EU regulatory uncertainties complicate prospects for its ad-centric revenue model. Digesting this complex asset carries significant risk—failure could prove catastrophic—yet avoiding such aggressive moves risks an equally painful hyper-to-hybrid transition.

Nearly all economic activity in match-based or “saga”-based metas follows the fundamental relationship:

Where:

• [latex]\text{Revenue}_{i}[/latex] is the revenue for the [latex]i[/latex]th player.
• [latex]\text{Attempts}_{i}[/latex] is the number of attempts made by the [latex]i[/latex]th player.
• [latex]\text{Gold Sink Per Attempt}_{i}[/latex] is the average gold spent per attempt by the [latex]i[/latex]th player.

Holding [latex]\text{Gold Sink Per Attempt}[/latex] constant, an increase in [latex]\text{Attempts}[/latex] leads to higher [latex]\text{Revenue}[/latex]. Similarly, increasing [latex]\text{Gold Sink Per Attempt}[/latex] while holding [latex]\text{Attempts}[/latex] constant also raises [latex]\text{Revenue}[/latex]. Many innovations in match economies operate within this framework. For instance, a “super light ball” mechanic increases [latex]\text{Attempts}[/latex] but reduces [latex]\text{Gold Sink Per Attempt}[/latex]. Events follow the same structure.

Basic Match Model

At the aggregate level, revenue across all players is:

Event Uplift

Events affect revenue by shifting the number of attempts or the gold sink per attempt. The marginal uplift from an event [latex]e[/latex] is given by:

Where:

• [latex]\Delta \text{Attempts}_{e}[/latex] is the increase in attempts attributable to event [latex]e[/latex].
• [latex]\Delta \text{Gold Sink Per Attempt}_{e}[/latex] is the increase in gold spent per attempt attributable to event [latex]e[/latex].

Effective Rewards

Events also distribute rewards, which function as a gold source in a simple gold-only economy. The effective rewards generated by event [latex]e[/latex] are:

Where:

• [latex]\text{Effective Rewards}_{e}[/latex] is the effective gold rewarded during event [latex]e[/latex].
• [latex]s[/latex] indexes different reward tracks or stages within the event.
• [latex]\text{Reward Amount}_{s,e}[/latex] is the gold granted at stage [latex]s[/latex] of event [latex]e[/latex].
• [latex]\text{Win Rate}_{s,e}[/latex] is the probability that a player reaches stage [latex]s[/latex] and claims the reward in event [latex]e[/latex].

Games generally find that engagement is strongest when effective rewards are high. Tracking either progression speed (win rate) or reward amounts amplifies any event’s effect.

Event Effect: Net Source or Sink

The net economic effect of an event is determined by subtracting the total effective rewards from the total event uplift:

If [latex]\text{Event Effect}{e} > 0[/latex], the event acts as a net sink, removing more gold from the economy than it introduces. Conversely, if [latex]\text{Event Effect}{e} < 0[/latex] the event acts as a net source, introducing more gold than it removes.

Not all events need to be net sinks. Events can function as planned gold sources, offsetting reductions in base saga rewards to maintain long-run economic stability. In this scenario, base saga victory rewards are reduced while event rewards compensate, ensuring that the total gold sourced per attempt remains stable but with different sources.

How Many Events to Run

For a series of events that act as net sinks, running multiple events concurrently increases the total gold sink. However, each additional event faces diminishing returns in increasing either [latex]\text{Attempts}[/latex] or [latex]\text{Gold Sink Per Attempt}[/latex], while [latex]\text{Effective Rewards}[/latex] remains relatively stable (win rate generally increases as rewards rise, though the rate of increase declines with each additional event).

Where:

• [latex]\text{Base Uplift}[/latex] is the uplift if the event ran alone.
• [latex]\text{Effective Rewards}_{e}[/latex] is the gold distributed by event [latex]e[/latex].
• [latex]\gamma[/latex] controls how quickly subsequent events’ net effect diminishes.
• As [latex]e[/latex] increases, [latex]\gamma^{e-1}[/latex] becomes smaller, reflecting diminishing returns for each additional event.

The optimal number of concurrent events occurs when:

Beyond this point, an event is a net source, which is not sustainable in the long run without other economic adjustments. This should be how teams assess event stacking—set up A/B tests to find the zero point; by all means, Royale Match suggests five to six events.

Conclusion

Match events should be rigorously modeled to understand their economic effects. Contrary to common belief, an event’s effectiveness is not solely determined by its ability to act as a net sink. Structuring events as alternative sources of rewards can be LTV-positive compared to embedding the same rewards in saga progression.

Balancing reward scalars—designing an event with the same Event Effect but significantly higher uplift and effective rewards—creates different economic trade-offs. Remember, these two are causally linked: offering more rewards increases attempts. Consider experimenting with holding effective rewards constant, increasing the win rate but lowering the prize.

Game economy designers have inadvertently revived the economic traditions of famed economist Irving Fisher. In Fisher’s world, the economy is a hydraulic machine, with refinement or “productive process” occurring as resources flow. The impact of major institutions takes on a literal effect, where a tariff might slow the flow of liquid or capital from one chamber or country to another. It’s a surreal approach and the early attempt at modeling game economies.

Irving Fisher studied monetary theory and price levels at Yale, where his obsession with equilibrium led to an extraordinary invention. In 1891, he unveiled his hydraulic-mechanical computer—a contraption of water, pulleys, and floats that physically demonstrated how prices find their level. The machine, described in his paper “Mathematical Investigations in the Theory of Value and Prices,” became the first working economic model, translating abstract theory into observable mechanics. It was part of a broader effort to formalize economic theory—something game economy design desperately needs.

Tools like Machination and Excel carry Fisher’s torch, albeit in digital form. Machination’s node-based diagrams mirror Fisher’s hydraulic logic: resources pool, flow, and transform through gates and converters. The Excel models do the same, but the piping and transformation process is more opaque (another reason it fails as a tool).

The economy designer is a new role that’s emerged only in the wake of live services played over the course of years. Live services, or long-time horizons, make virtual currency viable for lowering transaction costs (hundreds of IAP transactions mean credit card fees and awkward UX) and crafting complex time/money supply chains. This highlights currencies’ key roles as stores of value and units of accounting.

The economic model becomes necessary to understand these time and money supply chains, much as Fisher’s hydraulic computer was necessary to understand price equilibrium. If the only way to understand a supply chain is to solve something down it, then the economy model’s goal is to do just that. The result is an ability to understand unintended consequences between overlapping progression institutions and design the shape of key game KPIs.

The irony of modern game economies lies in their reliance on century-old methodologies. While Fisher’s hydraulic principles find new relevance in persistent digital economies, our tools remain trapped in the past. The field desperately needs reproducible, accessible modeling systems that simplify—rather than replicate—the complexity of modern virtual economies. Reproducibility and accessibility remain elusive, continuing to be the Achilles’ heel of game economy design.

In the 26 months since ChatGPT launched, Web3 has made more tangible progress than AI gaming applications. Any technology outpaced by Web3’s progress is in danger of irrelevance. And it’s for good reason: so far, AI’s visible footprint in games has been limited to a series of Cosmic Lounge press releases (I want my match-3 levels!) and a sixth finger in Call of Duty key art. Everything else appears to be cheap VC slideware or corporate investor day fantasies.

It’s true that gaming effectively subsidized the AI revolution by bankrolling Nvidia’s GPU development, ultimately creating a problem set whose solution unlocked accelerated AI growth. However, it’s failed to capture the subsequent gains, as code development is on a tear; real-time voice translations are consumer-grade, and medical diagnosing is advancing. Gaming has pioneered advancements in real-time 3D graphics, physics simulation, haptic feedback, and internet infrastructure expansion via demand acceleration and has taken wrong turns, too. Cloud gaming, motion controls, plastic guitars, and subscriptions don’t look like they work. Web3 looks like it will register as our next wrong turn, not because it’ll die, but because it’ll fail to grow.

When Do You Write Off the Forte Money?

On the demand side, most game applications are demos, while conversing with NPCs—the most rudimentary possible use case—is lauded as a compelling feature. For an industry that prides itself on the fusion of art and science, we must reclaim our heritage of pushing technological boundaries rather than simply implementing API wrappers.

On the supply side, the pursuit of optimized match-3 gameplay illustrates our current limitations. While we understand that level design significantly impacts retention, determining optimal sequencing amid countless variables remains a formidable challenge. AI promises to accelerate the supply chain, but the bottleneck has shifted downstream to fundamental design questions.

Web3 represents a fascinating detour that consumed substantial venture capital, yielding little except the employment of millions of Southeast Asian players who beat local wages through trading virtual value (true!). Whether the answer lies in Roblox text prompts or voice-command FPS demos remains unclear, but there is still room for a team to experiment. Gaming needs to return to Hoffman’s philosophy: “If you’re not embarrassed by your first release, you shipped too late,” and try getting embarrassed ASAP.

Matthew Ball’s 2024 gaming review paints an industry that had a good ride but is settling into stagflation. The Metaverse is nowhere to be found; instead, growth drivers like cloud streaming and GTA pricing replace it. If Stadia is the best we can do, we’re truly fucked. I wrote in agreement earlier this year with “Gaming’s Best Years Are Behind It.” I stand by the same pessimism, but there’s an essential factor that may well guarantee gaming’s growth: time.

Over the past decade, gaming’s largest achievement has been its pliability: it can adapt to any format—smart TV, smartwatch, smartphone, and increasingly health and financial apps. It’s not just a form factor; the business models bend too. F2P didn’t just take share from traditional box revenue; it expanded the audience of gamers. Now even Grandma plays Candy Crush. Without real momentum behind a new control factor (AR/XR), and with smartphone, console, and PC growth limits reached, the audience appears maximized. Surely, revenue will follow at some point.

The metrics demonstrated by early “gaming cohorts” suggest realtized growth that’s yet to revibrate among the DAU base. If we consider gaming itself a product, we observe that new cohorts are far more engaged age-over-age than in earlier generations. As a reminder, Roblox maintains 380 MAU, with about half the kids in the United States as active users. Roblox’s users are aging, and even if they churn from the platform, the more important question is whether they retain gaming as a whole, with spending power growing alongside income.

As the current Roblox generation ages and habits are maintained, new, more engaged cohorts will stack, changing the relative share of gaming engagement. Of course, monetization must follow as Roblox struggles to grow user spend. So far, they haven’t released monetization metrics as cohorts age.

The best source for time trends comes from the American Time Use Survey, which gives diaries to Americans and has coders classify entries. Among 15-24 year-olds, minutes played per day is over two hours, up from 1hr 40min in 2023. This minimizes the Roblox effect, as those cohorts just reach maturity enough to be counted 15 and above in the survey.

Not every cohort is up, however. Despite clear jumps in engagement, the new millennials (25-34) are not trending upward. In 2003, these cohorts would have been born in 1978 (Space Invaders), while the most recent age participants would have been born in 1998 (considered the greatest year in gaming of all time). It makes sense, then, that the 25-34 year-olds are poised next to see these most engaged cohorts enter the pool.

Meanwhile, the rest of the age groups are holding onto COVID highs but show no obvious long-term growth.

It’s important to remember that the coders for the time use survey have to follow strict rules on categorizing entries. The specific games category includes the computer, so it’s unclear if mobile games would fall in the same code:

32. Playing games on the computer (including Internet games): Code as
    Socializing, Relaxing, and Leisure/Relaxing and Leisure/Playing games
    (120307).

Still, it’s puzzling how stable the trends are for the rest of the age cohorts. Are 75 year olds really playing an hour of games per day?

At least multiple reports have picked up on the rise in young men’s gaming habits, enough to cause controversy as a scapegoat for lower labor force participation. This has doubles down on a shift in gaming comprising a larger share of total leisure time. In other words, as consumers exhibit more leasure time, we find that they spend it increasingly on gaming.

The holy grail is more longitudinal data of consumer preferences over time. Fortnite is celebrated as a graduation game from Roblox, but it’s unclear if that’s reality or if engagement instead withers away in place of other obligations. For gaming to remain relevant, it needs to continue to bend across age groups, with the NY Times Wordle as relevant as Star Stable in monetizing over the lifetime of a customer.

The warning signs are already apparent. Mountaintop Studios‘ venture-backed Spectre Divide has become a cautionary tale, with peak simultaneous users (PSU) dropping below 100. Supervive’s trajectory is similar, maintaining 4,000 PSU despite years of development, encouraging playtests, and the backing of the former League of Legends leadership team. Its salvation may lie in the East, given that 80-90% of Riot’s revenue flows from Eastern regions.

While the linear declines in Rivals and Path of Exile 2 (PoE) triggered alarm bells, the unfolding reality is more nuanced. Yes, engagement has declined, but from historic peaks. Both titles maintain a top-five Steam ranking and are inching toward PSU stability.

Video Game Insights‘ new daily active users metric highlights a fascinating disparity: despite nearly identical PSU curves, the gap in DAU between the two is stark. This isn’t a methodological quirk but reflects Rivals’ ripping through new players faster than a teenager through a PS5 on Christmas. PoE’s paid early access limits the casual “dating” that Rivals has seen. The upcoming Circana console rankings will prove particularly telling for Rivals and Eric Kress‘ grim pronouncement for the title.

Delta Force emerges as the dark horse, potentially overtaking PoE and Rivals through sheer stability. Its rock-solid stability is so stable it raises eyebrows; on the other hand, Steam’s military shooter audience—a unique demographic dominated by German detail-obsessed gamers who debate the historical accuracy of M1 Garand iron sights. Factor in China’s staggering 40 million pre-registrations across mobile and PC, the impending mobile launch, and Delta Force will have legs, even if the Western one is shorter.

For all the Q5 noise, Q1’s Palworld’s recent engagement bump reminds us to reexamine live-service go-to-market. We still haven’t figured out how to assemble variables like Early Access, F2P v. paid, console v. mobile, regional-locking, and Steam China into a coherent strategy. Chris Sides recently mentioned the “love-bombing” launch strategy, but this isn’t the answer. The industry needs a framework for testing demand before committing millions more to development. Mobile gaming solved this riddle years ago; 2025 must be the year HD catches up.

Famed Music Producer Rick Rubin said, “The audience comes last [in the creative process].” Game developers should heed this advice and avoid a “player-first” mentality. Creators must act in the best long-term interest of the game.

Games, like adolescents, need care and nourishment. Good parents make difficult, short-term decisions for long-term prosperity. Parents argue: “I’m just trying to do what’s in your best interests!” “Player-first” rarely extends beyond the immediate user base—insufficient for realizing potential. Games implementing strong monetization might trade short-term sentiment for capital to build features or acquire players. Simple survival benefits players, too.

Growth drives any game’s best interest, just as “human capital” drives a child’s future. Learning requires trade-offs but yields lifetime benefits that compound with time. This highlights survival’s importance. “Player-first” often means the current player base, not lifetime players. Developers recognize these trade-offs—invoking “player-first” mainly when giving away free content. Acting in the game’s best interest provides a better heuristic.

The player doesn’t exist at conception—the creative vision, not just the artsy, abstract vision but also the monetization vision, is a pre-player notion. While appeasing players is an input, it’s not the variable the creative process tries to maximize.

Consider World of Warcraft’s level squish in Shadowlands or EVE Online’s introduction of skill injectors. Both decisions sparked initial player outrage but served their games’ long-term sustainability. EVE’s economy needed new mechanisms for skill point circulation, while WoW required fundamentally rebalancing its level systems.

Star Wars Galaxies’ 2005 “New Game Enhancement” update shows the peril of overwrought reaction to player feedback—fundamentally altering core systems and streamlining its complex combat mechanics led to the game’s eventual decline, while EVE Online and WoW’s thoughtful but firm changes enhanced their longevity.

The most successful games maintain their creative vision while evolving through player interaction. Dark Souls never compromised its punishing difficulty despite vocal demands for an “easy mode.” Instead, it found its audience and influenced a generation of game design. Standing firm on core principles while thoughtfully acknowledging player feedback leads to lasting impact.

“Everyone has a voice, but not everyone has a vote” encapsulates the balance between creative authority and community engagement. It acknowledges player insight while preserving creative sovereignty. Moving beyond “player-first” allows for honest discussions about crafting enduring games.

When I wrote, Economics of Battle Pass Are Broken. Let’s Fix It in 2021; I argued that designers needed to play with ADMC (average daily monetization cap) to transform the pass from an engagement to a monetization vertical. Since then, every subsequent “innovation” has been a creative exercise in price increases.

Fortnite realized this long ago, which is why the direct store has ballooned in content and uses “the vault” to sell past pass-specific content; the 67% direct store, 33% battle pass revenue split leaked from the 2012 Epic v. Apple trial likely skews harder direct store since.

Two Options: d or fc

In the original model, increasing spend depth may be achieved by decreasing [latex]d[/latex] or the pass length, thereby shipping more passes over a given period.

[latex display='true']\text{average daily monetization cap} = \frac{\sum_{i=1}^N y_i + fc}{d}[/latex]

Mobile games mastered this in months, using low marginal cost, fractionalized, and storable content (currencies! resources!). For example, this was the exact strategy Marvel Snap employed at launch: passes refreshed monthly, and the game was stuffed with cheap content like titles and emoji stickers. This worked at creating consistent monthly revenue bumps. While there are trade-offs in shipping slimmed-down passes (50 tiers), it’s challenging to argue Second Dinner would make 3x the monthly average if they employed the traditional three-month battle pass instead. Respawn recently tried something similar, splitting battle passes into two season parts and dividing d by two.

Apex Legends Half Season Pass

HD games struggled as the pass became a content blackhole that spits little revenue in return. While there was always upsell on tier bundles with a base pass purchase, these were applied automatically, suggesting the player only forwent the more manageable parts of the progression curve rather than the higher XP requirement at late levels.

The new proposed solution is to grow [latex]fc[/latex] by adding another reward lane instead. Call of Duty’s BlackCell Pass, at 3x the cost ($30 versus $10), includes the standard pass, tier skip tokens (now as a storable currency to avoid the above problem), and ten additional rewards. It many ways, Activision just upsold a single usual $20 SKUs they’d otherwise put into the store as an entirely new pass. It’s a brilliant use of systems, wrapping content in a different monetization “sleeve” to maximize its value.

For its brilliance, Mobile employed a K.I.S.S. method, simply multiplying tiers into a new lane: another near-zero marginal cost innovation.

Three Lane Legends of Mushroom Pass (Two Paid, One Free)

Or adding multiple passes…

Legends of Mushroom Shroomie and Happy Farm Passes

The Experimentation Mystery

Another change has been to move the pass to a subscription: an automatic recurring charge. Fortnite implemented this in 2020 with Fortnite Crew, and while the program hasn’t died, it hasn’t precisely thrived either. Pokémon TCG Pocket is a recent attempt at this, offering an additional daily pack opening for active subscriptions on top of the typical lane system. It even boosts a two-week free “trial”!

An actual subscription seems obvious (and Apple sweetens the deal with lower tax if the player subscribes for a year or more) for battle passes, yet we’re seven years since Fortnite’s first pass, and the idea hasn’t taken off.

Analysts always praise subscription “sludge”: automatic charges that payers must actively unsubscribe to stop, and yet a neoclassical model offers more explanation. In a Beckarian view, players assess their time horizons and adjust the price for forgetting to unsubscribe, thereby increasing the “shadow price” of the pass. An easy experiment: implement the reoccurring charge but specify automatic cancellation if players don’t claim a monthly reward.

Games have also rallied around passes priced in real-world currency, not hard currency, to prevent players from earning enough hard currency from one pass to purchase another without spending real-world money. Brawl Stars did just this and was part of the platform that launched their recent wunderkind year.

The only countervailing trend has been a shift in HD battle passes like Fortnite and Call of Duty toward “Skill-Tree”-based reward selection. In this model, players select from a pool of rewards, and after collecting a given amount of page rewards, they are allowed to choose from a newly unlocked page pool. The practical result is players’ “short-cut” toward the final reward, reducing the effort needed to reach the last rare reward.

Fortnite’s “Skill-Tree”-based reward selection

From my conversations with people familiar with those initial designs, there was no profound economic reason for the change other than a sense battle pass was becoming dull, and choice gave players agency.

I wrote my initial Battle Pass series on the premise that Battle Pass evolution was only beginning, and seven years later, the industry has woken up to the realization that Battle Pass was too expensive for too little benefit. Players cannot pay for the pass in “time”: time does not pay VC, founders, or developers, and while k-factor is real, the inability to model an LTV on top of it suggests it performs as a hand-wavy argument. The industry’s obsession with engagement metrics as a proxy for monetization potential has been an expensive lesson in confusing correlation with causation.

Does a new Call of Duty map generate superior engagement compared to a new weapon? Which drives higher retentive effects: a new Magic The Gathering or Marvel Snap card?

In each question, retention is driven by more than the content but the systems that envelop them. In games, content operates and interacts in systems. In Overwatch, a mere balance adjustment—often implemented through a few keystrokes—ripples through other player choices. For instance, it may make mastering another counter-hero more critical. This way, a change that takes minutes to publish generates thousands of hours of strategic adaptation as players optimize their competitive advantage. However, a similar change, to say, a Vampire Survivors Hero, may not have as large of an effect, highlighting the system-dependent nature of the content.

Consider the curious case of Rocket League’s Flip Resets mechanic. Initially, a player could not jump in midair if they left a surface without jumping off it. Psyonix, the developer, thought it would be cool to jump after going up walls and allowed a player to “save” their flip, using it in midair. Now that players can jump midair, they can position all four wheels to touch the ball’s surface. This replenishes a jump (as if the car was touching a surface), allowing the player to jump again and create “swings.” The mechanic is now a standard and high-level player, driving thousands of hours of mastery as players learn it. Yet, the developers didn’t consider the change significant, so it wasn’t in the patch notes. The mechanic spawns from the interaction of the physics, map design (drive-up walls), and car abilities (jump); the systems operate on the content to make it viable.

The phenomenon adheres remarkably well to a Cobb-Douglas Production Function, where game systems (A) are technological multipliers enhancing player effort (L). This means players can achieve higher satisfaction in one game over another, given the same effort. Furthermore, the system’s design defrays content depreciation. A card game maintaining a meta of 300 cards presents a far more interesting puzzle than a meta of 30, as new cards suddenly make old cards useful.

In the headcount arms race, creating systems where a single piece of content is more valuable than alternative systems is one way to win. Investors understand that capital yields its greatest returns when deployed within the best-designed framework; the same lessons apply to content.

Call of Duty Black Ops 6 achieved the strongest paid Steam launch in franchise history. The franchise’s metamorphosis—from a modest war game that moved 5M units in 2003 to a cultural juggernaut shipping 20-30M copies yearly—is far more fascinating than marketing prowess—it’s a masterclass in industrial execution.

Like a Robber Baron Industrialist who strikes oil, Activision CEO Bobby Kotick’s primary innovation wasn’t in the product but in perfecting its extraction. The cascading studio system—a production architecture where three lead studios operate on offset three-year cycles—proved to be his equivalent of hydraulic fracking. Much less than “satiate” demand with an onslaught of entries, the annual releases grew demand.

Electronic Arts waited years before implementing strong support studios for Battlefield (and Apex!), and the cascading studio approach lived and died with Battlefield Hardline. The Battlefield franchise is a cautionary tale of hesitation, with Battlefield 1’s 21M copies the only time it outdid Call of Duty.

Kotick’s end-game is clear with Black Ops 6, as pieces coalesced around Modern Warfare III: horizontal and vertical franchise integration. Like Battlefield, players gravitate toward particular theaters of war, leaving them spread across different, non-monetizing franchise entries. A unified engine and single-app experience integrate the cascading model into a singular production line. This horizontal and vertical integration serves as another weapon in the Call of Duty “platform,” now functioning as a distribution center for diverse experiences: single-player campaigns, arena multiplayer, battle royale, extraction missions, and zombies—each feeding the central revenue stream.

As the game expanded, it doubled on the RPG foundations laid by Infinity Ward in 2007 to power each engagement pillar. Modern Warfare II “pro-tuning” created an almost infinite weapon permutations. Each annual entry embodied the “evolution, not revolution” mentality. However, consistent change, even if small, over many entries leaves a title nearly unrecognizable from its first entry.

Like Henry Ford’s production lines, Call of Duty has become gaming’s most robust engine—reliable, refined, and ruthlessly efficient—an anti-Ubisoft. While competitors chased innovation, Kotick proved that economies of scale are another path to dominance. While others hunted for the next breakthrough, Kotick built something far more valuable: a perpetual motion machine for manufacturing recurring revenue.

1. Distribution

The mobile “meta” is getting games in players’ hands post-ATT. The unintended consequences of Apple’s destruction of mobile acquisitions haven’t been to regain distribution control but for firms to find alternatives. “Content Fortress” ad networks, messenger apps, fake ads becoming white-lie ads, web stores, a return to brand marketing angles, and yes, web3, are all in play.

2. Distribution

See above.

3. Distribution

See above.

Joining the chorus of voices-heaping praise, Play Ventures put on a fantastic three days. Incredible event staff, an excellent venue, beautiful weather, intriguing conversations (shout out to Arya‘s team for a great fireside chat and interesting thesis), two Anton Backman jokes I’ll never forget, and sauna, of course.

FYI, it’s never seems to be “a sauna” or “the sauna,” just a “sauna” reminding everyone that this is an institution.

In tech writer Ben Thompson’s infamous aggregation theory, economic power derives from supply ownership—train barons owned the track and leveraged that power up and down the supply chain. The internet flipped the relationship: with near zero switching and distribution costs, suppliers instead aggregated user demand to leverage suppliers. Consider Amazon: aggregating millions of customers gained power over suppliers, compelling them to compete on its platform and play by its rules. In a post-ATT and “black hole games” world, gaming unwittingly embraced this dictionary, and distribution has become all-powerful.

Previously, games saw a lite version of aggregation theory, as platform owners like Sony and Microsoft owned the track (and thus users), allowing them to set the game’s rules. But something funny happened: cross-play, progression, and commerce crumbled console platform power as Epic leveraged Fortnite’s popularity to insert itself between platforms. Roblox grew so powerful that it bent Apple’s will, which granted it a “game-within-games” App Store exception. Roblox’s MAU nearly exceeds all game platforms combined, while Fortnite console MAU ranks first monthly. This consolidation of player bases has dramatically shifted the balance of power. The conquest continues, with Sony now allowing Discord voice chat integration on PlayStation consoles. This was unthinkable ten years ago!

Strong platforms exploit their position in and outside of games. Early MCU actors like Chris Hemsworth and Chris Evans were paid in the low nine figures on the notion that their careers would take off afterward. Marvel’s power came from massive distribution, reflected in these economic relationships. The same thing almost happened at last year’s Super Bowl: Artists who performed at half-time would pay the NFL instead of vice versa. While the NFL backed away, it’s another example of contracting battle lines being redrawn.

Roblox has succeeded in this for the last couple of years, and Look North World’s (LNW) Gundam deal signals it starting on Fortnite. Meanwhile, the Lego and Disney deals amount to Epic paying for the privilege of player access. While Gamefam and LNW pay for licensing, Epic, and Roblox pay nothing. Instead, they benefit from the engagement uplift and revenue share from in-game items, while brands access hard-to-reach audiences and enjoy adjacent brand monetization (toys, etc). The next question is when other large games like Brawl Stars and Call of Duty will extract terms from licensors given their 150m+ MAU (nearly half the size of the U.S.).

However, growth in these mega-platforms is showing signs of deceleration. Roblox’s growth has been essentially tier three, while Fortnite’s MAU has regressed, much less grown.

Ultimately, the next phase of gaming’s evolution will test the limits of aggregation theory. Success will depend on platforms’ ability to amass players and exploit their economic power.

At the Eras Tour in Stockholm, my thoughts naturally turned to Ms. Swift’s untapped monetization potential. Despite her billionaire status, she’s barely scratched the surface. By embracing “Taylor’s Version” labeling, Era-specific merch, and friendship bracelet culture, Swift could double her “extras” ARPPU from $50 to $100. ³

When the Eras Tour comes to town, media outlets scramble for clickbait headlines, and businesses concoct Swift-themed promotions faster than you can say “Shake It Off.” Brands quickly turn the occasion into a liveops event and co-opt the festivities. Along these lines, a local cookie shop created branded heart-shaped, glitter-coated “Taylor” cookies, but Swift received no revenue. “Free promotion” is a skeptical defense of using someone’s likeness for profit. Taylor has always maintained good-natured relationships with her fans, and pursuing legal action leaves a sour taste in everyone’s mouth. There’s a sensible solution hiding in plain sight: Taylor’s Version.

Post-Scooter Braun kerfuffle, *Taylor’s Version* became a rallying cry. Why not extend it beyond albums? Let businesses license the label, ensuring fans know Taylor gets her cut. Ka-ching: tens of millions in annual revenue.

The most striking part of the concert (beyond the concert) was the massive cosplay display. Each Era has a corresponding dress: cheerleading, orange riding hood, snakes. None of this was reflected in merchandise, which opted for the “Eras” brand itself, which has a distinct identity, rather than the coagulation of other Eras. It’s a monetization layup to let fans lean more into a specific Era and sell the merchandise. Fans want to explore more of the distinct Era identities.

The silly tip jars at cafes sometimes offer customers the chance to tip into a “Bulbasaur,” “Charmander,” or “Squirtle” cup, giving customers a sense of agency and identity while subtly encouraging them to catch ’em all – or in this case, tip ’em all. Where’s the Swift equivalent? It’s baffling that Swift’s team couldn’t come up with something more fun and engaging than the usual slew of t-shirts and tote bags. At a bare minimum, selling more Era-specific merch in a more fun way could increase conversion from 20% to 30%. ⁴

The friendship bracelet phenomenon (yes, I partook) begs for official Swift bling. Web3 taught us people love unique, tradeable items. Swift-branded beads and charms could net $20 per spender at 20-25% attach rate.

Lana Del Rey says, “Swift wants it more than anyone else.” Time to apply that hunger to brand expansion. Ditch the Wegman’s kids’ books; embrace Taylor’s Version of merchandising.

Economist Tyler Cowen notes, “Swift’s ability to capture consumer surplus is nothing short of remarkable.” It’s time for Taylor to shake off hesitation and fully embrace her economic potential.

Thanks to Laura Taranto and Matthew Emery for the feedback and editing!

Verdict & Prospects

As it stands, Squad Busters is not a game players will remember forever. But it can be one. 

Squad Busters demonstrates Supercell’s world-class ability to broaden the accessibility of midcore mechanics, in this case, roguelikes and MOBAs, while weaving an effective character progression consumable-based monetization model. Squad Busters promises to evolve into a stronger, more robust, structured sandbox than its seeming counterpart, Brawl Stars. Squad Busters is Supercell’s most accessible game yet and has the most sustainable revenue potential among their offerings, bar Clash of Clans. 

It’s also a crowning achievement in Supercell’s IP building. Squad Busters’ characters draw from other Supercell IPs, with endless and tantalizing future potential tie-ins. Yet the art style, gameplay, and wholly original IP elements (that theme music!) manage to expand and enhance the IP it’s based on. Potential cannibalization between Brawl Stars and Squad Busters is overblown. We should expect a portfolio-high download volume.

However, a lackluster learning loop, muted character design, and two-in-round currencies hamper the ability of Squad Busters to transition players to elder gameplay depth. Long-term prospects remain the game’s biggest mystery, as the game was never soft-launched long enough to measure so much as a D30 retention rate. While these are fixable issues, the team must solve them “live,” which is much more difficult if the team headcount scales. Ultimately, Squad Busters needs more refinement time. 

In the short run, and as the first game to go global under Supercell’s “Next Chapter,” it has a shot at helping Supercell achieve its highest annual revenue on record. That would be a welcome reprieve for a company that seems to be under mounting pressure from its owners, Tencent. However, solving Squad Buster’s core issues is paramount to avoiding the inconsistent performance of Clash Royale and Brawl Stars.

Supercell State of the Union

Supercell is launching Squad Busters globally after a six-year drought. Since their last launch in 2018 (Brawl Stars), they’ve killed six soft-launched titles and one announced title unofficially (Clash Heroes). This includes every member “trio” announced in April 2021: Clash Quest, Heroes, and Mini; Supercell canceled Mini in April after a three-year development cycle. The modus operandi were games with long soft launches that failed to elevate their KPIs to global-ready.

In contrast, after two short beta periods earlier this year and less than a month of soft launch, Supercell will release Squad Buster globally. This marks the quickest between a Supercell soft and global launch, although Clash Royale isn’t far behind with a 2-month soft launch period before a global launch in 2016.

Why the rush for Squad Busters? While it’s not entirely clear, Supercell has been undergoing radical transformation. Supercell “Cells” now mimic hierarchical management structures, including titled positions like “Producer” and “Product Manager.” This includes a significant headcount increase, with nearly 45 on Brawl Stars compared to their previous high of 20. “Small teams, small growth” is long dead.

Pictured: Timur Haussila, Supercell Game Lead, 2016 Montreal International Game Summit. The traditional model ended up becoming the Supercell Model.

Scaling has been widely successful, unlocking unbelievable Brawl Stars growth but more muted effects in Clash of Clans (CoC). However, CoC is likely year-over-year positive when combined with increased webshop revenue share. It’s difficult to line up headcount and revenue exactly, but we know Brawl Stars at least doubled in headcount from the summer of 2020 to 2024, while CoC’s team size doubled in 2023. The scaling status of Hay Day and Clash Royale is unknown, but job postings suggest they’re well underway. 

Labor’s Effect on Capital Takes Time

Quarterly Net Revenue, $USD, Sensor Tower

Changes sometimes cascade like landslides, so maybe a new game launch strategy is a part of the Next Chapter. On the other hand, it’s unclear what theory underpins the new approach.

Traditionally, games remained in soft launch to either fade away or improve KPIs for a global release. Soft-launch KPIs predict global KPIs, so companies only scale games with strong chances of scaling revenue. Instead, Supercell is launching Squad Busters globally with less data and seemingly more faith. This mirrors the recent launch of Match Factory, which is on its way to a $200M run rate.

The worst launch risk is a burned launch budget and perhaps opportunity cost in repairing and scaling Clash Royale. On the other hand, the upside is high. If Squad Busters performs similarly to Brawl Star’s launch ($250M) and Brawl Stars maintains its current position, 2024 is poised to be Supercell’s highest revenue year, breaking its all-time high of $2.3B in 2015. This short-run revenue bump may be behind Squad Buster’s launch push. 

Since Tencent’s acquisition over seven years ago, Supercell’s revenue has dropped 20%.  Tencent itself has been suffering, down from pandemics and even 2018 highs. Combined with Supercell’s cultural relaunch, it’s fair to wonder how much Chinese pressure played in Supercell’s Next Chapter.

Tencent’s Rocky Investment

Supercell Annual Revenue, $USD, Actuals 

As recently as a year ago, Eino Joas, Supercell’s top design chief, became Squad Busters project lead, even joining Ilka for the global announcement. That suggests real stakes are riding on Squad Busters, and a successful launch is the ultimate ratifying vote in Supercell’s Next Chapter.

Game Overview

Squad Busters screams accessibility, and its only miss is not pushing the accessibility to as many platforms as possible. The North American Supercell studio should be working on adapting a game like Squad Busters to the browser (hello, webstore revenue) rather than the original IP. 

Movement

As the game’s menus fondly remind us every round, players move with a virtual joystick and stop moving to engage in combat. Instead, the squad members automatically perform attacks or PVE actions (i.e., cutting a tree) within an “action circle” when the player stops moving.

The only other controls are the ability to “Turbo” (sprint) and throw in-round consumables like bombs and cannons. 

The controls constrain real-time player input agency in favor of mastery over effectively positioning units, selecting squad units, deciding when to disengage from combat, and how and when to use consumables.

Drawing from roguelike and battle royale, ten players compete on a shrinking map while building a unique squad by recruiting characters using resources collected through PvE and PvP combat. A player’s gems balance determines their ranking at the end of a match timer, similar to the Brawl Star’s Gem Grab mode.

Each battle starts with a random map selection, layout, and “gameplay modifier” to provide match variety.

• There are four “worlds” (read: biomes).
  • A player’s account level (“Squad Journey”) defines the current world and, thus, unlockable characters and maps.
  • It also segments the matchmaking pool, similar to Clash Royale’s arenas. 
• There are around three different layouts per biome (15 total maps).
  • Each world and world variation needs to include relevant PvE elements for characters unlocked in that particular world. For example, Marvis picks in-round carrots for gold and unlocks in “Royal World.”
  • Opening a locked character from an in-round chest is possible; collecting units is heavily incentivized! 
• Before the round starts, a modifier is chosen at random.
  • A modifier exaggerates rule sets: a Piñata may spawn with additional gold, the speed boost may become more powerful, or each player might start with a free Mega unit.

As rounds start, players choose one of three characters they’ve unlocked to build a squad. 

• The pool of available characters varies every match and between players in a match.
  • The pool includes fifteen randomly selected characters, including locked characters.
  • As of this writing, there are 27 characters, so a player utilizes 50% or more of their collection in a match. This will decrease as more characters are released.
• Each character has unique abilities and emphasizes different strategies. For example, Greg is the only character who cuts trees to reward gold and gems.

Match Opening

Players defeat enemy bots and other players to collect gold and keys, unlocking chests containing three characters to add to their squad.

Mastering the game involves learning fast-paced grind strategies to build massive squads quickly.

In-Round Progression

Players must choose which characters to pair well with one another and execute against the squad’s unique winning blueprint (collecting farming, combat, etc). Another mechanic, Fusion, introduces long-run consequences for early character choices.

Selecting a third duplicate character merges them into a stronger single (“fusion”) unit while decreasing the gold cost of the following chest open.

• Chest open prices rise with squad size, so fusing shrinks prices.
• The fused character is greater than the sum of its parts: 3.5x base unit health, 2x DPS, and varying ability upgrades.

Each character has a class (e.g., “Healer,” “Speedster,” etc.), but these have no practical effect.

Example Chest Open; Squad Member Selection

Over the 4-minute round duration or until other players are eliminated (“Busted”), ten players compete on a battle royale shrinking map to collect the most gems.

• A player’s streak level, or the number of consecutive top-five finishes, is prominently displayed.

End of Round Results Screen

The map is divided into ten quadrants, one for each player. 

• The map spreads boxes containing in-game consumable items (bomb, cannon, speed boost, etc) throughout.
• The center contains the gem mine and more powerful enemies, which reward more gems and coins.
• The center of the map includes a gem mine, taken straight from Brawl Stars, periodically spewing gems.

Gem Mine; Map Center

Map Design

Drawing from MOBA and RTS, the game maintains three phases: 

• (1) Early Game: The Grind Phase
• (2) Mid-Game: The “Jousting” and Recharge Phase
• (3) Late Game: Resource Control and “Final Clash” Phase

The in-round economy and unit design reinforce these phases. Amassing more squad members is always positive, and each of the three character rarities generally corresponds to a phase. High-rarity characters appear more often in high-rarity chests, which appear more frequently as round time elapses.

The marginal cost of a Squad member increases, so to continue to grow, players need to move inward on the map where stronger enemy bots drop more gold and gems. Vines encasing a shrinking map also guarantee inward movement into player-versus-player (PvP) conflict.

Enemy “density” measured in gold likely increases as players move inward.

The roller coaster of up-and-down price schedules creates the game’s second phase, where players skirmish one another before recharging for another bout.

Combat

PvP combat consists of simply not moving while atop other players’ units.

Attacking slows another squad’s movement speed, so there ought to be a limited margin for error when sizing up another player’s squad and observing how combat unfolds. If a player fails to retreat in time, the opponent will likely wipe them. 

As both squads dwindle, usually one player calls chicken and uses the “run” button, starting a hot pursuit. While attacking slows down an enemy squad, it only slows down a select few units of the enemy squad. These slowed-down units are focused and picked off, while the remaining squad members can create distance between their retreat and the attacking squad. Escape seems to have an extremely high success rate.

Lengthy Resolution Disincentives PvP Combat

GIF at 2x Speed

A successfully escaped player usually returns to the grind phase, rebuilding and retooling their squad on the price roller coaster. PvP should be high-risk-high-reward, but combat takes so long that players lose out on farming and the potential of another player joining the fray.

The end-of-round reward for placement compounds the problem. There’s so little to gain outside of expanding a Top-5 streak that players avoid combat despite shrinking map area and gems spawning from the map’s center. The payoff curve to placement is dull compared to something like F1, whose points incentivize fierce competition for top placements.

Points/Taps Per Place, F1 & Squad Busters

Less combat means the game eliminates fewer players, enhancing accessibility by maintaining a 50% win rate compared to the 1-3% win rate typical of battle royale games like Fortnite. Instead, winning means placing in the top-5 and extending win streaks. The rewards design re-enforces the win streak goal: nearly 30-40% of a rewards end of round rewards are sourced from win streak extensions rather than scoreboard placement. Positive audio cues start at 5th place (“Good job! 5th place!”),

The combat loop of attacking, losing units, and retooling drives in-match tactics. For example, if my opponent specializes in speed units, I need more turbo units to close the chase successfully. Or I may need more AOE characters to stop the skeleton summons of the Witch. Yet combat is chaotic and hard to read, so evaluating which squad strategies excel as the match phase progresses is difficult. 

Reading and assessing the strengths and weaknesses of seven to eight-unit squads in split seconds is impossible, so retooling feels aimless as it’s unclear what strategies are effective against which armies. Squad Buster’s “learning” loop needs more tuning, as it’s charged with carrying players to the game’s deeper Day 30+ mechanics.

Learning Loop

Each character unlocks a unique ability when they “evolve,” but the player would never know as they look and sound silent. 

For instance, one of Archer Queen’s abilities increases the attack speed of all ranged units in the squad, but there’s no discernible effect or UI icon in-game to suggest this is taking place. The same is true for communicating base stat information like critical hits to the degree they exist in the game. 

Where Art Thou Abilities?

While base character stats are visible during unit selection, the lack of an end-of-round report makes evaluating the squad’s performance difficult, and there’s no sense of squad “power score.” Introducing color of damage indicators (or damage indicated at all) would improve the learning loop.

For example, one of the Barbarian’s abilities is to evolve into an “Elite” version after slaying three enemies in the round, but again, the in-round feedback is silent. It’s unclear if the ability triggers when the character reaches a certain amount of damage dealt.

Another example is the Hog Rider’s passive ability to run faster on grass terrain. 

While the unit flashes as the ability is active, the visual readability of six to seven other units (Hog Rider is a mid-game unlock) makes it hard to draw the Hog-Terrain cause-effect. Conversely, the 2D Hog icon attached to the turbo button taught me the importance of Hog Rider in closing chases!

On the flip side, units have abilities that trigger conditionally, and an outer purple circle indicates the level of ability charge. The charge condition changes: one character may earn a fixed charge amount per enemy defeated (Primo) while others passively earn over time (Archer Queen). It’s difficult to nail each case, but the rings are effective.

However, unfinished in-match UX also fails to surface deep engagement features, like turbo attack speed. The turbo button’s attack benefit should not remain hidden, and simply changing the icon to the sword as players remain still is a straightforward fix.

While some of these issues may appear nitpicky, status effects and combat readability were an improvement area of focus for the now-defunct Clash Mini. Squad Busters doesn’t need to lose players in transitioning to a deeper strategic game. Tightening the learning loops should ensure Squad Buster’s accessibility promise runs through the game’s progression cycle.

However, while additional readability improvements will arrive in a future update, it’s precisely the type of work accomplished within a month or two of soft launch.

Character Design

Squad Buster’s biggest enemy is the dull character variety before Royale World, a Day 14 or later unlock. The consequences strike at the heart of its unique in-round progression and roguelike promise. 

Given the game’s low real-time input ceiling, Squad Buster’s long-run hook lies in compelling theory crafting as players build their squad, similar to roguelikes. Like card combos in Marvel Snap or builds in Vampire Survivor, players thrive on the ability to craft and execute imagined strategies. The randomness of the in-round characters or cards alters the win probability maximizing strategy, so players must adjust and solve a real-time puzzle. Theory crafting is best when the player’s imagined simulation of stacking characters or abilities becomes content onto itself, and this requires diverse and rule-bending abilities and characters. 

Expanding the character roster to fifteen at Royal World greatly improves the experience (and does so again at Beach World). It should be yanked to earlier levels, including many later characters too. Characters like Goblin (give coins) and Trader (convert coins to gems) combo uniquely, unveiling Squad Buster theory crafting. 

The characters need to continue to “bend” rules: the “action” circle size never expands, shrinks, or changes shape, the loot magnet range never changes, and even “slowdown” effects appear only as PvE elements like quicksand rather than character abilities. The stats system feels bare bones, and it’s never clear if units maintain something like critical chance damage. Supercell is probably holding a feature trick or two up their sleeve for launch, and something like autochess alliances or a slew of stronger character designs is needed. I’m rarely excited to test new strategies each run, and that is never the case in popular roguelikes like Balatro or Survivor.io.

Dual-Round Currency (Gold & Gems)

While players earn squad members by spending in-round gold, they ultimately win by ranking on a leaderboard of “collected gems.” Most enemies drop gems, as does the mine at the map’s center. “Busted” players will also spew their remaining gems after periodically shedding them when attacked. However, keeping track of both currencies is difficult.

When it was a Clash Royale clone, early designs of Monopoly Go suffered from the same dual in-round currency problem, but with only three other players compared to Squad Buster’s nine! A player’s cognitive budget is limited during intense brief battles, and the more real-time assessments a player has to perform, the more other info retreats to the important periphery. 

While the UI prominently displays a player’s gem-based leaderboard position, the relative lack of short-run gem impact makes the experience of gem collecting unsatisfying.

The win condition needs more experimentation; it’s not ready for prime time and would benefit from more liveops events testing.

Meta & Progression

Squad Busters marries out-of-round progression and roguelike elements. Even successful live-service action RPGs like Survivor.io have struggled to monetize the game’s core appeal and new abilities earned in-round, but Squad Buster nails the assignment.

In Squad Busters, locked characters may still appear in in-match chests, so players want to grow their character rosters to avoid missing out. The random chance of each character appearing means the player is incentivized to equalize gold spent purchasing character copies in the shop. Players are never sure which character will appear in in-match chests, so equal spending per character shrinks variance. The randomized character selection also prevents content bi-fiction: every player needs to master every character to be successful.

Squad Journey

Players progress along an account level (“Squad Journey”), which gates the key content: maps and characters. They earn “Portal Energy” either by evolving characters or clearing achievements.

Meta Flow

Character evolution is straightforward. 

Players earn “Baby” copies of characters. After collecting ten Babies, they convert them to one copy of a “Classic.” After ten copies of a “Classic,” players convert a copy to “Super.” Priced in Babies, the progression cost is exponential – from 10 copies to 100 copies to 1,000 copies at Evolution 3. 

Taping each age visualizes the number of copies players own, with units scrambling on and off placemats. The character upgrade screen UI is…strange and reminiscent of a cartoon abacus.

Character abilities unlock and sometimes existing abilities improve at each evolution stage, and stats improve, signified by a star count (Evo 1 is one star, Evo 2 is two stars, etc). The fourth evolution remains locked and is likely a launch item.

Each time ten units evolve to the next stage, the player earns portal energy, increasing their Squad Journey level (i.e., Star Road or a traditional account level system). This is the key gating mechanic behind the ability to earn new characters and levels. Portal energy level-up costs increase exponentially, and so do world lengths. The first levels increase at ten portal energy per level, while later levels increase by 100 per level.

Besides character upgrades, a looping set of basic “achievements” reward portal energy. The achievement’s design is akin to “play the game, and things will trigger.” The design is straight out of Brawl Stars. There’s no need to reinvent what works; the collection loop is satisfying and yet another chance to demonstrate Supercell’s top currency animations! 

Speaking of other Supercell influences, Squad Busters uses bots to boost average win rates over 50%, similar to Clash Royale and, famously, Marvel Snap. This feature is becoming standard for mobile PvP games, and even HD titles like Battlefield are using bots to soften PvP zero-sum brutality.

Economy

Eino Joas is fond of consumables, and they shine here. Squad Busters’ heavy monetization of zero-marginal cost mechanics while pay-to-win has never been stronger. In fact, it’s astonishing. Payment proximity to win probability has never been higher for a mobile F2P PVP game, bar maybe Yahtzee with Buddies Bonus Rolls. Despite this, prices are too low in the economy, while Mega units deserve stable monetization avenues.

Out-of-round, players fill a “Battle Backpack” containing four key types, each corresponding to a chest type. After waiting 60 seconds, players may activate a key in-round – the next chest opened is the one associated with the key. Crucially, the keyed chests require no gold to open, immediately making the player more powerful. This is compounded by “Mega” units, a special type whose probability of appearing in chests changes based on in-game factors.

One modifier event will grant players a free starting Mega Unit. Try before you buy!

Mega units are powerful squad members (~3-5x health, 2x DPS above average) that cannot merge. Drawing a Mega unit from a chest can significantly change the game. Interestingly, the drop rates of Mega Units are conditional on several factors:

• Having fewer gems than the round leader
• Having a smaller army than the largest army
• Unlocking both Mega units increases the chance of getting a mega unit. 
  • (For example, El Tigre might have a 3% drop rate, while Dragon Chicken might also have a 3% drop rate, meaning the chance of getting El Tigre OR Dragon Chicken is higher when both are unlocked).
• Round time elapsing (increasing drop chances)
• Epic and Fusion Keys also increase the chance of opening a Mega unit

Top-tier players already bemoan Mega Units randomness but also represent the game’s commitment to accessibility. In the same way that Mario Kart’s rubber-banding system ensures everyone has a chance to win, Mega units do the same here. 

Interestingly enough, Mega units are consumables. Every time players select one from a chest, they lose a quantity in their inventory. Top-tier competitive play guarantees play maintains some inventory, so it’s curious that Supercell didn’t provide a direct and consistent way to purchase Mega’s rather than rotating shop appearances. This is an easy avenue to grow consumables revenue, which appears too low.

The consumables model consists of tickets, streak preservation, key usage, and mega usage. Tickets act as reward “energy”: players who consume a ticket earn Babies copies at the end of a match. Squad Buster designers tightly control F2P progression, as a mere three refreshes every 24 hours. However, as the match-3 consumables-based model demonstrates, revenue is not only a function of price but how often players rotate through the turnstile. The more matches players play, the more times consumables usage triggers, and the more players need to re-stock. 

Various ticket engagement assumptions (three to twelve per day) and consumables usage rates range from $400 annually to $1,800. This feels particularly low. The game needs to experiment with more “default” booster usage: similar to the Mega units, spenders must “opt-in” to key spend by taping two screens during a real-time battle. Default key usage when their cooldown expires or some default way of activating keys is a solid way to “nudge” monetization.

Using gold prices from the daily store for price conversion, Mega Unit costs about $0.44 per use, while keys range from $0.03 to $0.13. This means the game can move monetization figures by rebalancing (i.e., increasing the Mega Unit spawn rate). But it’s also much lower than match-3, who might charge $1-2 for five extra moves. Growing the consumables sink per match will be a core growth area for the game’s revenue.

Streaks

Streaks, or consecutive top-5 finishes, are in the game’s limelight and are home to Supercell’s new “Jony Ive Pen”: the chest tap.

If players rank in the top half of the four-minute round by gem tally, they extend a “streak.” The game’s reward structure suggests maintaining win streaks is a significant progression source. When presented with a chest at the round end, players earn “taps”: each time a player taps on the chest, it may increase rarity or add Baby copies to the eventual open. Ladies and Gentlemen, Supercell has built a meta loot box.

The final cartoon-style Chest squishing always gets me.

Squad Busters incentivizes players to earn taps from “streaks” —the streak length functions as a vanity metric as it’s displayed on profiles and leaderboards. There are three streak reward tiers, and the maximum streak, at ten wins or more, grants three additional taps on top of the number of taps per placement on the leaderboard, which ranges from two (10th place) to four (1st place). A “maxed out” player would maintain seven taps per round end chest reward. The game cleverly mirrors match-3 style monetization, where streaks amplify the benefit of spending on extra moves. In Squad Buster’s case, they scale the gold price of streak preservation on a loss with streak level. Who knows, maybe Episode Race is the next Squad Busters liveops event?

Character upgrades need to form the backbone of the game’s monetization, and while it’s strange to see no gacha, hard currency (!) or soft currency level gating makes an appearance. At $2,000, the spending ceiling feels low, as does the ~$100 per Epic character price. However, given the exponential requirements, the cost curve will spike when Evolution 4 is released. This will likely coincide with more generous sourcing of copies, including “Classic” and “Super” copies.

Store

The game’s shop is dynamic, triggering based on heuristics like a player’s world level, but the seemingly self-imposed limit on spend velocity is entirely unnecessary. After purchasing a character, the store marks the slot as claimed rather than refilling the slot with other characters. Outside of consumables use, the economy effectively limits daily Gold spend. Putting aside characters (~$9), this translates to under $10 for all offers.

A yellow arrow indicates the character is ready for evolution with this purchase. Now, that’s combat readability!

At the very least, players should be able to re-roll offers for Gold, a sort of “inverse loot box” similar to FIFA Ultimate Team’s X-Ray packs. 

Given Supercell’s new focus on scale, Squad Buster’s design is a perfect runway to ramp content. While designers will need to limit the character pool randomly (or will they?) as it expands, adding new characters almost creates network effects since the abilities may interact to develop new strategies. The more characters, the larger the network of potential interaction.

Cannibalization 

Nearly all of MiHoYo’s Honkai Star Rail’s revenue came from Genshin Impact, a massive failure considering the marketing and development expenses.

Honkai Star Rail and Genshin Impact Failed to Grow Portfolio Revenue

This has caused concern that Squad Buster will eat from Brawl Star’s growth and compete for users. Yet, Squad Buster IP differentiates itself and could hardly be confused with Brawl Stars: bangin’ new theme music, characters from every launched Supercell game, and an ample dose of mischievous baby characters. 

There’s no guarantee that the title will cannibalize; the differentiation axis is important. For example, there’s no evidence of AFK Journey eating from AFK Arena players, and the same is true for Coin Master and Monopoly Go.

AFK Journey: Little Impact on AFK Arena

A Unique Sandbox

Squad Buster’s sustainability is a function of its monetization model (consumables), progression model (randomized), and core gameplay design (structured sandbox).

During Battle Royale’s (BR) ascent, teams questioned whether BR was a mode or genre. As BR celebrates its seventh anniversary, each mode attempt has languished (Battlefield Firestorm) or spun off into a monster hit (Call of Duty: Warzone). Now, BRs account for nearly half of FPS revenue and certainly a plurality. Much less than a mode, BR proved to be the FPS sub-genre. Ultimately, observers discounted the repeatability of BR’s roguelike and randomized loot elements.

The same can be said about Squad Busters, which, while inspired by Brawl Stars Gem Grab or Death Valley mode, provides a structured sandbox to Brawl Stars character platform. If Brawl Stars is a hero shooter like Overwatch, Squad Busters is a battle royale like Apex Legends. Each of these genres and games lives and thrives independently of one another.

If Squad Busters actually does stunt Brawl Stars, declining downloads will be a telling sign as they compete for similar player pools. 

Conclusion

Squad Busters is an accessible entry point to innovative roguelike and MOBA mechanics in a battle-royale-esque setting. It has elements to maintain and grow revenue for itself and for Supercell as a firm. Despite this, long-term retention still needs to be seen. The game’s lackluster learning loop, dull early game character design, and two in-round currencies are solvable problems before a global launch. However, doing so at a higher headcount level is a new challenge for Supercell. Will more headcount increase roadmap velocity or drown the title in bureaucracy?

If they’re successful, not only will Squad Busters be another entry in “games that will be remembered forever,” but Supercell’s Next Chapter reads longer than its last.

The Fallout TV series has received widespread critical and audience acclaim, with 65M viewers and Amazon Prime’s 2nd most-watched show of all time. The show appears to have cracked the secret transmedia success formula: top-tier IP, backed by a six-figure budget, distribution through a 200M MAU network, and widespread critical acclaim. Consistent with the transmedia thesis, engagement with Fallout games has skyrocketed, with *all* games experiencing significant uplifts. However, price cuts have ended, and according to VG Insights, revenue is down 50%+ from two-week highs in titles like Fallout 76. While the uplift has yet to return to equilibrium, a generous guess might land $40M in incremental revenue across the franchise before returning to a steady state. The final revenue impact may tally north of $50M. And yet, hailed as a transmedia success, Fallout instead shows the exact opposite: if this is the best transmedia offers, then the concept is creatively bankrupt.

The marginal revenue uplift from Fallout game sales amounts to little in the face of the show’s spend. And yet, transmedia’s biggest win should come from “rising IP tides.” The recent track record isn’t great: the Warcraft movie did little, nor did the Mortal Kombat, Assassin’s Creed, and Gran Turismo movies. The Halo and Witcher TV show’s creative atrocities hurt the franchises rather than help them.

Variety reports that Fallout cost $150M, which suggests a marketing budget north of $100M. $50M in incremental game revenue amounts to ~20% of the TV show’s total spend and likely less, maybe ~15% of its attributed revenue. Yet, the *expected* outcome should exceed 20% for the transmedia thesis to grow.

Transmedia isn’t a TV show based on a game more than Caesar salad is a side dish based on Italian cuisine. Instead, Transmedia’s entries should form a “core loop” – one media entry leads to another rather than mere adjacency. However, the Venn Diagram of viewers who consume differing formats is definitionally smaller than those who consume one media type. Creatives haven’t figured out how to optimize the format, giving consumers a collection of dishes rather than a meal.

As Jen Donahoe aptly suggested, the right word for the game-inspired movie and TV show is “franchise.” Here, the most significant effect is past entries building an audience for the next entry, not the next entries uplifting past entries. Five Night At Freddie’s is a prime example, using the property as a story distribution channel to the tune of a $78m opening. However, the nature of game-based uplift differs from books or music.

Games experience significant capital decay: Fallout 3 jumped +300 concurrent users (CCU), while Fallout 4 jumped +150k CCU. It’s hard to imagine similar declines between Harry Potter book releases upon a new film.

Zero marginal cost revenue is a nice treat, but successful strategies require scale, and transmedia has more to prove here.

Inflation, ads, and webstores reverse nearly every trend in gaming—and then reverse it again. These aren’t rounding errors either—they’re GDP-level phenomena (literally), and the inability to account for them is warping our perception of reality in the gaming industry. 

• Adjusting for inflation completely reverses the growth in mobile spending between 2020 and 2023, from +8% to -9%.
• Adjusting for inflation slashes Newzoo’s 2026 PC & Console forecast by 39%, from +15% to +9%.
• Ads revenue is likely a double-digit percentage of gaming revenue, and all we have to estimate is King’s quarterly earnings.
• Adjusting for webstores completely reverses the revenue decline in some midcore genres, casting cold water on the “move to the middle” thesis.

Inflation

Data.ai’s 2024 State of Mobile revealed that U.S. gaming spending grew 8% from 2020 to 2023 despite a rocky road. This was a relief after the headwinds of ATT and the post-COVID freefall. 

However, inflation changed dramatically over this period, with $100 in 2020 comparable to $118 in 2023 dollars. After adjusting for inflation, spending is down 9% rather than up 8%.

Newzoo’s new 2024 report follows the same pattern. They report a 52% increase in PC & Console consumer spending between 2015 and 2023. However, adjusting for inflation results in a more modest 32% increase, a -20% percentage point or -39% percent change.

The issue extends to forecasting, where Newzoo predicts a 15% increase in consumer spending from $93.5B in 2023 to $107.6B in 2026. ⁵However, assuming a 3% annual inflation rate, $93.5B in 2026 dollars is $102.3B. The inflation adjustment suggests a 9% real increase in spending between 2013 and 2026 rather than the 15% nominal increase, a whopping difference of 60%. These different outlooks must be adjusted to get a realistic picture of what’s ahead.

Ads

App Purchases (IAP) famously monetize a hyper-small proportion of users. In-game ads monetize the other 90% on top of freedom from the App Store’s 30% cut, but it’s hard to know the industry-level impact. Obtaining estimates is surprisingly complex. Unlike IAP, intelligence providers like data.ai and Sensor Tower can’t estimate ad revenue using credit card panel data; revenue isn’t collected at the point of viewing, instead following a sophisticated mediation process. The only coherent analysis comes from Eric Seufert’s back-of-the-envelope math derived from first-party King data. In his view:

Even if the underlying assumptions used in the article are fundamentally and severely flawed and the true extent of mobile games advertising revenue on mobile is half or a third of $100BN, it’s clearly a massive, important industry.

Eric Seufert, How Large is the Mobile Gaming Advertising Market?

Incestuous ad revenue further complicates the analysis, particularly hypercasual. These titles trade users and revenue between each other as opposed to “fresh” spend. 

And if that wasn’t enough, time series analysis is complicated by lowered eCPMs post-ATT. I suspect an increasing number of games employ ads, similar to the King example Seufert reports, as technological advances have made integrating SDKs easier.

No matter how it’s sliced, ads comprise a double-digit percentage of mobile game spending, likely all game spending. If ad revenue coughs, game revenue catches a cold. The problem is that we’re trying to diagnose the illness without a thermometer.

Webstores

Webstores have exploded in popularity in another attempt to escape 30% store fees. Aided by Epic Game v. Apple, firms may now contact users with information gained from the in-app experience to direct them to off-platform payments. Lilith’s recent SquadRPG, AFK Journey, sends messages via WhatsApp to those who opted in for 5% discounts for hard currency bundles.

AFK Journey’s New Off-Platform Approach

Playtika reported that webstores account for a gigantic 25% of revenue, and I’ve heard offhand that webstores reach 40% of net revenue in some products.

Playtika’s Off-Platform Payments Maxed at 25% of Revenue

We’d expect webstore usage to rise as a given player’s spending rises—heavy spenders benefit most from percentage savings. This assumption completely reverses the ATT hypothesis of a “move to the middle.”

ATT reached a majority of iOS devices in Q3 of 2021, and it appears that midcore genres experienced a significant revenue decline while casual increased. RPG went from comprising 53% of the cohort’s revenue in January 2021 to 46% in Feb 2024, a drop of $250M per month, based on Worldwide revenue from iOS devices. The fall isn’t as grand as for “Strategy,” which dropped about $100M monthly between the two periods. The gains for Match are modest, about $30 per month, but it eclipsed “Strategy” for the first time in January.

Match-3 for Success? Think Again

Laying growth in webstore revenue share reverses trends. Playtika reported 13% of revenue from webstores in Q2 2020, which has grown to 25% in FY23 Q4, a jump of 12% percentage points. ⁶ Applying a linear trend from 13% to 25% to reported revenue between Strategy and RPG tells a dramatically different story. Strategy revenue grows by 6% (nominally!) compared to a 15% decline, while RPG declines 12% compared to a 29% decline without a webstore adjustment.

To be clear, there’s a lot of hand-waving in these estimates.

• Is using a trend line from 2020 when starting in 2021 appropriate?
• Is Playtika’s webstore revenue share figure appropriate when considering other genres? Higher? Lower?

Most importantly, to poke holes in the “shift-to-the-middle” thesis, the share of webstore revenue from midcore genres must be higher than Match. Otherwise, all that’s shown is a level shift rather than a relative change in the revenue mix between genres.

We Hang Together or Separately

Where does merging inflation (revenue down), ads (revenue up), and webstores (revenue up) leave gaming as a whole? Quite frankly, I don’t know. But I do know these ticket items are huge movers of industry GDP. The absence of evidence isn’t evidence of absence, but we need to start building that very evidence. At a minimum, intelligence firms should adjust for inflation, while I know webstore revenue and more robust ad revenue reporting are on the roadmap. Next time you see an estimate, be the annoying guy in the back and ask, “is this adjusted for inflation?”

Call of Duty Mobile (CODM) provided a watershed moment for mobile, proving Western gamers are ready for virtual dual analog stick controls and heralding a wave of Eastern titles to the West. Or did it?

Four years after CODM’s launch, the graveyard of aspiring mobile shooters’ is extraordinary: Undawn, Lost Light, Blood Strike, Valorent (canceled), and Battlefield (canceled). These are high-profile games, too: Undawn reportedly had a 300-person team on a $130M budget to only amass $300k in revenue last month. Conversely, a few titles have succeeded, albeit primarily outside of the West: PUBG Mobile ($1.1B), Crossfire (~$500M), Arena Breakout ($50M), and Knives Out (~$109M). However, in the U.S., CODM accounts for 97% of revenue. Surely, this sets up Warzone Mobile for success?

CODM embraced an auto-fire mechanic to navigate the inherent awkwardness of mobile FPS controls. Dual analog sticks require both thumbs, and without shoulder triggers, mobile players need to lift one of their thumbs to fire. Instead, autofire will shoot once a target aligns in the crosshairs. While less competitive (top players use their index finger to fire, creating a c-shape with their hand), the title is more accessible. Much of this accessibility is awash in Warzone’s battle royale (BR) mode. Warzone’s maps are MUCH larger than CODM’s maps, meaning draw distances are much longer. This makes it harder to spot enemies, penalizing autofire players. CODM’s BR mode avoids these problems by switching to third-person.

Spearheaded by famed dev Timi (w/ATVI PMs), CODM flawlessly executes the Chinese liveops strategy: massive content and loot box monetization. This approach, mirrored by Apex Mobile (another Timi collab w/EA PMs!), seemed viable until transatlantic drama prematurely ended its run. Warzone is also a product of transatlantic drama but with a far more seductive pitch of simply translating existing HD Warzone content to mobile. This allows Activision (ATVI) to amortize cost across a wider swath of players and finally achieve Kotick’s Call of Duty “platform” fantasy. On the other hand, this comes as a double-edged sword, as the HD version constrains monetization mechanics and content of the mobile SKU. There’s also the overlooked question of the SKU’s target audience: HD gamers looking for COD on the go? That can’t be a giant Venn diagram, considering CODM is a superior product on nearly every front, including technically.

Roblox and Fortnite Mobile underscore the inevitable rise of virtual analog sticks in mobile gaming. As the largest Western shooter IP, COD paired with Chinese content quantity and monetization, created a category on its own. This should be an area of Western domination. However, Valorent and Battlefield’s inability to ship is a blemish on Western development. Still, riches await if publishers are willing to subdue their HD instincts and respect the mobile paradigm. Next up: Rainbow Six: Seige Mobile.

1. It's Distribution or Bust, and No One Cares

IDFA is gone; how is your game designed to maximize distribution? Game designers, not marketing managers, need to start building for sustained user acquisition. Adding an influencer as a cosmetic is a nice stunt, but it's not fundamental to the game loop. I was surprised at the dearth of these questions relative to the AI focus. If the Chinese refuse to divest TikTok, this question will become more pressing over the next year.

2. Web3 Evolved from Shipping No Product to Previewing Bad Product

Shrapnel, the web3 FPS developed by devs with actual game industry experience, was playable. It was...bad. The official launch is targeted for 2025, so there's time to fix the issues, but this is another hopeful breakthrough that's become doubtful. It's still progress for web3, which usually lacks products of any kind. Crypto Unicorns, another game with actual product, opted for a giant unicorn display statue instead.

3. There are Four GDCs, and I'm Doubtful They're Merging

There are separate GDC experiences for VCs, mobile, HD, and web3. Each has its own parties, personal, and drama. I keep thinking these worlds will merge, but it's been at a sloth-like speed. The continued wave of dual-SKUs (hello, CoD Warzone, and Delta Force) shows product signs, but the actual circles employees run in remain distinct. I'm keen to see the final tally for this year's attendance, but I sense a new record.

4. A Lot is Riding On UEFN

No one cares about Roblox because the 70%+ tax means studios will never scale. The top of the market, Adopt Me, does $50M in revenue, and it's all downhill from there. UEFN allows an off-ramp to distribution on EGS or Steam and a much more favorable cut. Beyond anything else, UEFN will enable creators to prove product-market fit when raising another round, but Epic needs incredible roadmap execution. Key items have already slipped.

More of the event here, and my talk should be live soon…

1. Cultural Norms Matter

There’s far less “sharing” in Turkey than in Finland or Sweden. Anyone with a Finnish ID number seems to have access to Supercell’s dashboards! Turkey has largely avoided non-competes, and it must stay that way. King single-handedly has hamstrung Swedish mobile game development with its notorious three-month non-competes, which, after a three-month notice period, amount to six months for an employee to start elsewhere. France is dealing with a similar issue, and it’s to their detriment. Much of Silicon Valley’s success is due to the ban on non-competes, something that’s gone over the heads of subsidy-happy governments.

2. Agglomeration Effects are Real

Economists are fond of the power of location to build industries through virtuous success cycles and knowledge spillovers between firms. Turkish development feels like 90% puzzle or hypercasual, and it’s no wonder, given Peak, Gram, Dream, and Rollic’s success. But Turkey needs to do more to hedge on genre category, as hypercasual’s story teaches us. Falling KPIs from Spyke Games’ Tile Busters are a warning sign.

3. Politics Threaten the Entire Ecosystem

Upcoming elections on March 31st made the tension palatable. Stable institutions and respect for the rule of law are essential to economic prosperity, and I wonder how much more Turkey could have grown over this period with stable institutions. Nothing encourages capital flight more than political instability, except, well, inflation, and that’s not looking good either.

4. AI is like Teenage Sex: everyone talks about it, nobody really knows how to do it, everyone thinks everyone else is doing it, so everyone claims they are doing it…

AI feels obvious, but I’ve only seen it used for concept art, which is hardly an essential part of the funnel. Its use is predicated on integrating into engineering pipelines and throughout different parts of the stack. Tomi Huttula‘s Cosmic Lounge is the most open and aggressive use of technology; everyone should see that talk (online soon!). Starting from the ground up with AI feels like the right play, and it’ll take longer than expected.

5. Everyone’s Accepted Industry Contraction, but No One Has Structural Answers

Mobile and HD are shrinking, and it’s not clear we’ve hit bottom. While there was a focus on new strategies, the elephant in the room continues to be shrinking time played. Time played needs to explode to match COVID numbers, and no one has a path to get there. I wrote about this in Gaming’s Best Days Are Behind It, and will continue too. As the EU seems to have thrown in the towel, primarily concerned about regulating USB cords, I’m becoming increasingly convinced that our saving grace is continued global economic growth, mainly from the U.S. and Tier 3 regions.

Matthew Ball’s excellent piece, The Tremendous Yet Troubled State of Gaming in 2024, unpacks a paradox: gaming seems prosperous yet is marred by unprecedented industry layoffs. Ball’s findings reveal a simple reality: playtime is down over 20% (3.5 hours per week) from its COVID peak. This decline, further seen in declining inflation-adjusted spending, cements gaming as an entertainment product with a capped growth potential. Without new strategies to reclaim time to share, gaming’s best days are behind it.

Despite COVID hangovers, Ball remains optimistic about gaming’s prospects, citing opportunities in streaming, in-game advertising, and transmedia. Apparently, the mid-00s breakthrough of OnLive, Burnout Paradise, and Uwe Boll wasn’t enough. But other growth suggestions include “‘UGC platform developers’ (i.e. Roblox developers) shifting to standalone titles.” This resulted in Steam’s top-charter, Lethal Company, the product of a former 21-year-old Roblox developer. UGC needs to play a central role in returning gaming to COVID highs.

Game design is a technology; like other knowledge-based technologies, it improves the more people work on it. Ideas have sex and cross-pollinate into emergent breakthroughs. User-generated content (UGC) results in idea orgies, and it’s given gaming its top genres: MOBA and battle royale. However, few franchises have added in-game UGC creations outside the party genre (Stumble Guys, Eggy Party, etc), and even fewer have viable monetization strategies. Unfortunately, there’s little indication this is changing, and instead, the Youtube-like models of Unreal Engine for Fortnite (UEFN) and Roblox dominate.

Two in three U.S. kids play Roblox (!), a platform that trades at an astonishing 57 times EBITDA. UEFN is growing and maintains an aggressive (and public) roadmap of features. Yet, cracks emerged underneath the surface: Roblox’s growth is from tier-three regions, hinting at future slowdowns, while first-party content dominates UEFN. As Ball astutely observes, technology shapes content, and UEFN and Roblox remain 3D joystick experiences. It’s hard to imagine them fostering experiences like Wordle. For gaming to flourish, it must embrace the variety of platforms already in people’s lives.

One of gaming’s crowning achievements is its ability to convert Tetris-like pieces into experiences using computers, phones, watches, airplane TVs, and Peleton bikes. However, its contortions haven’t always yielded long-term success. Despite selling over 200M Wii and Switch units, Nintendo couldn’t convince its newly diverse audience to play anything beyond Wii Sports and Mario Kart. Smartwatches were a dead end, and fitness devices resulted in closets of Pelotons piled on Wii Fit boards. Social media is gaming’s biggest contortion opportunity, but failed gaming divisions at Snapchat, Facebook, and TikTok don’t bode well.

If gaming can’t redefine its post-COVID future, its best days are behind it.

Phil Spencer, Microsoft Gaming CEO, named Johanna Faries the new president of Blizzard, closing a pivotal moment in the ATVI-Microsoft merger. Faries faces the challenge of revitalizing Blizzard amidst declining revenue and engagement across its franchises. The situation is dire enough that Faries inherits a needed rebirth of Blizzard. The Blizzard of old is dead, and several causal agents are responsible.

The Ship of Theseus is a thought experiment about when a ship remains the same after trading its planks one by one with another ship. The same experiment applies to Blizzard, who has replaced a good deal of its workforce after they left for new start-ups. Over the last decade, venture capital siphoned copious amounts of Blizzard employees into new start-ups, aided by California’s law banning non-competes. The VC-backed combination of Dreamhaven (fmr. Blizzard founder), Second Dinner (fmr. Hearthstone team), Magic Soup Games (two fmr. Bizzard Presidents), and Frost Gaint (fmr. Starcraft team) likely have more years of Blizzard experience than Blizzard itself. While sad for Blizzard, like DICE and Bioware departures before it, veterans start new studios, growing the area into a hotbed of talent. Venture capital has made it easy for frustrated employees to take corporate exits.

Blizzard hasn’t been able to use these exits to reset processes and spur live-service growth. Overwatch is sputtering, and mobile growth remains muted. Diablo IV successfully launched but won’t move the needle on long-run monetization, opting for traditional content expansions. Arklight Rumble proved Blizzard could ship quality products and mobile-first monetization, but the lack of mobile UA remains concerning.

Blizzard’s most damning indictment is the over 50% growth in Call of Duty over the same period. Kotick discovered Call of Duty’s value extended beyond developer Infinity Ward, rotating three studios on the franchise in 3-year dev cycles to hit cascading yearly releases. But much less than “satiate” interest, annual Call of Duty launches made players hungrier for content. Blizzard never exploited economies of scale to maximize franchise value.

While both venture capital and an inability to grow the franchise’s lifetime played a role in pulling the trigger, Blizzard has the opportunity to rebirth itself as a franchise steward of its original IP.

Blizzard’s rebirth should focus on growing existing IP by leveraging studios and headcount outside the Irvine office. Diablo Immortal’s $1B success proves Blizzard PMs can guide eastern development. They can do the same for the Chinese autochess market: a Warcraft alternative to Riot’s top-10 Chinese grossing Team Fight Tactics. The same is true of Overwatch if Valorent Mobile proves to be worth anything. And World of Warcraft Mobile has long been rumored, and with Eastern monetization, it stands a chance at challenging Korean and Chinese MMO titans.

Ah, web3: a past marked by rug pulls, North Korean heists, and prices more volatile than Gamestop shares. However, for all its misgivings, audiences still capitalize web3, with prices far from zero. Parallel, a new web3 CCG hit a half-billion market cap considering token and NFT prices, and yes, it’s pre-launch. Despite flaws, Parallel looks cool, and games like Sipher are turning heads, while Sorare is already established. Chris Heatherly is brewing something with Mystery Society, giving ‘Among Us live-service’ a jolt. But for all its hopes, web3 has been blocked by distribution and confined to the browser. Epic Game Store, a trusted brand, now carries web3 games, even with an Adults Only rating. Finally, Web3 is playable on mobile, with the compromise of giving Apple 30% and tacking on an equivalent user-facing tax. All this festers into 2024 as the year web3 ships a chart-topping product; it can’t keep failing upward. Right?

As Eric Kress reminds us, new platforms and genres represent new opportunities. Disequilibrium persists before market leaders establish themselves. Look no further than Palworld, an early Survivor-Crafting genre entry, where a key hire came from a part-time convenience store worker. New platforms grow audiences and, ultimately, the industry. If you want gaming to win, you want web3 to win because that means games win.

Despite slow infrastructure advancements (hello, Forte) and high transaction costs, platforms like Immutable‘ IMX are emerging default solutions. Remember, Steam wasn’t easy to use on launch, but its exclusive Half-Life distribution made the ‘juice worth the squeeze.’ Web3 founders tell me they want to ‘onboard the first billion users to crypto,’ but this means designing something worth surviving the ‘web3’ part for; web3 must make its content juice worth the onboarding squeeze. This challenge extends to tokenomics, where simplicity should be a priority over complex economic models.

Web3 still needs to find an ‘off-the-shelf’ tokenomics meta. Teams shouldn’t need economic manifestos to launch web3 games, but virtual goods get complicated when they carry a real-world price. It’s been hard to figure out tokens: a hard currency, an annuity, and a growing stock all-in-one, but if it needs to be anything, it’s yield for owners. Firms must integrate concise ways they ‘share the means of production’ with stakeholders in exchange for capital. Or avoid the ordeal entirely and only permit account selling. At risk of self-harm, games shouldn’t need a game economist to succeed.

Web3’s journey, led by inexperienced pioneers, remains challenging. Distribution is no longer a barrier, but content and tokenomics remain immature. Success lies in creating concise tokenomics and delivering compelling content worth the web’s gripes. Web3 likes to remind skeptics that ‘it’s so early,’ which is true, but 2024 must be the year it gets a bit older.

After four days, Palworld, an early-access survivor game described as “Pokemon with guns,” grossed ~$108M over three days en route to the 2nd highest concurrent player count for a paid title on Steam. It’s an outstanding success for Pocket Pair, a 50-person Japanese team with $7m raised. But like early access hits Valhiem, Battlebit, and Temtem before it, an inability to quickly deploy capital hampers compounding return. Instead of these titles having their best days to look forward to, they’re in the rear-view mirror, the equivalent of peaking in high school, but this time, you get millions in cash for the ordeal. Teams should follow the Kikta Studio model: prove a trajectory and let someone else build the bow. When founders sell games, they retain control of their destiny and find themselves newly rich with modelable capital, allowing games to grow into their best selves.

Despite explosive early access launches, Valhiem and Battlebit grew back to as little as 55% of their early access peaks, while Temtem managed 10,000 players for global launch. 4 to 50 studio headcount support the games, compared to 400-800 devoted to Candy Crush alone, a count common among top live service titles, including Apex (4+ support studios) and Rainbow 6: Siege (Ubi Montreal: 4,000 employees). Even with lucrative venture deals, capital takes time to be deployed, given early studio maturity. It’s much easier to grow 100 headcount of 1,000 employees than to grow 10 headcount of 100 employees. At the very least, full-title sales outweigh marginal costs, allowing the title to become a “lifestyle” business that founders indefinitely chip away at. Likewise, No Man’s Sky delivered an impressive roadmap for a single-player title, enough to support a moderate margin for 50 headcount studio.

We don’t yet understand the headcount to support a live-service survival crafting game. It’s unclear how “productive” content is in buying units of retention from players. Millions of players will see the inaccessibility of choosing a resetting server for 30 days or more in 2024 – there’s a reason top Steam survival games Rust and Ark have yet to scale to consoles meaningfully.

Blizzard should be all smiles, given their announced genre entry. Survival-crafting accruing breakout after breakout; where there’s smoke, there’s fire. Palworld seemed to discover the trojan horse: Pokémon. Boy, players really want a Pokémon MMO, as Temtem and now Palworld remind us.

However, it’s a conversion funnel, and Palworld’s most valuable contribution is wild marketing assets at the top. Focus groups must have gone crazy when a researcher put an AK-47 in Pikachu’s hand. The trailers didn’t disappoint either, with one memorable scene demoing Pichu’s assembling M-16 rifles. The next part of that funnel is the difficulty of converting early hype into lasting retention. Palworld will need more capital to get there.

Imsominaic leaks reveal Playstation lost 3M MAU over the last three years. At a time when gaming is at its grandest heights, consoles have failed to seize the moment. Sony and Microsoft have forgotten their role as platforms, purchasing dead-end franchises instead of fostering third-party innovation. Meanwhile, Steam gained over 90M users, a ~35% increase over the same period, with Valve founder Gabe Newell downing New Zealand pies as third-party development rakes in billions as Valve barely lifts a finger.

Sony’s lavish string of single-player blockbuster titles like God of War, Spider-Man, and Zero Horizon Dawn certainly have positive ROI, but another ten years of $300M roulette spins (budgets also leaked) hardly feels digestible. It’s a near miracle these titles performed critically well as they have, illustrating why Sony executives rush toward live-service investments. For its part, Microsoft squandered Gears of War and Halo while Minecraft lingers in the shadows of Fortnite Creative and Roblox. Console is no longer where titles like Geometry Wars, Rock Band, and Tony Hawk found their home. Instead, Steam has become the beacon for innovation. 

Vampire Survivors, Sons of the Forest (the top-selling Steam game of 2023), and Valheim all started in Steam Early Access. Sony and Microsoft are so focused on stacking their bench with billion-dollar acquisitions that they’ve forgotten they’re a platform, with the lack of modding and robust early access programs telling. Bill Gates reminds us, “A platform is when the economic value of everybody that uses it, exceeds the value of the company that creates it.” Nintendo, Sony, and Microsoft likely flunk this test. 

Steam’s success is so domineering major publishers crawled back to Steam, with Apex’s Steam launch netting Valve hundreds of millions since its launch. Even Blizzard was desperate enough to release Overwatch 2 on Steam, while Microsoft gave up and made Halo a Steam product in 2019, with Sony scoring another 3M in Zero Horizon unit sales (also leaked). 

The picture becomes a breaker with the advent of Fortnite Creative and Roblox. While both have console homes, they shift distribution and discovery to UGC platforms, weakening console control. They both play on mobile, and more titles find the dual-SKU strategy viable. Genshin’s mobile gross exceeds console, with Honkai Star Rail arriving on PlayStation six months after its mobile release.

At the very least, Microsoft dies on a semi-coherent total addressable market growth strategy in the cloud and Game Pass. The best Sony can muster is a dilapidated PSP VPN revival and a VR headset reminiscent of Virtual Boy. Sony’s mismanaged live-service approach isn’t enough to grow consoles.

To regain their lost crown, both giants must embrace their roots as the first and best platforms for discovery and development, a role they’ve neglected for over a decade.

MMOs and most web3 titles suffer from a fundamental design flaw: capital depreciation (or lack thereof). The result is a Zimbabwe-style hyperinflation economy so prevalent Redditors plead to find an MMO that hasn’t suffered. Even EVE Online, with its own Head Economist  ⁷. suffered inflation bouts. It’s a looming threat to the web3 space, only mitigated by the fact that web3 games haven’t survived long enough to grapple with it. Luckily, there’s an answer that combines the best solutions from MMOs with the “permanence” soul of web3: a staircase tax.

Reddit Find Friedman in All the Unexpected Places

Gold Farmers Always Win

As Milton Friedman reminds us, “Inflation is everywhere, and always a monetary phenomenon,” a principle that extends to MMOs. When players generate or mine assets, holding all else constant, the supply curve moves outward, and price falls. With enough time, price converges to $0. Axie Infinity saw its NFTs’ reproduction rate exceed one, leading to a surplus that drove prices down.

The Supply Problem

Real-world goods suffer a similar fate but benefit from capital depreciation. Wearing a T-shirt repeatedly degrades the quality and changes the nature of the good; it becomes stained, stretched, and shrunk. Holding all else constant, producers maintain price equilibrium by equating the supply of the good with the rate of capital depreciation. The circulating supply of T-shirts remains constant if five wear out and five are sold yearly. This isn’t the case with digital goods – a Taylor Swift song .mp3 is the same on the first listen as the millionth.

MMOs have historically dealt with the problem of inflating supply by introducing ‘durability,’ where items wear out and become unusable after repeated use. For example, in World of Warcraft, players slay monsters, collecting items and currency. Since these items are tradeable, each time a monster is slayed, the global supply of the items the monster drops increases. Supply is, therefore, a function of drop rates, monster respawn time, and the quantity of players. Players must sink currency to repair the item, which shifts the supply curve inward. While durability balances the economy spreadsheets, players may find constant repairs as convenient as frequent stops at the gas station. A staircase tax avoids item consumability while removing the item from market circulation.

Sadly, No Premium Gas Available

A staircase proposes an escalating tax rate on a given item at every n+x trade. For example, an item might start at a 5% tax rate, but after the second trade, the tax rises to 10%; after the 3rd trade, it rises to 15%…, and so forth, until it reaches 100%. At a 100% tax rate, the item is effectively soulbound. The tax could take on several designs: increasing at a fixed or set rate or changing the number of trades that trigger a tax increase.

Example Staircase Tax

The item becomes less likely to be traded at each trade, gradually removing it from circulation while maintaining utility. To maintain price stability, these need to be set in such a way as to counteract the rate of new supply.

A staircase tax isn’t only for web3; it’s something any game with tradeability should consider. By escalating taxes on successive trades, this mechanism simulates the real-world principle of capital depreciation without resorting to the unpopularity of item durability or the design quagmire of destructibility. More work is needed to determine the “optimal” set of staircase variables, and if web3 lives long enough, it might be able to tell the tale.

At 350 million monthly active users (MAU), Roblox boasts a user base rivaling the combined size of Steam, Xbox Live, and PlayStation Network. However, this impressive engagement comes hand in hand with an exorbitant 70-80% platform tax. In any case, Roblox is too big to ignore. Its bigness forces Roblox and non-Roblox developers to grapple with a pivotal choice: to depart from or align with the platform. Roblox departers jeopardize the very audience responsible for their success, while Roblox-joiners scoff at the prospect of surrendering nearly 80% of their revenue to the platform. But there’s a third door, a Micheal Scott win-win-win, where players, Roblox, and developers win. It’s time for Roblox SKUs.

Talent is increasingly grappling with platforms’ growing ability to control supply via their massive scale of demand. It’s a classic story of Ben Thompson’s Aggregation Theory. Taylor Swift briefly left Spotify for undervaluing her catalog, only to come back within the year. No concessions were reported. Microsoft tried paying top Twitch steamer Ninja $50M to join their Mixer platform, only for it to explode within the year; Ninja failed to bring viewership. And, of course, Marvel famously paid their early actors hundred of thousands of dollars, not millions. In each case, the platform helped to create talent, not vice versa. But there’s a curious effect at the tippy top of the talent spectrum; some creators have never been worth more.

Only recently, TV showrunners like Shonda Rhimes were raking in $300M, and Joe Rogan scored a cool $200M on Spotify. In each case, the platform and the talent evaluate the joint and separate production functions. Clearly, Shonda Rhimes can produce more with Netflix’s near-limitless resources on top of instant global distribution to millions. While Shona may not capture all of the joint production gains, she captures enough compared to going independent. But is low-capital podcasting the same? Andrew Wilkinson argued that Spotify ripped off Joe Rogan despite having just signed a nine-figure exclusive podcasting deal. Rogan traded upside for security and lost hundreds of millions in expected revenue. Rogan had been sitting on oil he refused to extract – a little arm grease would have let him own the listener relationship but, more importantly, capture LTV gains. Swift, too, had to reckon with reaching 500M fewer fans after removing her catalog from Spotify. In each situation, talent had to weigh their value against the platform: does the talent make the platform, or does the platform make the talent? Roblox developers aren’t asking this question enough.

Wilkinson’s Rogan Pro/Con List

Roblox developers grapple with an economic structure designed to cut off growth; for every $10 spent, developers only see ~$2.25. However, unlike a particular App Store, the platform’s benefits are clear and vast: Roblox covers server costs, handles payments and customer support, provides an engine, and, most importantly, an audience of 300M MAU. Once you consider the opportunity cost of Steam / the App Store at a 30% platform fee on top of another 25% in server costs, the Roblox tax shrinks to a more reasonable ~20-30% toll tax. 

As David Taylor’s excellent piece models, the top Roblox game, Adopt Me, annually earns the equivalent of one month of a top mobile game. This is around $300 million for Adopt Me, but developer Uplift Games sees just $77 million after Roblox’s cut. Adopt Me likely lands at ~$0.32 monthly ARPU on 60-80M MAU. If they maintained similar monetization, Adopt Me could lose more than 20M MAU or 25% of its audience and break even on Steam or the mobile platforms. At some point, the platform tax causes enough pain to consider alternatives.

Adopt Me’s Gamble

Axing Roblox is too risky, but shipping additional SKUs on mobile and HD platforms is a more evident and safe play. At $70m+ in annual revenue, a port should cost no more than 10% of revenue. Roblox developers own their IP, so they have legal rights to do this. In some ways, this strategy mirrors the rise of webstores from mobile publishers to save on platform fees. This argument also works in reverse. Not only should Adopt Me launch a mobile SKU, but some developers should consider a Roblox SKU. Mainly Steam’s new wunderkind, Battlebit.

In February of last year, va ideo of a high-fidelity Roblox shooter, Frontlines, shocked players and onlookers; Roblox’s engine could do more than squares and circles. 

But despite the fanfare, Frontlines maxed out at 6,000 PSU. Meanwhile, on Steam, voxel-based first-person shooter Battlebit rocketed up the Steam best-seller charts as a $15 premium game. From a team of only four, Battlebit maintains 60,000 PSU. The acquisition and monetization profiles could not be more different; Battlebit counts 2M in lifetime sales at $15 LTV, compared to Frontlines’ 42 million visits at a factional pennies LTV. While Battlebit will indeed decline, the game has grossed at least $30M, likely far beyond Frontline’s lifetime revenue. As it turns out, HD experiences don’t work inside Roblox, but Roblox experiences might work in HD. Q234Battlebit follows this model, as does Unturned, another Voxel-based Steam hit from developers who started prototyping in Roblox.

Roblox should make it easy to bring Battlebit and Unturned from Unity to the platform. While Roblox’s proprietary engine is as much of a technological as it is a corporate requirement, Unity plug-in or “port layer” brides some of the gap. If the platform can run high-fidelity titles like Frontlines, it can undoubtedly support more diverse art styles with broader appeal.

As Roblox continues to grow, the platform faces pressure to either increase revenue share with top developers or risk losing its biggest hits to greener pastures. Conversely, Roblox faces an opportunity to bring hits like Battlebit and Unturned onto the platform. With billions in revenue on the line, Roblox must find a way to balance platform economics with creator incentives before restless developers leave and take their valuable intellectual property with them.

The free-to-play (F2P) revolution will not be televised, but it will be downloaded. Naraka: Bladepoint, a consistent top-10 Steam performer, transitions free-to-play this week. The new tally counts seven of the top ten Steam games as F2P, with four starting life as a premium before switching to F2P. Major Western publishers composed free-to-play holdouts, but that’s crumbled in the last five years as Halo, Overwatch, Rocket League, and Destiny switched to free-to-play. The question for live-service franchises is now not if but when.

What was a sprint on mobile has been a decades-marathon transition in Western HD (PC & Console). Executives at major publishing houses have long harbored fears about relinquishing the consistent revenue from premium games. The apprehension traces its roots to mid-2010 wounds when F2P iterations of Battlefield (Heroes), Call of Duty (Play4Free), and FIFA (World) failed to take off. But the reason for mobile’s success mirrored HD’s failures; “it’s the distribution stupid.” In 2013, digital and physical sales of HD games were equally prevalent, a 50/50 split. By 2020, the digital market share shot up to 90%, leaving physical sales to a mere 10%. The limited extent of the digital HD market capped free-play success, while Steam only opened its doors to F2P in 2011, and full two years after Apple introduced the IAP,

In an economics fable as old as time, the relentless pressure of competition inevitably forces the hand of industry giants. As new firms enter the market, they drive prices down. Titles like Fortnite and Apex Legends forced PUBG’s F2P transition, a move that’s proven sustainable. Publishers have even woken up the idea of F2P as a sword rather than a shield.

Old premium behemoths are facing new F2P challengers. Ubisoft’s XDefiant has emerged as a bold contender against Call of Duty, while miHoYo devours Nintendo’s F2P apathy. Run down the Steam top seller list; a premium price smells like a F2P invitation. Are Rust and Dead by Daylight next to enlist in the F2P revolution?

Apple’s VisionPro rejects gaming, but gaming doesn’t have to care. The game’s industry position has grown so extensive and favorable that it can survive and thrive on any platform, even when initially overlooked.

As web3 VCs are jolly to remind everyone, gaming typically serves as the initial use-case for novel technologies. This was undoubtedly Meta’s conclusion leading to purchases of 17 VR developers, including the developer of Beat Saber, the most popular VR game to date. But a first-use case or not, gaming envelopes most platforms it touches to the point “Will it run Doom?” became a meme. We’ll see the same adaptation if VisionPro succeeds or fails.

Apple’s core competencies took center stage during the WWDC presentation, irrespective of VisionPro’s future success or failure. In the blink of an eye, they rebranded “metaverse” to “spatial computing.” Apple proved they guide the narrative around their products and their ability to avoid the kind of PR missteps that have plagued competitors. Subtly, their crowning achievement of the day was avoiding looking stupid. So far, no picture of an Apple executive waring the Vision Pro has been taken or released to the press – that’s hardly an accident. Compare this to Zuckerberg’s infamous aisle shot, and it’s evident that “reading the room” is a marketing culture asset. The rush to release Quest news to avoid being buried by VisionPro is another reminder of the fear their marketing team strikes in the hearts and souls of companies everywhere. If Apple decides to enter a category, there’s a good chance they’ll come to dominate it.

Apple’s never particularly cared about gaming: no updates or major game announcements on Apple Arcade, bar writing checks for DOA VisionPro Arcade ports during WWDC. I hope game developers bring their microscopes because they’ll need to read the DAU charts of a $3,500 device. Still, Apple announced a macOS port of a game From a Couple of Years Ago. So at least there’s that. But while Apple may never care about gaming, even on VisionPro, they can’t escape it. A staggering 70% of App Store revenue now comes from games. As Eric Seufert argued, it’s the Games store, not the App store.

Games seemingly address VisionPro’s most glaring issue: its unclear use case. The early reviews have indicated a flawless experience, but how the technology solves novel consumer problems is unclear. Steve Jobs started the original iPhone presentation by framing the product as a response to the lack of customized digital keyboards. Before the iPhone, physical mobile keyboards were clumsy experiences on tiny screens with bulky hardware. Tim Cook frames VisionPro as a new form of computer interaction closer to the next iteration of the mouse and keyboard. If AR-based visualization, eye tracking, and finger-based flicking improve productivity for workers by only a couple of percentage points, VisionPro could follow the path of the original PC: starting as an enterprise product as costs fall to widespread consumer levels. If a worker generates $100,000 a year in revenue and improves output by 3.5%, VisionPro’s price is justified. 

While Apple’s VisionPro strategy remains a question mark, Meta’s approach to VR gaming has proven a struggle. A live service hit or even moderate success eludes the platform. One Epic Games Store revelation was to package live-service games as the “free game of the week,” including the most successful live-service game of the decade, Grand Theft Auto. Players frequently launch these titles, equating to thousands of platform interactions over a player’s lifetime. Generating a series of one-time hits is far more challenging and uncertain than single-title liveops. 

Not only does the platform forego live service, but it also imposes steep fixed costs on VR players. Simply getting mobile users to flip the phone landscape is challenging enough, much less repeated use of a cumbersome 4-hour battery headset. It’s no wonder VR’s most significant growth moment came when it was untethered from the PC into a standalone device. 

Despite Apple’s dismissive stance on gaming, the medium has proven its adaptability and resilience time and again. Gaming has survived and thrived through every new interface, whether it’s the joystick, analog pad, or Apple TV remotes. So, while Apple may not recognize and cherish the medium, games find a way to flourish and dominate their platforms.

Hypercasual solved a novel problem for *all* mobile titles: marketing. Like web3, hypercasual always seems to be caught in its drama with recurring reports of death or life on top of wash trading driving genre GDP. But beneath all the bullshit, hypercasual discovered something profound: how to get BILLIONS of players to download games at sub $1 CPIs. In 2022, hypercasual accounted for 29% of ALL app store downloads and cleared over 1 BILLION monthly downloads. The numbers represent real player action – they see something in hypercasual ads they don’t see in others. On a rough click-through basis, players might be 5 to 10 times more likely to tap on hypercasual ads then midcore ads.

Hypercasual’s IP playbook hammers on reducing cognitive complexity; gameplay and art are straightforward to digest. Navigating a medieval fantasy game teeming with knights, dragons, and wizards requires more cognitive resources than understanding blank-faced stick figures that run. Hypercasual titles like Mob Control or Aquadpark immediately drop players into a simple, satisfying, active input gameplay. 4x’s gameplay might take 30 days to emerge, while squadRPG players are saddled with three days of reward collection before the actual game beings.

For all of hypercasual’s success, the walls are closing in. Apple’s IDFA deprecation murdered ad revenue, leaving hypercasual’s primary source of revenue high and dry. The shift to “hybridcasual” may sound like a marketing ploy, but it’s a compelling and distinct thesis: acquire players at minimal CPIs using the hypercasual playbook, then retain and monetize them with proven meta models from Clash Royale, Coin Master, and others. The only challenge is that midcore is burning the other end of the candle.

With the surge in CPIs, midcore marketing has leaned into avoiding gameplay in ads or extending “fake ads” into real D0-D3 gameplay. I never thought I’d be matching-3 in a squadRPG like AFK Arena or drawing circles to save a dog in an auto battler like Mighty Party. 4x titles are terrified of showcasing actual gameplay ads, instead focusing on nebulous concepts like “power level” delivered by no-name influencers. Hypercausal, conversely, demands ads show gameplay; there’s nothing to hide when the core is readily understandable within seconds.

The stage is set: will midcore become hybridcasual faster than hypercasual can become hybridcasual?

Supply-chain economics is as much a game design responsibility as it is production. Core design influences the production gap between Blizzard’s Overwatch heroes and, say, Ubisoft’s Rainbow Six Siege (R6:Siege) Operators. Ubisoft has shipped an average of .51 new Operators per month since launch compared to Blizzard’s .21 new heroes, more than double the pace. Blizzard will pack far more lore, cosmetics, and unique gameplay into each hero at the time of release – it’s part of the Blizzard “quality bar,” but maybe heroes should release faster and with fewer “features.” In faster character designs, weapons are sometimes disintermediated from character units – Apex and Valorent release weapons separately. Part of player empathy is servicing players’ content needs, with many legacy AAA studios refusing to tackle the challenge. But it’s at their peril; players ultimately choose the optimal mix of quality and quantity, not design directors.

Liveops’ responsibility is to reduce the constraint of the lowest common denominator in the supply chain. If melding a steel car frame takes three months but painting it takes six months, a car will take at least six months to produce. Game balance is commonly cited as a limiting factor for expanding supply, suggesting game designers play an integral role in optimizing production. Magic The Gathering ships over 1,000 new cards a year, a feat granted by their ability to retire old sets; only the most recent three are available for ranked play at any one time. Those are design decisions, and game designers are front and center for these choices.

The designer must also decipher the game’s retentive content. And, preferably, make more of it. For instance, the last decade has shown a weak preference for new maps in FPS games. Maps were the centerpiece of any DLC entitlement pass (read: Battlefield Premium) but now feel dated, with the nearly one map needed to maintain battle royales providing an extreme contrast. Designers would claim maps encourage “mastery of the terrain,” but the empirical evidence rarely backs the theory; small one-time retention bumps are common for map packs. The hundreds of characters in MOBAs and squadRPGs are no accident; it’s a supply-side response to demand.

There’s a unique bias toward quality as it satisfies a sense of self-pride; quality is integrity, while quantity is gluttony. But design that sustains is design that survives; not everyone wants to starve themselves daily for a Michelin dinner. We all need snacks. For studios that adapt, like Respawn, the reward is immense; Apex has hit all-time highs on top of year-over-year growth for the third straight year.

The reverse-Blizzard thesis has arrived. Firms like Blizzard, Supercell, Valve, and even Apple thrived on popularizing but not inventing mechanics and genres. Blizzard’s next fresh franchise, a survival crafting game, draws heavily from predecessors like Ark: Survival Evolved and Rust. Those titles failed to scale to mainstream adoption; if players need a server browser for play, it’s not ready for the mainstream. But both Supercell and Blizzard have struggled to scale their studios to meet modern content demands, and that’s left them open for a sort of reverse thesis: drop production values and win on the supply side. 

No game exemplifies this better than Path of Exile. Blizzard’s failure to establish and maintain a live-service team for Diablo III gave birth to Path of Exile, a game that recently reached an all-time high of 1.5 million DAU and is nearing a $100M run rate. To be clear: Diablo IV will far outgross Path of Exile, but it’s unclear if Diablo will win in the long run. And lest we not forget, Grinding Gear’s cost profile is likely half of Diablo’s team with a marketing budget that looks like a rounding error to Activision executives. While hero shooters outside of Overwatch (OW) and Rainbox 6: Seige died with Battleborn, Blizzard’s cancelation of OW’s PVE mode and the two-year content gap between OW 1 and 2 left blood in the water. Will upstart developers bite?

Netflix CEO Reed Hasting infamously claimed, “The goal is to become HBO faster than HBO can become us.” A similar paradigm is playing out between AAA developers (HBO) facing low-production values but high-content supply developers (Netflix). But unlike linear entertainment, AAA developers have been sharpening a secret weapon to swat away nibblers: UGC.

Fall Guys become an overnight success thanks to its fresh, accessible design and widespread PS+ distribution at launch. Never intended to be a live service, Mediatonic, the game’s developer, was compelled to build one in-flight. Over the last two years, the plane never reached sufficient altitude, and the reverse-Blizzard thesis reared its head in Stumble Guys. Scopley swooped in and ramped content, with rumors swirling the game is making some $30-$40m+ per month in ads alone. And don’t forget, on top of mobile; Stumble Guys maintains a cool 100k DAU on Steam. Mediatonic’s recently announced UGC mode for Fall Guys is their attempt to stem the tide of supply-side competition.

As Battlefield Portal, Halo’s repeatedly delayed Forge, and Fortnite Creative launched, these AAA developers must look to their next challenge: incorporating high-quality UGC content into the core experience. To ensure the success of the AAA UGC approach, UGC must not be viewed as a separate or “branded” experience but should be indistinguishable from first-party content. As always, Valve seems to hold the answers that the rest of the industry often overlooks. Perhaps this time, the industry will pay attention.

Chinese games have gone global. This week’s launch of Honkai: Star Rail and Starlight 84 marks China’s emergence as a AAA developer for Western audiences.

In the past, top Chinese games such as Honor of Kings, Fantasy Westward Journey, and QQ Speed struggled to meet Western expectations or compete with their Western counterparts. As with our usual economic growth story, manufacturing and outsourcing accumulates valuable human capital and spur agglomeration effects, letting developers move up the supply chain. Original IP development sits atop that chain.

The adoption of mobile gaming and sophisticated virtual “dual stick” controls in the West has aided Chinese success. GenZ is being raised on dual-stick Roblox mobile, and Call of Duty Mobile’s dual-stick success was a revelation. The explosion of anime in the West hasn’t hurt, either. Driven by platforms like Netflix and Crunchyroll, the growing popularity of anime in the West has contributed to a receptive audience for anime-style art. This cultural influence is evident in other Western preferences; I write this across from three ramen shops in Stockholm. Ten years ago, there were none.

Next stop? HD. Naraka: Bladepoint has consistently ranked among the top 10 titles on Steam, although it has yet to gain significant press attention. Meanwhile, the 800-pound gorilla in the room, Black Myth: Wukong, will be released next year.

Although tensions between China and the West may be simmering, including rifts between Netease & Blizzard and Timi & Respawn, the regions have never better served each other’s audience.

The world’s least powerful CEO won’t lead Supercell’s next chapter. Supercell is superscaling, and with it comes bureaucratization and hierarchy. 

CEO Ilkka Paananen believes investing in their proprietary game engine is crucial for the company’s future. One way to scale is by increasing each developer’s productivity. The engine functions like a platform with “out-of-the-box” tools to avoid re-building the wheel for every title. The challenge, however, is that each game and genre has its own “wheel.” Supercell’s most ambitious project, Clash Heroes, opted for Unreal instead of the in-house engine, raising concerns about its effectiveness.

The economics of proprietary engines haven’t worked. Engine development is expensive, with costs amortized over customers. The fewer the customers, the more expensive development. This worsens when combined with new studio locations and remote work. EA’s Frostbite isn’t exactly setting the industry ablaze, and the multi-title property engine graveyard is stuffed to the brim with tombstones.

A stronger central engine also means a stronger central engine team with increasingly prioritized and controlled roadmaps by someone “else.” Will Clash of Clans or Clash Mini get priority for their requested engine feature? Politics is born anew.

Even on the labor side, centralized roles like “Chief Game Officer” have emerged. While we don’t know the responsibilities of this role, we’ve seen Pixar-style creative councils quickly devolve into soft gates, as King employees will vouch.

Supercell’s Next Chapter is a beautiful analysis of its next moves, but it wasn’t written by the “world’s least powerful CEO.” Perhaps changing this title is a part of the Next Chapter.

Snap is doing…ok! It peaked at 10K Steam PSU and 14M mobile downloads. But for all the innovation and press hullabaloo, shouldn’t Snap have higher expectations? After all, Vampire Survivors, from a team of one, wrapped up the year moving 2.5m copies at $5 a pop, with zero budget. And lest we not forget, Snap’s launch marketing war chest is now dryer than an FTX’s office party. Downloads have fallen from ~800k a day in October to ~50k in January; DAU will soon follow suit. It’s too early for a title of Snap’s caliber to see DAU softness!

Second Dinner’s successful unshackling of the mobile design paradigm is also the source of Snap’s struggles. Vertical progression or the ability of players to increase their power level ⁸ guarantees a relationship between time spent and progress. The more a player plays, the more the player will progress. No such connection is guaranteed in horizontally dominated games like Snap. 

Outside of RPGs, players acquiring more and more powerful items in a PvP title amounts to HD design sacrilege. Guns in Call of Duty are at least intended to offer different experiences, but not more or less powerful ones. The same applies to Operators in Rainbow Six Seige or Heroes in Overwatch. Horizontal progression bets the progression house on players “mastering something new,” like the timing of a new Rainbow Six Operator ability or the optimal lineup of an Overwatch team after introducing a new character. CCGs, too, embrace “mastering something new” by giving players an expanding collection of cards to solve novel puzzles. Ben Brode proclaims,

“Part of the fun in Marvel Snap is being put in new situations that you have to puzzle your way out of. If you’re in the same situation over and over again, there’s less puzzling, less thinking on the fly, and less problem-solving. So I think having a lot of variance there is really important.”

Source

But puzzle-solving is a skill with sharp diminishing returns; most players reach puzzle-solving skill equilibrium in weeks, not months, and certainly not years. In a PvP game, the inability to improve at puzzle solving (signified by a 50% win rate) signals the end of progression. The player’s response to a lack of progression is churn.

King’s now-defunct experimentation team operated on a central thesis: if we want to understand how something matters, remove it. Things will break in interesting ways if it’s as essential to the player experience as we’d expect. This is akin to removing the vacuum pump ⁹ from a car – the car may still function, but the vacuum’s absence reveals connections and mechanics not evident at first sight. So what happens if we remove the ability for players to pay? How does the ability to pay matter? How is it wired to progression?

Match3 players spend hard currency in three ways: extra lives (arcade quarters), extra boosters (win probability), and extra moves (win probability). Spending tied to increasing win probability forms the overwhelming share of hard currency sink. Match3 level design operates a sharp oscillating difficulty, so players spend to flatten the troughs and crests.

If players can’t pay, they can’t buy extra boosters or moves, so their win probability declines. And without progression, players churn. We’d likely see the same result for FIFA Ultimate Team or any power progression game after the Netherlands banned loot boxes – player churn.

However, pivoting to vertical progression in Snap is beyond the pale and simply infeasible. Previously, I argued,

[…] cards don’t have RPG mechanics like health or armor to use in micro-dosing vertical power, and the card abilities are too carefully designed to absorb a radical re-balancing of card power between every player. 

Snap’s scaling solution is the same as Magic the Gathering: double down on horizontal progression with new modes like drafting (three birds, one stone?) or Commander-type modes. The enhanced randomness of the mode distributes “luck” more equitably and expands the type of puzzles for players to solve. The top constructed and draft players differ in Magic the Gathering, reflecting different skill sets to succeed. It’s not clear that Snap’s newly announced “Battle Mode” will create a new pool of winners.

Industry rumor has it that Clash Royale had a furiously high reactivation rate. The brutality of an average 50% win rate ensures churn is high. But a computer doesn’t have feelings, so why not have the player beat the computer instead of other players? And so Supercell did the obvious thing, and around three years ago, introduced bots into Clash Royale matchmaking. It solved the 0-sum PvP problem while eliminating matchmaking times; there’s always a bot to play against! Surely retention improved, and players were better off, so what gives?

The 18th-century philosopher Immanuel Kant tells us it’s wrong to lie. Always. Like always. Kant’s categorical imperative details a moral action as one that, if “universalized,” does not result in a “contradiction.” If all people did was a lie, it would become impossible to function as a society; therefore, lying, in any case, is wrong. That’s extreme¹⁰, so consider an addendum: if universalizing an action results in negative externalities, it’s immoral. Consider fake mobile game ads. Fake mobile game ads raise CPIs even in firms that deploy truthful ads. Fake ads create player skepticism where previously there was none – is this ad reflective of the actual gameplay or made up? Players don’t have enough information to know, for example, that Gardenscapes frequently runs fake ads, nor should players have to.

Under the categorical imperative, if every mobile game ad were fake, no player would tap on mobile game ads, and the industry would implode. For this same reason, we should reject unmarked PvP bots as ethical: universalizing PvP bots cheapens the player’s experience at the cost of PvP games everywhere, regardless of if they use bots or not. 

Youtubers produce brazen conspiracy videos like, “100% PROVING THAT CLASH ROYALE HAS BOTS IN LADDER!” because developers fail to inform players if (a) the game uses bots and (b) if the player’s opponent is a bot or actual human. Such a disclosure might harm the player’s experience; as devilish as it sounds, beating up on those with emotions is more satisfying than those without emotions. But the illustration above shows players need to be fooled into thinking they are playing a real human for bots to be particularly effective. Worse still, there’s a clear and present expectation from players that their opponent is an actual human. Sorry developers, “we didn’t say one way or another” isn’t a compelling counterargument.

Clash Royale, Call of Duty Mobile, and Marvel Snap use bots, perhaps extensively. Developers should mark bots as just that – bots. It’s the right thing to do for the player and the right thing to do for the industry. 

I have a lot to say about Marvel Snap; it’s a breath of fresh air in a mobile market that’s smelt stale for too long. Snap dedicates itself to taking advantage of mobile rather than treating it as a tax. However, for all its innovations and wins, I wonder if the Snap team outsmarted itself with a monetization model that will cap its success.

Despite structural issues, Snap isn’t far from snapping a top 20 revenue spot. But beyond the dizzying amount of design innovation and truly outstanding animation work lies a monetization system so soft it makes marshmallows blush. Instead of glaring at Hearthstone or MtG Arena (~$300m – $600M per year), Snap should set its revenue sights higher to Clash Royale’s glory years (~$700M). At ~1.5M DAU, this suggests a $1.27 ARPDAU target. That’s an extremely tall order that requires moving both the numerator and denominator.

With so much to celebrate and discuss, I split covering Marvel Snap into three parts: Game Overview + Game Economy & Monetization, Progression + LiveOps, and UX. Here I cover Game Overview + Game Economy & Monetization.

Context

At the start of 2018, Ben Brode and Hamilton Chu packed up and left Blizzard after more than a decade each to start Second Dinner. After ~4 years of development and ~6 months in soft launch, the former Hearthstone Design Director and Executive Producer finally released Marvel Snap in Oct 2022. The Collectable Card Game (CCG) has already amassed a whopping 6m+ downloads with publisher ByteDance (see: China expansion and license), while the game’s apparent “stacking DAU” is an encouraging sign. While mobile engagement is tricky to track, Steam actuals suggest ~100k DAU against a PSU of ~8k. It won’t surprise me if Snap has already surpassed 1.5M DAUs between PC and mobile platforms. It’s probably a sign of the massive uplift developers get from macOS development. It’s a great start and suggests an extremely bright future for Second Dinner while encompassing everything Blizzard mobile should have been and will likely never become.

Marvel Snap’s Early DAU Stacking (Steam)

Game Overview

Improving lane-based play isn’t the only mechanic Snap weaves throughout its design, drawing from retreat mechanics in Air, Land, and Sea, card design from Marvel Smash Up, and session time learnings from Valve’s recent failure Artifact. Nearly every design, technical, and UX decision is presented to the alter of session time. The adherence to the mobile medium is a heavy pivot from a Hearthstone team whose past mobile releases haven’t exactly won player applause. And despite Ben Brode claiming Marvel Snap represents all he knows about CCGs, it’s also an exemplary understanding of what he appears to know about the mobile and digital medium.

Marvel Smash Up Card Design

The Need, The Need for Speed

After the massive success of Clash Royale, there was a flood of speculation that real-time mobile PVP was the next great mobile design paradigm. However, real-time PVP has barely evolved since Royale’s launch six years ago. And, in the handful of successful cases, it’s come from HD, not mobile franchises, like Fortnite, Call of Duty, and Apex. Even synchronous CCGs like MtG Arena and Hearthstone are hardly considered “mobile-first” titles, and I suspect their revenue splits endorse that view.

The critical difference between Marvel Snap and others is that Snap designers didn’t sit down to make a great CCG; they sat down to make a great mobile digital CCG. The game’s marketing frequently refers to the speed of play, and perhaps even the game’s title is meant to be a play on speed (over in a Snap?). So many choices are geared toward reducing average match time:

• Players play simultaneously rather than in turns
• Match time has an upper limit enforced by:
  • Turn time is capped at 40 seconds (Hearthstone clocks in at 75 seconds)
  • 6-turn limit
  • A “Snap” mechanic which encourages players to retreat before turn six
    • This reduces average match turns to ~ five turns, shaving more match time
• Three rather than seven-card opening hand reduces decision space in early turns and aids speed of play (I’m unsure why they don’t shorten the early turn timers).
• No ability to mulligan

The UX gets credit here too:

• Ultra-fast matchmaking (3 seconds or less)
• Vertical, rather than horizontal presentation (how users naturally hold phones)
• Snappy animation

Some handy APIs and curious players help us compare match parameters across CCGs; here, the comparison inches Snap closer to Clash.

Match Time Parameters

Just Card Design Things

In nearly every card game, physical or digital, players fiddle with their in-hand cards, rearranging them in some particular order. It functions as an itch, the equivalent of a walk-up routine in baseball.

But Marvel Snap prevents players from arranging and rearranging card order in hand. I pondered why only to come across a card whose effect is to enhance the power of your leftmost cards. With the ability of players to change hand order, players could guarantee the card effect pairs with the best in-hand cards.

Consider another card, Daredevil, which lets you “see your opponent’s plays before you make your own” in a simultaneous play card game. Or there’s Agatha Harkness, which plays your turns for you. Yes, you read that right. There’s simply nothing Second Dinner card designers won’t mine for a mechanic, and they fully take advantage of the digital format possibilities in a way Magic is only starting to. It can feel like Second Dinner designers are rummaging the proverbial mechanic couch for pennies only to come up with twenties. I don’t know if this is a stable design foundation for a 10-year game, but it sure as hell is a fun one. I feel like I’m problem-solving in real-time with weird and exotic tools that surprise and delight. However, I suspect set-based mode restrictions will appear sooner rather than later.

Can card design be too fun? No, no, it can’t.

Naturally, the other design table stakes are improved. For example, a single card apparently solves the mulligan mechanic. Limited lane space is consistently played up as a mechanic. Card effects may destroy other cards in your lane, manifesting as positive and negative effects. “Opening up lane space” lets powerful cards take the place of weaker cards. The game’s lane events (revealed in each of the first three turns) layer a liveops mechanic on top of a Texas Hold’em Poker flop. The game’s limited turn count reduces the need for large decks; instead of 40 or 60 cards, players only need twelve single-card copies (no duplicates) to field a deck. It translates to less “draw screw” as players draw ~75% (9/12 cards) of their decks every game after considering starting hand size and turns limit. It’s also here monetization, and progression problems begin to rear their head.

Game Economy and Monetization

While Snap needs tailor-made monetization choices to pair with its peculiar design, I think the team outsmarted itself.

Snap will be (and already is) plenty successful, but every parent wants to see their kids reach their maximum potential, and game development is closer to raising kids than you might expect. Growing monetization into all it can be is a part of a game’s “flourishing.” A closer mirroring of the Clash Royale meta, more robust sharding implementation, and a drafting feature would drive ample revenue and sound ethics.

Game Economy

Players who want new cards receive them from…purchasing card packs? No! Instead, players increase their “collection level” by “frame-breaking”; spending “credits,” and a randomly distributed card-specific XP called “boosters,” to upgrade cards cosmetically. Completing frame breaks reward collection level points, which in turn, unlocks new cards, variants, and boosters. Players purchase gold which may buy a daily limit (!!!) of credits or cosmetic card variants. What I’d call the core economy loop revolves around collecting materials for frame breaking, frame-breaking, and unlocking new cards via collection level (which function as more material for frame breaking).

Value Flow

The Value Flow catalogs a bizarre way to pretend not to sell horizontal gameplay. It feels as if a misplaced sense of ethics encourages Snap to contort its design into a Reed Richards-esque figure instead of simply using card packs to distribute new cards. Instead, players are 5-6 hackneyed steps removed from USD, translating to new cards. In a more Royale or even aRPG model, randomly distributed boosters from packs would “fill” a given card until the player has collected enough shards to unlock that card. Sharding still allows developers to implement controlled RNG into unlocks as well. Here’s an example from Idle Heroes.

Snap essentially does this for boosters and frame breaks, but sharding should be expanded to new card acquisition. Instead, the team implemented high-stakes loot box special events at ~$300 per new card. Assuming Second Dinner was planning on time-based price discrimination, I don’t think they got the price wrong, just the distribution. Apex Legends faced similar initial battle pass criticism, but the issue wasn’t over the pass itself, but rather the difficulty (and thus implicit price). While diagnosing between structural and systems design problems is no easy task, I think the Snap team threw the baby out with the bathwater.

Structural or Systems Problem?

Sharding helps modularize and marginally distribute large single-value durable goods (you have it or you don’t). This is crucial for a game with 12-card single-card-copy decks, 80% less than Magic the Gathering. The nth+1 card value, where n is deck size, drops off dramatically. Players want to maximize the collection of those who take the field, not those who sit on the bench, and large deck sizes encourage players to chase a wide array of cards.

Growing monetization into all it can be is a part of a game’s “flourishing.” But design decisions made to avoid Reddit’s wrath and “fit in” with low-level Youtubers feels like Lindsey Lohan showing up with an oversized polo to loiter with the Mean Girls at the local mall. Community really does lead us astray, and it’s high time developers, and publishers stop confusing “community” with social media. A game’s community is anyone who plays the game, not just internet commenters. Devs, if the subreddit to our game is an open tab in Chrome, we are Lindsey at the mall!

Was Second Dinner so afraid of loot box comparisons they purposely worsened the booster experience? “Booster” packs are there for the taking! The lack of box/pack animations or UX feels odd for a game dripping in production value. And it’s only animation that separates random end-of-match booster rewards and caches from loot boxes or card packs. Adding pack opening animation helps further player understanding instead of an abrupt end-of-round flow, which doesn’t communicate or excite the player about which card will be selected for booster rewards.

Vertical Progression Isn’t Possible

It’s worth reflecting on why Clash Royale’s model wouldn’t directly translate to Snap. Clash breaks from traditional TCGs with “pareto efficient” upgradable cards; increasing a card’s level increases card stats with no adverse effects. A card’s level presents a massive swing in win probability.

This option is barred for Snap: cards don’t have RPG mechanics like health or armor to use in micro-dosing vertical power, and the card abilities are too carefully designed to absorb a radical re-balancing of card power between every player. Sorry, Deconstruter of Fun crew, this one doesn’t make design sense.

Cosmetics

I’ve written about cosmetics before, so I won’t say much more about Snap’s cosmetics system. It’s worth adding that frame breaking isn’t a cosmetic system as it is a cost-saving mechanic to fuel progression. It’s also curious to see Snap break from every other CCG in not building gameplay-related rarity into cards, which we’ll return to in the next post. The real cosmetics come from card variants, but so far, these feel flat with no vertical differentiation beyond a change in 2D art. This isn’t enough. Variants need meta-collection progression layers on top of variations in logos, animations, etc. Second Dinner must massively up its game if they want cosmetics to form Snap’s monetization backbone.

The roadmap appears to deliver on this, but it’s odd to see Snap drop both frame-breaking and rarity for features like Titles and Card Sleeves. It’s also another area that would benefit from more robust sharding. Second Dinner still gets credit for expanding the total monetizable surface with these horizontal cosmetics. Watching every other CCG fumble with player identity and status in a hypercompetitive PvP game has me reminiscing about Shaq’s inability to hit easy free throws. Although, I’m pretty sure no one is hack-a-Shaqing these other titles.

Battle Pass & Store

Instead of card packs, the battle pass does much of the heavy monetization distribution lifting. Let’s apply our trusty ADMC (Average Daily Monetization Cap) model to Snap. And low and behold, Snap comes out on top, next to Dota 2.

ADMC

While Snap does offer rewards past tier 50 (notably loot boxes), players cannot purchase them with gold bars. This is a missed opportunity to monetize cache packs and provides a low-PR risk method for Second Dinner to consider in the future.

The decision to price the pass in USD rather than gold avoids giving away a lifetime of rewards for the cost of a single pass. This was good early learning from the soft launch, where passes were sold for gold.

The battle pass sets the expectation cards will not only be included but rewarded at tier 50, helping drive more tier sink and conversion. Prima Facie, the battle pass, rewards a whopping 3500+ credits; we’ll break that into eventual collection levels and cards in the next post.

The limited rotating store is unnecessary; I’ve covered why these models struggle (and under what conditions they succeed) here. Limiting gold to credit conversations to about $5 daily is the only additional monetization crime. I’m sure this resulted from lengthy discussions between product managers and game designers, and it looks like the designers won. My only retort is that the game and players suffer under the credit purchase limit.

Drafting

Consumables drive “topless” monetization caps. All games contain a finite amount of durables, but consumables are repeatedly purchasable. CCGs naturally extend themselves to one of gaming’s great consumables: drafting. Drafting lets player challenge their problem-solving skills by having them assemble a deck of cards from random pool cards and requires that every opponent do the same. Drafting rarely concerns durable rewards and provides a unique problem test of wit, creativity, and on-the-fly system thinking.

Novel additions like Magic’s designer-driven Cube drafts also wrap a LiveOps opportunity inside an engagement mechanic inside a monetization mechanic. Snap needs to consider moving drafting to its more immediate future, and its 12 card decks are already optimized for drafting speed. While the lack of launch profile cosmetics is forgivable and perhaps even profitable (moving launch dates represents opportunity cost!), I was surprised to see drafting missing. If loot boxes are the lost cause, drafting provides a more structural solution to boost ARRPU.

Final Grade

Snap made several bizarre choices to avoid selling card packs. And while the horizontal expansion of cosmetics (card sleeves, titles, variants) exceeds comparable CCGs, expanding rarity along these dimensions feels like an easy win. A full embrace of card packs and a more robust version of sharding represent more expensive but higher revenue opportunities. The battle pass delivers as long as a Second Dinner maintains a monthly release schedule. Drafting feels like an engagement, liveops, and monetization trifecta, and I’d love to see it bumped on the roadmap.

The following post will look at Progression + LiveOps with a final post on UX. In the meantime, keep snapping! (On turn one, no less).

Previously confined to PC turn-based play, mobile free-to-play transformed 4x into sprawling persistent world MMOs. The combination was and continues to be electric; what other genre can admit to $1M LTVs? Each X,  EXplore, EXpand, EXploit, and EXterminate invites an unparalleled level of feature width and depth. ¹¹ Its LTVs are not driven only by spend depth but also by copious amounts of long-term retention. Like long, long-run retention.

The “10-year game” has gone from a CEO fairy tale to reality. A couple of weeks ago, Clash of Clans (CoC) celebrated its 10th anniversary by grossing over $480 million in the last year alone. But longer lifecycles present new and unexplored challenges. 4x faces a genre-level conundrum: given vertical progression and persistent worlds, what is the most effective way to manage server populations?

Shards

4x games typically funnel new players into “mirrored worlds” segmented by language and real-world region. Each shard shares a similarly sized population cap and “static” elements like progression speed or world map size. Tiny shards (< 10,000) like those in Star Trek Fleet Command create laboratories for experimentation and mistake. Even on 1M MAU, that translates to 500 parallel worlds or shards. On the other hand, a few games have instead opted for a single shard shared between all players. Managing single shards, like those in EVE and Dual Universe, is technically tricky, with over 200k DAUs sharing a single real-time and persistent server. Still, the shared sense of history from a single shard setup binds a community together in a way little else can.

Server persistence combined with vertical progression introduces unique time-based effects; additional doses of vertical progression increasingly stratify players across power levels. As the vertical progression cap increases, it becomes increasingly expensive for new players to win. For example, it may have taken 1,000 hours / $10,000 to reach level 100, but now takes 3,000 hours / $50,000 to reach level 150; “catch-up” is expensive for new players. Imagine starting a marathon hours after other competitors, only to find out the organizers have extended the finishing distance by several miles. And to be sure, extending vertical progression caps are a necessary condition for retaining elder players; nothing kills a game quicker than no new content.

Spent Cap Model

For any 4x game, consider the vertical ceiling or the max amount of progression as the sum of all individual durable timers (i.e., how long you have to wait to upgrade a building).

[latex display='true']\text{Vertical Ceiling} = \Sigma{DurableTimer_i}[/latex]

For most games, the vertical ceiling follows an exponential curve. Not only will leveling up a building require more and more time, but the total number of upgradeables increases as players progress. On the one hand, this suggests absolute vertical power increases exponentially.

But on another, the returns to time in terms of vertical power percentile decrease (i.e., a given hour or dollar of spend buys less percentile increase as the absolute amount of time/dollar spent increases). It’s easier to go from rank 500 to 400 than rank 100 to 1.

The Seat of Power

Most 4x games revolve around a “seat of power.” Players band together to control the seat, with a single representative from the clan executing against the clan’s interests when on the seat. In some sense, this creates a fixed supply of “victory”: there can only be a single holder of the seat of power. The seat itself gives the holder a vast array of powers. For example, in March of Empires, ruling players control everything from tax rates to peace shield rules.

Ceilings and Seats

We can imagine the vertical power percentile representing the time and costs to obtain the seat of power. The probability of obtaining the seat of power is conditional on the population size of the shard.

[latex display='true']P(W_i \mid S_i)[/latex]

Where:

As [latex]S_i[/latex] increases, the top “bid” to hold the seat of power increases, and the probability of any one player obtaining it decreases. And as we know from above, the “efficacy” of a dollar declines as the absolute amount of spending increases. On the other hand, increases in [latex]S_i[/latex] also increases utility from holding the seat of power. Consider the relation

• [latex]S_i[/latex] is the active player size of the [latex]i[/latex]th shard
• [latex]W_i[/latex] is the seat of power for the [latex]i[/latex]th shard

[latex display='true']UW_i = UW_i(S_i)[/latex]

with

[latex display='true']UW_i' = \frac{dW_i}{dS_i} > 0[/latex]

where [latex]UW[/latex] is the utility from holding the seat of power. I’ll go out on a limb and claim it’s more fun to rule more people. The key consideration for firms is to find the optimal point that balances increases in shard population, thereby increasing the utility from the seat of power against the decline in probability any one player obtains the seat of power.

However, there’s an additional design space to consider in managing shards over time. Until now, we’ve considered a static problem: what is the optimal shard size? But given the persistence of shards, we must also consider the design space of merging and re-filling shards.

Consider a round-robin style model: new players are assigned shards from a looping list. This ensures a consistent influx of immigration, but players entering elder shard face difficult prospects. By then, top players have accumulated massive amounts of capital.

A former colleague suggested an alternative model: “shard bursting,” or filling up shards in a series, i.e., the following 6,000 new players might all be assigned to shard “A” while the next 6,000 are assigned shard “B.” Bursting hedges against the “round-robin” model by creating a relatively even start line from which players rush to the seat of power. On the flip side, brutal early retention funnels (sub 20% D1 retention) leave bursted shards looking something like modern Japan. With no fresh blood, the shard population will dwindle over time.

Another model: “the open border model,” struck me recently as another exciting possibility. In this model, players move between shards freely and at will. As when examining countries in the real world, players could access data on production, activity, and, most importantly, war when examining shards. The model incentives an equilibrium wherein all of the shards have the same probability of any one player obtaining the seat of power. Such that [latex]P(W_1) = P(W_2) = P(W_3) \cdots[/latex]. If a shard’s population caters, new players will enter the shard in an effort to maximize [latex]P(W_i)[/latex].

I’m not actually sure which model produces the most robust retention rates, maybe we should auction the whole thing off! But I do know the problem is here to stay, and the open economies of web3 only complicate the dynamics. Web3 4x designers certainly have a curious task at hand!

Economy design maintains a UX component. It’s vital for the designer to draw the cause <> effect loop between action and reward for the player. Currency animations, played when players claim or complete a task for a reward, stitch the acts into an experience. Wallet amounts don’t magically increase; the currency flows from the claim button to the wallet’s UI location.

The animations amount to a classically conditioned injection of dopamine, not so dissimilar to the original story of classical conditioning-centering on audio-visual reward cues. In that story, Russian scientist Ivan Pavlov presented a stimulus (not a bell!) before feeding a group of dogs. He would collect and measure saliva volume samples after the trigger but before giving the food. Eventually,

[…] Pavlov noted that the dogs would often begin salivating in the absence of food and smell. He quickly realized that this salivary response was not due to an automatic, physiological process.

Based on his observations, Pavlov suggested that the salivation was a learned response. Pavlov’s dog subjects were responding to the sight of the research assistants’ white lab coats, which the animals had come to associate with the presentation of food. Unlike the salivary response to the presentation of food, which is an unconditioned reflex, salivating to the expectation of food is a conditioned reflex. [1]

Acquisition of a currency is a step removed from acquiring the items the currency can purchase—the time between acquisition and disposal of currency forms anticipation.

Principles of Great Currency Animation

Great currency animations:

1. Show the currency directly entering the wallet UI location
2. Ensure the currency reward is hard to ignore
3. Make the player feel rich while not cheapening the currency signal (i.e., animate more currency than rewarded)
4. Match animations to powerful audio cues

Mobile Legends Adventure, above, doesn’t showcase a wallet drop-down and violates 3. (notice the number of purple diamonds, it’s less than 200—the actual reward). The audio leaves much to be desired as it only sounds like a single gem has been earned (4. fails).

Game Dev Tychon

The sheer volume of icons is impressive, but the art communicates low-value “points” instead of currency. The currency flow from the computers to the UI location demonstrates the type of assets each programmer accrues. Unfortunately, the sound is lightweight, and thus the reward feels light.

Spyro

The gems convert to bouncing numbers on collection, providing a visual firework of sorts. The gem sounds are traditional, but the sound scoring effect is dated. Spyro may have very well helped pioneer the “zone of attraction” effect whereby a player “sucks in” currency around them.

Clash Royale

The audio design is top-notch, and so is the UX flow. Players always hit the main menu after receiving gold, so the connection between chest reward and gold balance increase is tight. The sounds indicated a TON of coins. Suddenly I’m Scrouge diving into a pile of copper coins. The only downside is seeing the gold form a straight line in the animation, thereby minimizing the total visual impact.

Runner-Up: Beatstar

Beatstar’s UI oozes sex. It’s as if someone who exclusively works from home in red velvet bathrobes developed it. The UI art merges with animation; coins flip and spin, reflecting light when they enter a player’s balance. The artist again uses the motif of spinning objects with track stars, mimicking the opening of the Paramount “Mountain Credits.”

Oh my, the music. Everything feels like it’s reverberating from a subwoofer, almost underlining the importance of the reward. Each currency gets a unique sound profile so players can differentiate on sound alone.

Winner: Brawl Stars

Surprise, surprise, it’s another Supercell win, but to be fair, Brawl Stars deserves it. The currency pauses, mid-animation, ensuring players admire it, before directly adding the currency to the power point bar. The amount of currency, in this case, 13 units, is the exact amount added to the power point bar. It’s impossible to miss as the currency spreads around the character portrait (“Poco”) rather than form a direct line, as we saw with Clash Royale.

The audio suggests chips instead of coins, a refreshing break, but the improvement of the power point bar suggests an almost score-like element.

Currency plays a peculiar role in economic activity. Classical economists are fond of claiming “money is a veil”; it abstracts away the underlying economic activity it helps coordinate. However, we know all too well that money is only a veil until it’s not. Monetary economics is justifiably a subdiscipline, so it makes sense for “tokenomics” to emerge as something similar to blockchain. And like early monetary economics, tokenomics finds itself suck in a strange mercantile stage of development wherein a token’s highest order is to appreciate rather than facilitate transactions. Something akin to increasing net exports and not letting money “leave the system.” But more specifically, modern tokenomics falls prey to three problems:

1. Fixed Supply & Fixed Supply Schedule
2. A Means, Not an End
3. Make-Believe Ownership

Fixed Supply & Fixed Supply Schedule

Price stability builds trust because it is predictable. A fixed number of tokens and a fixed schedule in which those tokens enter circulation handcuffs the ability of monetary policy to respond to inflation or deflation. When stability collapses, so does trust; look no further than the recent wave of stablecoin disasters.

In the spirit of considering money as a veil, consider the effects of fixed supply and supply schedules on something like Magic the Gathering. In this parable, imagine Wizards of the Coast announcing a fixed number of card packs for the next set and a schedule to sell those very packs (ex., one-hundred packs for sale one week after release, three-hundred the week after). If there were many players and thus high demand, prices would be prohibitively high. Without the ability to purchase cards at a reasonable rate, players might churn to other games; a high price isn’t always the profit-maximizing price. On the flip side, too many packs for a given number of players could mean low prices. It’s hard to imagine Wizards of the Coast maximizing profit if they “minted” cards at the cost of production rather than exploiting their monopoly on card supply; P doesn’t equal MC outside of perfect competition.

Optimal monetary policy adjusts based on the conditions of growth and contraction, while most tokens prevent any such possibility. 

A Means, Not an End

Economy designers generally model currency as a medium between activities rather than an end to itself, and economists do the same. Instead, however, cryptocurrency white papers spend more time on token distribution and allocation than redemption, ignoring that a currency is only as valuable as it can purchase. 

Packy McCormick, an equally entertaining and informative writer, deep dives into the business models of start-ups in his weekly newsletter, Not Boring. In Braintrust: Fighting Capitalism with Capitalism, he chronicles the model of Braintrust, a talent platform. Late into McCormick’s explanation, he documents the cryptocurrency “twist” of Braintrust. By using and referring talent to the platform, users can earn tokens. It’s never clear, however, what these tokens gate or purchase. Users on the platform may price services in terms of the token, but they’d only do so if the token has purchasing power in real terms. Packy writes that Braintrust may offer “hints at future benefits for token holders; these might be things like educational content, free software, or coaching.” If so, the token’s price reflects the value of these services. I find it excruciatingly difficult to believe these services are worth much, so, in turn, the token shouldn’t be worth much. Starting a fiat currency from the ground up is a daunting task, but the most important precept of any token-backed project should be to grow demand for assets priced in the token.

Blockchain games do better here by exclusively first listing goods in terms of the game’s tokens; the token’s price reflects the demand for these goods. In the case of Braintrust, the token’s price reflects a future expectation that users or the firms will find a use for the token. Rather than composing a minority of a given whitepaper, explaining why there’ll be demand for assets priced in the token should be the analysis.

Make-Believe Ownership

A stock represents both a governance right and a claim on dividend payments from a firm. More recently, markets segmented governance rights between Class A and B shares. B shares vote on governance issues while class A shares do not. Most blockchain games took note and immediately segmented governance rights into separate a token altogether. In this way, there is some sense of ownership. But so much of this token class stands on shaky legal grounds. Are blockchain governance votes legally binding? Over what issues do token holders have autonomy? Can they fire a CEO? Increase a token’s supply? Again, the entire issue of ownership is murky.

On the other hand, no form of token entitles an owner to dividends. If it did, tokens would fail to pass the Howey Test and be subject to SEC regulation. As mentioned above, a token’s current price represents present and future expected demand for the assets redeemable by the token. This is a wildly different value proposition from a traditional share which derives value as a claim to a stream of dividend payments. Far from “aligning incentives,” the different incentive structures result in different outcomes. Redemption incentivizes the price of the underlying assets to increase; this isn’t always in the best interest of the underlying product or service. However, dividends incentivize the growth of a firm’s profitability. Profit incentive finances and signals supply expansion. The supply expansion makes everyone better, whereas redemption results in something like what we see with the NIMBY’s homeowner movement. In translated terms, it might make sense for someone who holds a hypothetical governance token in Magic the Gathering to vote against issuing new packs of cards. It’s not hard to see why this would be destructive to the game.

Forward

As blockchain applications grow and fluctuate, monetary policy will increasingly come under a microscope. Vague regulatory guidance and short-run design decisions to build trust have forced awkward token design. Moving away from these constraints will produce sustainable applications while aligning incentives between token holders and users.

It’s incredibly gratifying to help set people on the right track, given I was asking for the same help years ago. There’s so little on game economics it forces aspiring Economists to cold-call people with the title on Linkedin. This blog’s mission is to grow the conversation and those participating in it.

Along these lines, a student interviewed a fellow Game Economist for their master’s thesis. They shared the questions with me, and they were a lot of fun, so I’m reposting them here with permission.

What would you say is the biggest difference between the virtual economy in MMOs and the real economy?

Let me talk about virtual economies more broadly. In some sense, there’s no invisible hand in virtual economies. A virtual economy does not exist without a central institution that creates it. A designer needs to craft each inch of a virtual economy and how it functions.

In the real world, the environment is exogenous – a given, with each agent optimizing around his or her utility preferences. An agent is still optimizing around their utility preferences in a virtual world, something we have ample empirical evidence for (thank you, Edward Castronova!).

If game designers were to design the real world, they could alleviate things like hunger instantly. But a game-developed world wouldn’t mean a lack of pain or difficulty. On the other hand, we observe optimal flow state comes from overcoming challenges.

If we were to think about MMOs versus other virtual economies specifically, it’s the prevalence of player-to-player trade-in MMOs. This gives rise to floating prices facilitated via auction houses and provides a whole new dimension of engagement.  

Why do you think players value virtual goods more than real ones?

I can’t entirely agree with the premise of the question. The same Magic the Gathering card will command a higher price in physical than digital form. This will shift as Wizards of the Coast brings more events online than offline (i.e., the Pro Tour) and expands their online offering (Cube Draft! Commander!).

What type of goods is the most desired in MMOs and why? (Functional, hedonic, social)

Done correctly, whichever goods the institution incentivizes players to desire the most. A game where I could not see other players’ cosmetics would probably sell fewer cosmetics than a story-based DLC model, ala Destiny. Empirically, when I observe most MMOs, vertical progression or gameplay affecting items are desired the most—progression gates content that arranges a social hierarchy.

How is a competition created in games like WoW, FFXIV, or Lost Ark?

There’s the meta-layer of progression mentioned above, which adds a competitive element to the social hierarchy. Progression provides straightforward UX to “move ahead”; it’s incredibly meritocratic or at least “time-o-cratic” in many ways. It’s unclear how Lindsey Lohan moves up the social ladder at her high school in Mean Girls. In MMOs, the answer is clear: spend more time grinding!

Some MMOs have dipped into real-time combat-based elements like Alteric Valley in WoW or arena’s in Lost Ark. This re-arranges the social hierarchy to not depend on time but on “twitch skill” with a decent helping of planning. What equipment and skills players enter the Valley or arena with are dependent not only on time spent grinding but also on the planning players undertake on skill tree choice and gear equipment.

Why do you think some players rather spend money on buying an already leveled-up account instead of playing the game themselves?

Sometimes the Wikipedia article is better than the actual movie. Buying high-level characters let spenders jump straight to the top of the hierarchy. They may be missing out on the “experience,” but everyone has a soft spot for shortcuts. 

Would you say that high-level players play a role in making new players buy the virtual currency in games?

It depends, again, on the institutions the designer has created and what that institution incentivizes. The more social matters, the more high-level players probably play a role in monetization. But a game like Candy Crush has much less social. If you were to remove Facebook Connect from Candy Crush, I don’t like revenue would change much in an A/B test. I love Gardenscapes, but some players on level 1,000 do not change my propensity to spend as a new player.

How can someone create a fun virtual economy for players when economics is one of the most boring studies?

I could not disagree more! Economics is the most riveting social science out there. Economists commonly refer to “the set of glasses” the discipline grants after careful and dedicated study. The economic model helps me organize and think systemically about my world. There’s a satisfaction in the ability to take a random observation and build a model of understanding around it. Something that was chaos becomes order.

The best Game Economist is someone who brings to bear the tools of economics on games. It’s what Gary Becker does for crime or Richard Posner for law.

What game has the best economy, in your opinion? 

I think of the best economies as ones that maximize the net present value of the game. Economies here don’t refer only to how many currencies a game has or the price of items. Economies also refer to things like the supply of content in a game. I have a couple of nominations: 

Valve’s Steam Marketplace

Valve deserves a lot of credit for bringing player-to-player trade to a broad scale. The auction house and UGC design have given huge revenue and engagements tails to the games that have adopted it. Valve’s titles continue to have strong engagement, including TF2, a 15-year-old game that survives on community rather than first-party content. 

Clash of Clans / 4x titles (Game of War)

Selling vertical and horizontal progression is expensive since creating supply is costly (i.e., new level cap, new MOBA character). Clash of Clans (CoC) monetizes based on timers for creating troops. CoC treats a given troop as consumable; a single troop has only one attack use before it must be re-generated. The hard currency players can spend on reducing the generation time for new troops is a nearly infinite depth sink. 4X games use a similar loop while adding on consumables like peace shields. The result is LTV in the seven figures. The lack of consumables and expensive card creation in Clash Royale has significantly harmed its long-term prospects.

Match-3s

Perhaps an odd choice here, but match-3 economies are well understood and tuned. Match-3 developers understand the “meta” of the genre: how do I optimize level difficulty against the speed at which I can create new levels? There’s beauty in the rigorous optimization problem, and it seems to be going well. Match-3 grabs 21% of U.S. iOS spend, with players approaching decade-long relationships with a single title. Candy Crush recently celebrated its 10,000th level.

Netflix is seriously ramping its games division. As of February 2022, I scrapped ~25 game or game-related titles on Linkedin. Excluding Night School (+15) or Next Game (+125), Netflix Games probably approaches an internal headcount of ~50-60. With both studios, ~150-200 Netflix Game or Games related employees. At 14 titles, we’ve seen of things like Dungeon Dwarves and Krispee Street, highlighting small indie-based narrative adventures rather than big-budget AAA affairs. On the contrary, I’ve yet to see a pundit to suggest Netflix get involved in the recent game M&A spree. However, in an industry turning hard right on live service, Netflix is justified in turning left.

Subscriptions spread cost among participants; the higher the denominator, the lower the cost per subscriber. From The Economics of Game Pass Demand a Playstation Launch:

The supply-side economics of subscriptions are fairly straightforward: get massive scale and distribute costs. Netflix has over 160M subscribers meaning a $100M production only costs each user $0.63. There’s zero marginal cost (unlike Spotify), so the more subscribers Netflix has the less a piece of content costs on a per user basis. At $11 per user per month, Netflix can spend over $1.7B on content each month and break even. In fact, as recently as 2019, Netflix spends close to break even. In the long-run, Netflix has a incredible loop: content brings users, these users lower per user content costs, this accelerates more content spend, which brings more users. Rinse, wash, repeat until diminishing returns take hold.

The challenge with game subscription services is that players rarely play more than one game at a time; there isn’t a thirst for new games when League of Legends and Fortnite are dropping free weekly content. This is, after all, the point of live service: foster long-run retention with a particular game. But live-service economics washes away in a world of linear-based gameplay and limited content updates. Instead, games like Control or Tomb Raider look like Too Hot to Handle or Peaky Blinders: driving {y} number of viewer/engagement hours for {x} cost. In this framework, games are nothing more than another type of growth content to feed the subscription loop. In an interview last year, Netflix’s Chief Product Officer all but confirms this notion (emphasis mine):

And we also feel that our subscription model yields some opportunities to focus on a set of game experiences that are currently underserved by the sort of dominant monetization models and games. We don’t have to think about ads. We don’t have to think about in-game purchases or other monetization. We don’t have to think about per-title purchases. Really, we can do what we’ve been doing on the movie and series side, which is just hyper laser-focused on delivering the most entertaining game experiences that we can’t.

Reed Hastings claimed the biggest competitor to Netflix was not HBO but rather Fortnite. Initially, I dismissed this as Valley babble, but with 66% of kids ages 9-12 playing Roblox, he may be onto something.

In 1989, political scientist Francis Fukuyama wrote an essay asking if we’ve reached The End of History? Often mischaracterized, the essay argues democracy and free markets represent the last evolution of political-economic systems. Fukuyama writes:

What we may be witnessing is not just the end of the Cold War or the passing of a particular period of post-war history; that is the end point of mankind’s ideological evolution and the universalization of Western liberal democracy as the final form of human government.

Blockchain may well represent game monetization’s End of History.

Following a Kuhnian process, game monetization has been through its own evolutionary process. But by all means, the evolutionary process has not been linear. The transformation from coin arcades, fixed-price titles on-at-home consoles, DLC, and free-to-play has happened in uneven bursts. Lest us not forget, between 1981 and 1984, arcade revenues halved from $8 billion to $4 billion. The saving grace was Nintendo’s 1986 launch of the NES that buoyed an arcade comeback. Not everything is up and to the right. We’ve had unique detours, including per-game subscriptions ala World of Warcraft and a never-ending flirtation with ads in all forms. The last two decades appear to showcase an accelerating rate of change. The time between the launch of the Xbox Live Marketplace (2005) to Farmville (2009) was a mere four years, with Candy Crush launching on mobile two years later, in 2011. Free-to-play has only grown since but hasn’t entirely delivered the finishing blow to fixed-price titles. Instead, the industry has grown; as we know, rising tides lift all boats, so we find different game experiences supported by different models.

However, Kuhn observes the ability of old paradigms to survive if they can integrate the questions posed by new paradigms. Blockchain gaming isn’t about “true ownership” or “decentralization,” which are pseudo-socio-political causes irrelevant to the gamer. What started with the Steam marketplace is crystallized and popularized in blockchain games: markets.

Technology Doesn’t Always Work Backwards

I’ve been particularly hard on “true” ownership, but as the tweet that opened this series opined, something is interesting about blockchain underneath all the craziness. That interesting something has been staring us in the face. If the last year has shown us anything, it’s not just markets; it’s so obviously markets. And not just the video game kind, markets in everything.

Public and open readability of tokenized assets enables the beautiful world of price discovery to bear down on games. It might not be revolutionary, but it’s undoubtedly evolutionary. Yes, individual games can adopt the Steam Marketplace or build their own auction houses as EA has done with FIFA. But doing so is prohibitively expensive and fraught with cross-platform complications. Asking, “what can blockchain do that normal games cannot?” is not a particularly useful or informative question. It’s the sort of question your parents asked when you desperately argued for them to drop sending snail mail in favor of emails. Or perhaps we can revisit naysayers suggesting the iPad is just a larger iPhone. And yet, the iPad is precisely that. I’m not sure Apple planned for the iPad to become check-in machines at workplaces or cash registers at coffee shops.

Instead, it’s useful to ask, “What does blockchain make cheaper or easier than alternatives?.” There’s room for faster horses just as much as cars.

Little developer action is required to enable on-chain assets to list on OpenSea or Mintable. The public readability of blockchain allows for multi-homing across various marketplaces. Making a token or minting a new on-chain asset for auction is surprisingly easy. The technology is there, but the design, cultural resistance, and lack of internal corporate advocates have hampered auction houses from advancing sans blockchain; gaming has a bizarre relationship with player-to-player trade.

Everquest, MMOs, and Something, Something Diablo

The vast majority of games operate under autarky or near monopoly-monopsony relationships, the paradigm we’re operating under when buying and selling items in most RPGs. But beginning with Everquest (1996), MMORPGs carved out a small genre expectation that has run for decades. MMORPGs provide a natural genre-feature fit; markets are particularly useful when participants have a range of comparative advantages. As we know, MMOs enable everything from fishing to chopping wood. Auction houses never became a do-or-die feature despite on-paper justification as an engagement mechanic. For example, it took Runescape as much as three years after its launch in 2007 to implement one.

The existing MMO markets traded in virtual currencies; participants could only spend currency earned via in-game activities to purchase items. Diablo III provided a mainstream break away from this model; players could sell items for real money.

Nearly every conversation about auction houses mentions Diablo III in passing. The comments’ substance amounts to, “auction houses are interesting, but Diablo III’s failed.” It’s become something of an academic exercise to point this out. And while Blizzard did remove both the real and virtual money auction houses from Diablo III, they were far from a failure. Conversely, they were a smashing success. 50% of players regularly engaged with the auction house (!), with some players bringing in enough to supplement full-time incomes. Sound familiar? Jay Wilson, Diablo’s game director, further conceded the auction house performed as intended to combat account fraud.

Diablo II revealed players were already trading items but doing so off-platform, resulting in scams (repeated in World of Warcraft, mind you). Given all this, why kill the auction houses? Blizzard didn’t want to sell progression.¹²

Over the years, I’ve spoken to designers from the Diablo team, and I’ll immediately ask about the auction house. While stories differ and memories are hazy, there’s repeated mention it caused quite a bit of strife on the design team. “No feature is worth breaking up the band for,” a designer recalled. Fair enough, considering Diablo III is one of the best-selling games of all time. Looking back, it’s a very early-2000 design philosophy and quite silly considering what we now know.

Valve and the Auction House Gap

Essentially a research institution, Valve is always lightyears ahead of current industry thinking, whether in battle pass, VR, or auction houses. It’s worth pausing and considering Blizzard announced the auction house in late 2011, while Valve launched the Steam marketplace only a year later, in 2012. And, of course, Gabe contacted Yaris before Valve launched the marketplace.

Similar to battle pass, we can capture the value of marketplaces in a couple of different ways. For the post’s purposes, let’s consider strictly firms rather than players. A simple way to consider auction house value to the firm is as the difference between auction house price and the next best pricing scheme:

[latex display='true']MR = \sum (\text{auction house price}_i \times TF) - OC_i[/latex]

Where,

• [latex]MR[/latex] is the total increase in revenue from an auction house
• [latex]\text{auction house price}_i[/latex] is the auction house price for the [latex]i[/latex]th item
• [latex]TF[/latex] is the marketplace fee (%) for selling the item
• [latex]OC_i[/latex] is the next best price by an alternative mechanic for the [latex]i[/latex]th item

Over and over, we’ve seen just how glaring this difference is. What developer would know to price an item at $10,000? The more significant the gap, the more auction houses make sense. The advent of live service games is growing this gap the previously unimaginable heights.

Lifetime Utility

When players engage with a giver game measured in years rather than days, time horizons shift. Game assets are fixed upfront fees that provide a stream of benefits over time; the longer the time, the bigger the net present value of the dividend. From Game Companies Are Not Tech Companies IV:

The efficiency of any monetization or pricing system is the degree to which it can correlate marginal cost (MC) to marginal benefit (MB). In the above examples, we fixed the price. Fixing the price makes sense, given that the utility curves flattened out. But the more the curve refuses to flatten, the most decorrelated MC to MB becomes as Time continues.

The 140 character history of game monetization is a cat and mouse game between the ability of games to grow LTU and pricing mechanics trying to catch up and appropriately price those LTU gains. Auction houses represent the final form of this cat-and-mouse game. Prices can automatically adjust to increases (or decreases!) in the value of in-game assets. In-game assets, however, only cover part of the transformation. Sitting adjacent to auction houses is the ability to invest games directly via tokens—to be discussed in a future post.

Blockchain is making a real play to bring markets in games to the mainstream; whether or not the feature survives in strictly blockchain form remains to be seen.

But economists have had a flirty relationship with games; Edward Castronova’s Virtual Worlds: A First-Hand Account of Market and Society on the Cyberian Frontier remains one the most downloaded papers in SSN history. And well… there’s not much else. I’ll put a stake in the ground and claim Castronova as the world’s first Economist to study video games. He generally gets lost in the discussion to the most talked-about Game Economist, Yanis Varoufakis.

When Valve signed Yanis, and it was at Gabe’s academic behest, news articles sprung up left and right. Despite producing a whooping three blog posts during his tenure, his name resurfaces as a fun fact wherever game economics discussion appears. The Varoufakis story stuck because it brings validation to the industry, “Look, a real economist studying games, isn’t that just wild?” And Varoufakis is certainly as real as they get; he would lead Greece through the Eurozone crisis with a spice that would make citizens of Arrakis jealous. But I think this serves as another reminder that the industry suffers from a strange inferiority complex. From Game Companies are Not Tech Companies:

Games never seem to get the mainstream or broader tech circle legitimacy many think they deserve. Despite operating in the Valley and major tech hubs, game companies do not reach the crazy evaluations of FAANG (ATVI has a market cap of $63B, while Facebook sits at $215B). Furthermore, public intellectuals like Tyler Cowen or Ben Thompson rarely discuss games as a share of their commentary (not the case with Matthew Ball – but maybe he drinks too much kool-aid). Even internally, game firms seem to be chasing subscriptions off the heels of Netflix and Spotify.

We continue to see the same sentiment whenever an MMO has a “real economic problem.” Despite economists and economic problems acting as validation stamps, there’s a genuine utility to economists in games. Hopefully, this blog serves as a small piece of evidence for that utility (it certainly isn’t my Twitter).

Since Varoufakis, there’s been a smaller and much quieter wave of tech Economists who have fully converted to Game Economists. It’s important to point out the evolution from Economist to a Game Economist is distinct and essential, a process I’ve struggled with myself.

Like Castronova and Varoufakis, my imagination nearly imploded at the thought of games meeting economics. Economists crave the ability to run crazy experiments with large sample sizes at low costs. Games seemingly provide the perfect laboratory to test micro and macroeconomic theory; look no further than Levitt’s King paper or World Of Warcraft’s virus outbreak. However, after working in games, my mindset started to shift. Instead of asking what games can do for economics, I started asking what economics can do for games. A sort of discipline puberty, if you will.

So What Can Economics do for Games?

A few years ago, I applied the gaming company I worked for to participate in the University of Chicago’s Field Experiment program. The program links academic economists with companies in the pursuit of running large-scale randomized field trials. The academics hope to get a published paper based on a field experiment, while the companies hope to get top-tier analysis at no charge. It was an absolute blast to visit Chicago and meet the father of field experiments, John List, and Bates Medal winner, Steve Levitt. But the program also reminded me how tunnel-visioned economists and companies could be. Levitt showed absolutely no intellectual curiosity in the breathtaking results of his three million player study. Massive quantity discounts did not only fail to increase revenue but failed to convert non-payers into payers. The implications of such a result are vast and wide-ranging, yet when I talked to Levitt, he considered the experiment a “failure.” He considers it his only “failed” field experiment with a company. It was a soul-crushing exchange, but his incredible partner and paper coauthor carried on the torch the King-side. I tremble with excitement at what the Experimentation team at King has done since, a shining example of Popperian science.

Later on, Levitt mentioned speaking to none other than Valve but felt they were mostly trying to “show off” in conversing with him. Something along the lines of “see how complex and real our economies are?” I guess the industry’s inferiority complex rears its ugly head again.

The explosion in demand for Game Economists and projects of Varoufakis and Castronova have something very particular in common: floating prices. Traditional games operate on a monopoly-monopsony model: one buyer and one seller. But here’s a shocker for readers: Economists like studying prices, but for prices to reflect information accrual, they need to be the result of barter. Hayek famously points this out in his The Use of Knowledge in Society essay:

The whole acts as one market, not because any of its members survey the whole field, but because their limited individual fields of vision sufficiently overlap so that through many intermediaries the relevant information is communicated to all. The mere fact that there is one price for any commodity—or rather that local prices are connected in a manner determined by the cost of transport, etc.—brings about the solution which (it is just conceptually possible) might have been arrived at by one single mind possessing all the information which is in fact dispersed among all the people involved in the process.

Studying prices and their relation to specific events provides critical insight into what has happened and how to act. Prices are signals.

And prices are precisely Castronova and Varoufakis’s work: Castronova studying Second Life‘s open economy, and Varoufakis doing the same for Team Fortress 2. If there is anything blockchain games will mainstream its auction houses. They are notoriously hard to manage, but the gains to do so are vast; ask the FIFA Ultimate team developers for evidence of both claims. On the other hand, games are taking their first steps into monetary policy with tokens, and it’s been… rough (more on that later in Sense & Nonsense in Blockchain Games).

Let’s stop here to note anyone can contribute to game economics, but I don’t think everyone who does is a Game Economist, just as a mechanic can make a car better without being a driver. Despite this, the blockchain gaming move to open markets means it’s helpful to delineate between a Game Economy Designer and Game Economist. A Game Economy Designer is generally not a Game Economist, but a Game Economist can be a Game Economy Designer (but isn’t always). Sound confusing? Let’s discuss.

A Game Economist has some shared and unique skills compared to a Game Economy Designer (as does an Economy Designer to an Economist!). A Game Economist should have the usual economic toolkit: marginal analysis, causal inference, econometrics, model building, coding (!!!), etc., but bears the toolkit down on games. They should be masters of vital economic concepts while evangelizing them to a team, both in the game’s design and production.

Economists Rank the Most Important Economic Concepts

The Game Economist marries economic theory, data science, and a game design into polygamous glory. A great Game Economist throws in product management skills for good measure. I’d propose a simple litmus test for a Game Economist: could this individual be employed as an economist outside of games, for example, at Amazon, eBay, or in a university? This isn’t meant to be some sort of elitist creed, but instead, be clarifying. The differentiation matters: opportunity costs make Economists a lot more expensive than game designers due to opportunity costs ¹³.

We still need to build the foundational economic models of games, just as Becker did for crime or Buchanan for public choice. The dump truck has barely arrived, much yet started to pour the concrete. We’re still mixing at the factory! The most important thing we can do now is recruit, retain, and engage a new wave of Game Economists.

But blockchain warps the traditional meaning of ownership into a ship of Theseus problem; each blockchain-based game can define ownership in its context. Suddenly, ownership becomes a design space in itself, making it all that much more crucial to dissect and analyze.

There are three stylized notions worth exploring:

1. Ownership versus Licensing
2. Item Possession versus Key Possession
3. Obligation to Honor

Does a License Count?

A few years back, John Deere faced a legal challenge over preventing customers from repairing tractors. The key legal question was the extent of ownership.

John Deere’s license agreement prohibits new tractor owners from tampering with the “Security Measures” on embedded software. Right to Repair advocates want the ability to modify this software to perform repairs and modify their machines. After all, farmers own their tractors, shouldn’t they be allowed to edit or replace their software? 

Ask a bystander what it means to own something, and they’ll say something along the lines of “the ability to do whatever you want with it.” But a definition like this would leave out the 58% of homeowners under Homeowners associations (HOA). These associations sometimes prohibit things like painting your house pink without authorization. HOAs intend to protect the value of the community from negative externalities; a pink house or a boat in a neighbor’s driveway can reduce the value of the homes around it. Despite the restrictions, we’d still consider someone the “owner” of an HOA house.

Deere and HOAs demonstrate that ownership is not a single “ability” but rather a bundle of “abilities” or rights. Such a view is consistent with Adam Smith’s conception of property. And while blockchain can facet such abilities, the abilities themselves derive from a socio-legal institution rather than a technological one. In nearly all cases, blockchain games don’t confer “true” ownership of an asset beyond the ability to buy and sell. It’s definitely something, but perhaps it’s a bit of a stretch to consider that single ability “true” ownership.

Being on the blockchain guarantees little; the design space is vast. And that’s a good thing. It just might not lead to the world cryptonites ¹⁴ imagine.

Sorare and the Metaverse Fantasy

In a claim repeated so many times, I’ve nearly become queasy; cryptonites imagine a future where players “import a CS: GO skin into another game” or move assets from title to title.

Supposedly, this sort of possibility builds the metaverse (I’m still confused on the whole metaverse thing). But most blockchain games do not have the legal or technological framework for a “game IP as a platform” to occur. Take Sorare as an example, a firm I’ve written about before. A player cannot take “their” football cards and build a new mode or FIFA-like game utilizing them. “Ownership” does not imply an unlimited license or even an exclusive one to the token owner.

Blockchain ownership doesn’t mean copyright, but it can. The demands of external IP holders limit Sorare, and nearly all games using 3rd party IP will likely be limited in the same way. The “interoperable” world fantasized by cryptonites will require original IP or a new middle ground of copyright rules. A new copyright paradigm may come forward, and to be fair, we have seen layer-two games like KittyRace, based on CryptoKitties, emerge. Sky Mavis has something of the sort, but it’s more of a reactionary framework than a leading one.

Curiously, despite all of the money sloshing around blockchain games, we’ve yet to see a layer-two game studio. Instead, we’ve seen internal IP worlds and shared asset experiences, ala Riot and the Marvel Climatic Universe. Illuviuim has already announced a second game, and Sorare/Ubisoft are experimenting with a one-shot league – both products use or will likely use a player’s existing collection of blockchain assets. But an independent studio would be a significant validation of the blockchain gaming future many articulate.

Finding the Right Analogy

Beyond copyright skepticism, there’s technical skepticism to “true ownership.” The assets players “own” are not stored on the blockchain. It’s simply too expensive (more than $100 per GB and rising) to put large files on each node. Some firms employ IPFS, but centralization and trust rear their ugly head again.

A better way to think about blockchain ownership is as a unique and independently verifiable key to the storage location of a limited license asset. Sort of like the keys to a Public Storage warehouse holding a USB drive that can work on some ports. Not quite as sexy as “true ownership,” but closer to reality.

Under No Obligation

Aristotle argues that politics concerns the struggle for recognition. Ownership follows a similar challenge. As mentioned above, it’s a socio-political concept rather than the technological one. An under-appreciated ability of blockchain games is to reject their assets. For example, the card game God’s Unchained does not have to honor the attack and defense on an on-chain card. The God’s Unchained game engine can prevent a player from using a particular blockchain card in the game altogether. The engine could even translate a given card token into an entirely different card! Characteristics of items on the chain (to the degree that such things are even on-chain) are excellent suggestions, but no one is under obligation to utilize them. And again, this is a feature, not a cost. How would game designers re-balance a game if characteristics were immutable?

Materially, a Magic: The Gathering card is a piece of flimsy cardboard. It’s only valuable because it has meaning in a highly desired card game in which Wizards of the Coast maintains monopoly power on supply. Wizards of the Coast has convinced millions of players only to recognize the “real” thing rather than someone printing out an entire deck of cards on printer paper. As they say in politics, “perception is reality”; it’s just unclear to me that ownership over a Magic the Gathering Arena card is any less “true” than ownership of a God’s Unchained card.

Central authorities govern value and, therefore, price. Centralized organizations can hold a significant degree of value governance in a blockchain-based world, and our priors should suspect they will.

Wide Open

The design space of blockchain “ownership” is vast and comes with far fewer guarantees than one might be impressed with by reading blockchain advocates wax and wane over the future. The implicit definition of ownership advanced by many blockchain gaming firms is the ability to buy and sell particular assets. Even this can be barred, however. While blockchain gaming firms have made a point to have tradeable assets, this is also open to design! A firm could tie an asset to a specific wallet address or design a smart contract that taxes the next transaction at 1M ETH ($4.3B USD). While many firms are not choosing to bar market transfers, the ability to do so is an important feature.

It’s not that blockchain advocates don’t know many of the points I’ve made above, but they do underappreciate them. As Lars Doucet has pointed out, noting these challenges does not diminish blockchain’s long-run viability, but advocates need to grapple with them.

Sorry readers, I caught the blockchain bug. When 30-person firms get $4.5B valuations, it’s time to pay attention, regardless of one’s priors. Everyone seems to have a Blockchain hot-take, but I’ve found most to fall short of ruthlessly integrating the implications and design space of the technology. Blockchain players (who are also, by definition, investors) insist, at every turn, that it’s so obvious blockchain gaming is the future. Advocates cite specific benefits – true ownership, play-to-earn, “aligning incentives between players and developers,” and decentralization. A smaller crowd expresses skepticism about what blockchain solves for gamers, and if blockchain has the implications, advocates think it does. This mounts a normative and positive element to discussions of blockchain. On cue, Benedict Evans offered a measured assessment of the situation:

Here’s the thing: I want to investigate that very notion: are there interesting technologies and great ideas buried in blockchain gaming? Over the last few weeks, I’ve spoken to several blockchain companies and studios to investigate these questions myself. Instead of chopping another log for the hot-take woodchipper, I want to split blockchain posts into examining individual “web3” gaming pieces. A subject like tokenomics demands more than a passing mention to appreciate and dissect truly. Regardless of blockchain’s future viability, it’s undoubtedly true it faces a whole host of juicy problems: token design, NFT asset design, smart contracts, auction houses, and the like.

A Win For Game Economics

For a Game Economist, answering those juicy problems hits nearly every button: monetary policy, institutional design, and property rights but doused in video game marinade. You’d be crazy not to find this stuff fascinating. But it’s also clear to me that if blockchain gaming takes off, so will the continued evolution of the designer. We’ve already seen live-service games create a class of “economy” designers and a relative decline in level design. In the type of open economy that blockchain games enable (trade is possible), game designers must start to take formal economics seriously – a large part of blockchain design depends on predicting and understanding the movement of prices. Raising equity via token design becomes a new problem for designers. While we’ve seen slow growth in the number of Game Economists, blockchain gaming will demand a rapid rise in designers with price theory training.

First up, an evaluation of “true ownership.”

A good deal of odd and folk-lore design priors float around gaming; my two favorites are free hard currency and time-limited cosmetics predicated on FOMO or fear of missing out. The FOMO model suggests developers ought to stuff their game with time-limited content, once the timer is expired the content is gone forever (or for a long time – a year or more). I’ve argued paper-thin theory holds up free hard currency, but small revenue stakes drape it as a marginal issue, whereas the FOMO model has a far more significant impact on the bottom line. High-stakes decisions demand more substantial evidence than low-stakes decisions, and we don’t yet have one for the FOMO model beyond “Fortnite does it.” Is there a persuasive and substantial case for FOMO?

Consider two battle pass models: one in which pass content is never sold after its initial introduction and another where it’s always available for purchase; Fortnite provides a real-world example. In the four years since its introduction, Epic has launched 17 (!) battle passes. A given battle pass caps spending at $160 ($10 Pass + $150 on tiers), so multiplying the passes together results in a total spend Depth of $2,720 ($160*17). As Fortnite development marches on, total spend depth grows by a tune of ~$600 a year or by about four battle passes worth ($160*4).

The effective revenue from a given player is a function of spend depth against the likelihood of “filling” that spend. If a given player spends $40 on a battle pass, we might say they have 25% spend Fulfillment ($40/$160). The spend “fill” notion helps expand on our prior model of cosmetics and battle passes.

For any given player [latex]i[/latex] at time [latex]t[/latex], we can model [latex]\text{Expected Revenue}[/latex] just as we would Expected Value, as a function of Total Value or [latex]\text{Total Spend Depth}[/latex] against the probability of the event or in this case [latex]\text{Spend Fulfillment}[/latex].

[latex display='true']\text{Expected Revenue}_{ti} = \text{Total Spend Depth}_{ti} \times \text{Spend Fulfillment}_{ti}[/latex]

If we allow players to purchase a given battle pass at any time, rather than the three-month duration, they run for, spend Depth explodes. Not only does spend Depth explode, but it grows over time. In a time-limited world, spend Depth is always constant. It’s striking to visually examine how Fortnite’s spend depth would have grown had they offered older battle passes to purchase.

Fortnite Battle Pass Spend Cap Models

A player acquired to Fortnite a year after its launch faces $640 in total BP spend cap against only $160 in a FOMO model. At a constant spend fulfillment rate, say 25%, player LTV is wholly altered; $160 versus $60 in expected revenue ($640*.25 vs. $160*.25).

The standard objection to the no-FOMO model is that there are no constant returns to growing to spend Depth for a given player. Proponents of this argument claim there is an inverse relationship between spend fulfillment and spend Depth. Instead of spending Depth staying constant at, say, 25%, it declines for each $1 increase in spend depth.

Diminishing returns to scale is an entirely reasonable objection to the model, but it’s not enough to falsify the no-FOMO model. The variable we care about maximizing, expected revenue, only declines when the rate of decline in spend fulfillment is greater than the rate of increase in total spend cap. To visually illustrate, we can model the change in expected revenue with changes in spend fulfillment.

Fortnite Spend Fufillment Against Expected Revenue

At 100% spend fulfillment, a player purchasing all 17 battle passes and tiers rings in expected revenue of $2,720. A player needs only 5.8% spend fulfillment against this total to produce $160 in expected revenue – the total possible spend under 100% spend fulfillment in the FOMO model.

Yes, the no-FOMO model may produce lower spend fulfillment, but it’s tough to imagine the rates of decrease outpacing the increases in spend depth and thus expected revenue. Imagine if HBO decided that Game of Thrones would only be available during certain weeks rather than on-demand. Overall conversion would spike during the period, but it’s obvious that’s not the revenue-maximizing solution for the content. We could certainly debate the viability of time-limited discounts or loss-leader items to drive traffic, but what the game industry is doing is paramount to book-burning: pillaging hundred of pieces of content every three months.

The Acutally Missing Out Objection

The second objection to the FOMO model is time-fixed effects. During any BP three-month period, there could be, say, 1M unique players. But fast forward a year, and a new three-month BP period may see 2M new unique players. The FOMO model doesn’t expose new players to the prior pass content, and thus they never had the opportunity to spend on it. Sorry, you never got to see Star Wars, kid, but we’ve burned all the copies.

At the very least, Fornite and Call of Duty Warzone have moved to higher item carry in their rotating stores. What was once a quaint eight items has exploded to some ~50 items.

Fortnite’s Growing Store

We’ve also seen the introduction of a vault to bring back old content.

Miss Me? Returning Cosmetic Content

Gaming firms are learning. In addition to Fortnite and Call of Duty Warzone, Valorant leans into the same concepts. Of course, Valve figured all this out years ago with limited selections of FOMO content backed by an auction house. The case for FOMO has never been strong and only weakens larger an audience is for a given title.

In-round progression completely alters the salary profile game designers pay players. If we think of rewarding player action with progression as a wage rate, in-round progression radically alters the incentives. Instead of optimizing for the long-run, in-round progression presents a 30-60min time horizon to players.

The Late Majority

Even without engaging the opposing team in direct combat, merely taking action under a game with in-round progression means to compete. In-round progression suggests that every choice is strategic. If the opposing team finds a method to progress faster, their win probability increases via increased vertical power. PvE environments like those found in Hades face a different but still game-altering change of incentives. Hades rewards players with Boons or character upgrades that persist for the duration of the dungeon run. While progression is not a race in this context, the acquisition and choice of Boons becomes a focal point for player success. And because progression resets at the end of a dungeon run, designers have the freedom to make progression in a round feel significant and powerful without fear of players hitting a level cap or out chasing difficulty.

Out-of-round progression is distinct from by in-round progression virtue of persistence across sessions. An MMO like World of Warcraft progresses players by having them gain permanent power from levels and equipment; there is no reset.

The RTS genre first played with in-round vertical power but struggled as a genre to find a way into the mainstream lexicon. If we imagine game design as a sort of “technology,” we can model the diffusion of ideas similar to technology. Following the same S-shaped diffusion curve, the RTS genre is an early adopter, and the MOBAs are an early majority.

Technology Diffusion S-Curve

The late majority is upon us. Both Battle Royale (BR) and the emerging Roguelike have seamlessly integrated in-round progression concepts into their sessions.

BR and Rougelike

An underrated feature of Battle Royale is the looter-shooter power creep across a session. Games like The Cycle take this further with explicit leveling within a round. Gear in Battle Royales explicitly contains power ratings; some guns and shields are better than others. But even beyond gear, Battle Royales are also incorporating in-round currency.

In-round progression pairs perfectly with the Roguelike genre of repeated dungeon runs. Hades has sold more than 2M copies on Steam to critical acclaim following over 1.5M copies of Slay the Spire; queue the indie paradigm shift.

In-round progression has touched an increasing number of genres to great applause. It remains to be seen how the “Laggards” will deploy the mechanic.

One of my favorite folk-lore design positions is the notion of giving away free hard currency. Without hesitation, every product manager I’ve worked with will off-handily assert the “drug-dealer” model. Supposedly, after experiencing the wonders of hard currency, players will be more likely to spend real money on hard currency rather than just enjoying the free stuff. While free giveaways may sound nefarious, it’s a common strategy in nearly all walks of commercial life: streaming services, cars, software, and Costco all employ trial mechanics without moral confusion or controversy.

Despite strong priors displayed by PMs, I’ve never seen empirical evidence supporting the claim. But more importantly, I’ve never seen the argument entirely fleshed out. For example, does the theory suppose I’d maximize conversation if I gave players $100 worth of hard currency? Or will a mere $1 of free hard currency suffice? Should I send players hard currency on a regular schedule, or is an initial allocation sufficient? In true freetoplayeconomics.com fashion, we need a model!

There are three critical variables at play:

1. The amount of free HC
2. The rate of HC sourced
3. What the HC can purchase

The Amount of Free HC

Consider a game economy with $500 in total spend. If we initially allocated $500 in hard currency to players, real money spend would drop to $0. Of course, no PM would suggest as much, but testing the endpoints helps form the shape of the optimality curve. Presumably, low levels of HC are needed to maximize conversion, but too much and conversion returns are negative. There is a global optimum somewhere in the early part of the curve; the amount of HC sourced here maximizes conversion.

If sourcing HC has a significant effect, we need to experiment to find just the right amount sourced - remember there’s a negative area to the curve. Do we know where we’re currently operating on the curve? 

Rate of HC Sourcing

Instead of an initial allocation, we could conceive of giving players hard currency on a fixed or variable schedule. After all, the theory claims that getting players into a “routine” is key to the conversion. Psychology research supposedly demonstrating “routines” are sticky, but it is a misreading of the evidence that lacks external validity. The research draws a sharp line disambiguating routines, habits, and rituals. But, turning to habit research, little of the work introduces friction into the habit formation process. In other words, how much do I need to pay players or persons to give up their habit?

Moving players from spending free currency to buying real currency presents significant “friction” into the habit; it’s not clear whether the habit is sticky enough to resist this. It’s as if proponents of the idea envision a slingshot propelling users to buy, maintaining the habit with the introduction of real-world payments. However, ever-present friction can decay the power of habits over time.

Furthermore, players and people can swap out parts of the habit supply chain. For example, if a Starbucks were to close down, a given person may switch to another local coffee shop to grab morning coffee rather than travel to a Starbucks location further away. Similarly, the habit HC sourcing creates is not spending HC but spending free HC. While I’m highly skeptical that the rate of HC sourcing positively affects conversion, it’s still a plausible lever that stands the theory on its legs, requiring falsification.

How Compelling HC is to Spend

As we’ll see in the general model, HC spending resembles a free trial. A trial allows consumers to gather new evidence and convert it to complete the purchase of a product. More generally, a trial helps potential consumers solve an asymmetric information problem. Despite a given firm’s enthusiasm that everyone and anyone needs their product, consumers have a skepticism that this information may be accurate. But the actual product plays a significant role in determining the level of skepticism, with new items or product categories meeting the most skepticism. Tesla has famously claimed it has a test drive conversion rate of 10%, far above industry norms, and showcases a wide variety of outcomes from trials. It follows that not every game would benefit similarly from a trial; instead, the effects localize to the type of product purchased with HC. We’ve generally observed that cosmetic games have lower ARPU and conversion rates, suggesting that they may benefit less from a trial than gameplay-based games, such as Battlefield versus FIFA Ultimate Team.

Towards a General Model

I model a given trial [latex]LTV[/latex] or [latex]LTV_t[/latex] as a function of the expectation gap, [latex]EG_t[/latex], subtracted from the satiation of the trial, [latex]S_t[/latex].

[latex display='true']LTV_t = EG_t - S_t[/latex]

[latex]EG_t[/latex] measures the trial’s ability to communicate valuable information about whether the consumer values the product above its cost. Again, we assume trials solve an asymmetric information gap between producers and consumers. [latex]EG_t[/latex] can take on negative or positive values; it’s entirely possible consumers come away less likely to convert after partaking in a trial. Producers should actively guard against games that don’t “trial well” or showcasing the wrong type of content. Yes, some movie trailers hurt the movies they trial for. But even if producers choose the correct content, there’s a risk of satiating players with too much product. EA Play includes a 10-hour trial program: players have 10 hours to play any title, including new releases, at no additional charge. However, trial play is disabled for games like Jedi Fallen Order, a single-player linear storytelling game from Respawn. Gamesradar estimates the title takes 14-18 hours to complete; players could complete more than half the game with the EA Play trial alone. Large values of [latex]S_t[/latex] turn the entire trial [latex]LTV_t[/latex] negative and, producers should pull these from the market.

Takeaways

I’m unsure if offering free hard currency maximizes conversion, but I have seen firmly held free currency priors with little empirical evidence to support them. If free currency trials are as powerful or influential as we think they are, they deserve to be examined and optimized.

Game monetization discussions tend to focus on what to monetize broadly (gameplay or cosmetics) as well as how to price it. As someone might imagine, these are crucial and foundational discussions to have. So naturally, therefore, it makes sense to invest a lot of human capital into optimizing them. Increasingly, however, I’ve become convinced that just making lots of stuff trumps all other optimizations. Instead of being an afterthought, supply-side considerations deserve to be front and center.

Game analysts are too focused on analyzing player data which is readily accessible and has a pre-established process for collection. Rarely do analysts venture to understand cost over revenue. Yet, it’s impossible to do cost-benefit analysts without cost. Instead, we relegate finance departments to the task. But, of course, these departments, just like analysts, only look at one side of the equation. I’ve yet to see a dashboard showcase return on spending: how much revenue did we make, on, say, cosmetic hats, relative to what we spent on them. These sorts of metrics can have a profound impact on production given that labor is fairly substitutable. It’s doesn’t cost much to have artists produce more banners than profile pictures or shift from 30 common rarity items to 10 epic rarity items. But to understand these trade-offs and make optimal decisions, cost data is crucial. Not only does this benefit ongoing games, but cost data can play a pivotal role in the initial design of the game economy.

Developers pay a lot of lip service to quality over quantity, but this isn’t a meaningful or detailed position. For example, consider if King were able to release one great Candy Crush level a month or ten far worse levels. Should King choose the one with quality over the ten quantity? Our usual model dictates the decision that maximizes net-present game LTV, and in this case, it’s the 30 worse levels a month. We should all be intimately familiar with this logic: after Game of Thrones ended, did we churn from HBO or Netflix? Which one are we still subscribed to today? Netflix makes a lot of B-grade stuff, HBO makes a few A-grade things. It’s not a mystery to which one is winning. The posturing of quality over quantity without critical examination is driven by social desirability bias. Demanding only the best, at the cost of all else, signals high status. We commonly see this in restaurant advertising and medical care discussions. 

I’ve suggested the battle pass + item shop paradigm is lackluster at monetizing effectively. Despite this, Fortnite continues to rake in billions every year despite this. The simplest explanation for its sustainability is the incredible pace of content releases. There’s always something new going on, from Marvel heroes to Marshmallow concerts, on top of a steady stream of battle passes and map/weapon updates. Building and optimizing the live-service supply chain is just as important as the demand one!

The vocal members of any community tend to dominate the feedback. And as we know, this minority is not representative of the whole. Most anyone I’ve talked to will freely admit this, but we commonly ignore it when referring to the “community” of a game. For example, job postings for Community Managers emphasize social media management. 

The overwhelming supply and velocity of social media mean it dominates qualitative feedback. Yet, the community shouldn’t only refer to Reddit posters or Facebook commenters. Instead, a community is composed of everyone who plays the game. The Community Management role ought to solicit, organize, and describe the silent majority’s viewpoints. Unfortunately, firms relegate UXR divisions to the task. As the name might suggest, UXR divisions are busy with UXR studies. Too often, it’s only large firms that can support multi-member “Consumer Insights” divisions. These divisions conduct surveys and focus groups, gathering “pull” feedback (reducing the barrier to express an opinion) rather than “push” feedback (social media).

Speed and Feedback Loops

Peer-reviewed research on Twitch streamers highlights the feedback loop that makes streaming so compelling. Streamers report the near-instantaneous feedback as a primary motivation:

However, there is a second part to these experiences, in which the “chat goes crazy.” It rapidly fills up with messages like “LOLOLOL” or other humorous phrases and emoticons specific to the game or the stream. For instance, “RAISE UR DONGERS” is a popular phrase on the streamer Imaqtpie’s stream. This feedback lets everyone share in the emotional high of the moment. In addition, it reminds everyone that they are part of a unique group of people that saw something special as it happened. Streamers also reported that viewers expected them to have a webcam so that they could share their emotional reactions to these events through facial expressions.

Any social media user can empathize with the experience of a comment or like lighting up the notification bin. It’s a minor hit of endorphins at variable intervals; short feedback loops are divine, especially with other humans. I always end up revisiting Aristotle’s original claim that we are socio-political beings as a motivating factor. It’s no wonder then you’ll find nearly all developers alt-tabbed into their game’s subreddit on a studio floor. The viral cause-effect loop means social media dominates how studio members convince of qualitative feedback. Unfortunately, delivering the exact velocity of input for “silent majorities” remains expensive. Vocal communities push feedback; for silent majorities, it must be “pulled.”

Building Strong Priors

I can’t underscore how minuscule the Reddit audience is for notable titles. Consider the below sample using publicly available Steam and Reddit PSU data. When factoring in PS4 and Xbox players, Reddit represents anywhere from .5% – 4% of players. If we scaled to all every social media platform, it likely represents less than 10% of players.

It’s Also a Biased Sample

Even if Redditors were independently drawn and distributed (i.i.d), the sample is small enough to be susceptible to sampling error. It’s not that vocal communities are useless, far from it. Instead, the developers need to understand it represents a data point rather than the data. Each data point should be weighted relative to the strength of the evidence. This starts with more precise language. For example, the “Community” didn’t say x or y. Instead, “x posts with y upvotes said this or that. This represents x% of our total players.” Context and scale matter. Generalizations can sweep this under the rug. Investing more in in-game surveys or clever ways to poke at silent majority views is desperately needed as well. 

In-game data, social media, CI, UXR, product management, design intuition, benchmarking…all help paint a picture of what’s going on and what to do next. If we want to make better games, we have to be vigilant in sharpening and understanding the tools at our disposal. 

Nope, they are mostly used to obfuscate real prices and can’t store value

— Nicola Visonà (@VisonaNicola) May 25, 2021

Even within gaming studios, I encounter a similar sentiment. Proponents of this position claim “firms use virtual currency pricing as a behavioral trick!” With virtual currency, there’s a disconnect between the real-world price and the virtual currency (VC) price. While true, there’s no intentional trickery going on. Instead, the reasons for VC are quite straightforward:

1. Improve Payment UX
2. Forgo First Party SKU Certification
3. Avoid Fixed Credit Card Fee

Improve Payment UX

The huge transaction volume of a given player unpins all three reasons. A spender might transact hard currency 10-20 times in a single day, and certainly more over a lifetime. The first-party payment flow is expensive, a customer must enter a passcode or update payment details for each transaction. Additionally, the customer receives a separate email receipt for every transaction. The first-party flow is obtuse and cumbersome, disconnecting the player from the game UX. And the cost continues to grow as the number of transactions grows.

Forgo First Party SKU Certification

First-party platforms like Xbox, App Store, Steam, etc require separate listings for each new SKU. And for each new SKU, first parties independently QA the SKU for critical bugs. Again, having to do this for every item priced HC is unreasonable. This could mean thousands of listings and QA submissions. These platforms also limit the number of SKUs, with the App Store having a ceiling of 10,000 SKUs. Roblox alone has 40 MILLION games, they can’t all submit first-party SKUs. Instead, the Robux system ensures only a couple of SKUs are required.

Avoid Fixed Credit Card Fees

Credit cards charge both a fixed and variable fee for each transaction. The fixed fee hovers around $0.10 per transaction, while the variable free hovers around 1.2%.

It’s even higher for Paypal and Square.

Consider a large spender making 50 transactions per day at $0.10 per transaction and playing for 200 days ($5 a day, $1000 LTV). In this case, 100% of the LTV goes to transaction fees, the developer doesn’t break even on a gross margin basis! Combined with the 30% platform tax, a first-party-only transactions model would set an effective floor at $0.10.

“Well, Why Not Make VC to Real World Money Exchange Equal?”

Developers try too! Many games peg their base SKU at $1 USD to 100 HC. But developers also like to pass bulk savings onto players, so they discount at higher purchase amounts.

Don’t forget that the entire world doesn’t run on USD. Pegging at nice, round multiples in one currency guarantees it won’t be the same in other currencies; exchange rates are not equal.

Virtual currency isn’t some nefarious behavioral economics trick. On the contrary, it’s a pro-player! For a look at why multiple virtual currencies exist within a given game see Optimal Area Currency with Milton. This is one of the ideas that inspired the creation of the blog five years ago!

UGC platforms have hailed the rise of the Creator Economy™. Roblox, TikTok, and Youtube have democratized the creation of content, abstracting the costs of getting content to market. But we’ve assumed UGC cuts out content gatekeepers in favor of entrepreneurs. Everyone gets a warm glow when “the little guy wins”. And by all means, this appears to be true! Creators can single handily craft and distribute TikTok videos in seconds, not days or months. The barriers to Roblox creation are higher, but it’s a far cry from the rigamarole of traditional game publishing. The effects of UGC are profound: despite easing requirements with Early Access, Steam hosts 50,000 games to Roblox’s 40 million. It’s such a gap I made this handy chart to underscore the difference:

Total Games by Platform

I wrote in The Content Problem and the Death of Level Designers that content satiation is a defining characteristic of gaming. UGC opens the content furnace to anyone. Yet I’ll argue this equality of opportunity doesn’t imply equality of outcome. The Creator Economy could be the future of consumption, but it isn’t the future of employment. This has implications for how firms payout and develop creator-driven platforms.

Payouts and Superstar Effects

My favorite interview exercise is to estimate the yearly profit of an Uber Driver. It’s unimportant for the candidate to arrive at an accurate estimate. Instead, it’s important to see if the candidate can break down a complex problem into component parts. The conclusion of the exercise underscores the revenue ceiling for drivers. An Uber driver is a rivalrous resource, only one person at a time can consume it. Digital software faces no such constraint. Not only can millions of players play League of Legends at any given time, but the game improves with scale. This isn’t replicated in the case of the 5* Uber driver. High-quality drivers can’t scale beyond ferrying a handful of passengers at a time. It’s the same, for say, barbers. Generally, labor-intensive goods have trouble scaling and are tugged along via Baumol’s cost disease.

Software unlocks the ability for laborers to scale their labor at near-zero marginal cost. Rosen points out in Economics of Superstars, “…a performer or an author must put out more or less the same effort whether 10 or 1,000 people show up in the audience or buy the book.” Adding, “the implied scale economy of joint consumption allows relatively few sellers to service the entire market. And fewer are needed to serve it the more capable they are.” Empirical evidence points to this as the explanation of why CEO pay has risen so much. Pay ties to firm size and a CEO’s ability to scale their talent across a big revenue function. And which firms have grown to be the biggest? Software firms of course!

Intuitively, it makes sense. A 1% increase in $500m is $5m, while a 1% increase of $50m is only $500k, a 900% difference between the two figures.

Not Many Winners

Economists describe unequal distributions using the Lorenz Curve, and its child the Gini coefficient. A Gini of one suggests a single individual holds all the wealth while a score of zero means everyone has the same amount of wealth. While we don’t have robust Roblox evidence, public data on CS:GO spend helps measure the Gini for F2P products. It’s reasonable to assume Roblox developer revenue faces a similar skew, if not more unequal.

Gini Comparisons

Rex Woodbury, VC and Creator Economy writer, suggested Roblox was minting millionaires:

Roblox is minting teenage millionaires: this year, it will pay out a quarter-billion to its network of 345,000 paid developers.

The S-1 filing from Roblox unravels this notion, only .0000357% of Roblox’s 7m active developers earned $100,000 or more. The 1,000 true fans model is a nice heuristic but it doesn’t appear to meaningfully scale. The theory rests on the diversity of human interests being wide enough to serve numerous and powerful niches. It turns out human interest follows wide buckets. For example, I love first-person shooters as well as World War One history. I should be playing Verdun then, right? While Verdun hits both of these categories, Apex Legends has far stronger shooter mechanics so I choose to play it. The vastness of the Creator economy is governed by the degree of segmentation of interest. The niche interest needs to be strong enough to overcome the benefits of mainstream appeal. How many niche interests can a form of entertainment serve? In the case of Roblox, not many. It’s a tautology to say $10 a month * 1,000 subs equal $100,000 in yearly income, not a prophecy.

Roblox PR has toured the case of NewFissy, a 50 person studio dedicated to the single Roblox game Adopt Me. But here’s the thing: a 50 person studio seems to be the near ceiling of the platform. The drop between the first and second most popular game on Roblox is stark. Consider the current PSU measures from RTrack:

Roblox Peak Game PSU, May 28th

Both Ran Mo and Rohit describe distributions like these as problematic. I’m less sure if these outcomes are good or bad than I am that they exist. And I don’t believe more platform revenue will spread income to the “tails”. Rising tides of platforms do lift all creators, but they don’t lift all creators equally. In fact, a given Creator’s income growth is probably tied to the Creator’s capture of platform growth. I suspect top Creators capture an increasing share of platform growth over time. Rising tides of platforms do lift all creators, but they don’t lift all creators equally. There’s a difference between saying “everyone can be a creator” and “anyone can be a creator”.

Isreal Kirzner cemented entrepreneurship as the fourth factor of production. And it’s not for everyone. Consider the income composition of Influencers, a unique form of Creator:

Hammering out brand deals is a labor-intensive task! As an alternative to variable income and sweat-based brand deals, Influencers can sign with firms. Firms help centralize and negotiate all these transaction costs. We see something similar on Roblox where freelance contractors have become studio employees. I guess Coase was right. On the other hand, does this make 50+ person game studios like those of Adopt Me part of the “Creator Economy” or just another studio?

Firm Response

UGC lets everyone compete, but it’s not clear it extends the width of the podium rather than the height. Platform earning structure plays a huge role in picking winners and losers. Getting this structure right helps minimize the principal-agent problem. If UGC platforms depend on a few Superstars then programs should be based on finding and cultivating Superstars. It’s not clear to me that “The Creator Economy Needs Fatter Tails“. Does this make billion-dollar creator funds from Snapchat or Tiktok the right approach? As they say, “it’s the incentives stupid”.

Platform payouts need to do two things: (1) tie revenue payouts to platform income and (2) beat creator opportunity cost. Without (1) the platform will lose money in the long run, and without (2) the platform won’t attract any creators. Both of these goals suggest high-stakes competition-based payouts or declining marginal income taxes as platform policy. High-stakes competitions are routinely used for infant UGC platforms, but declining marginal income taxes are less used. The idea is similar to Steam’s new revenue split for publishers – the % share to Valve declines at high revenue amounts. The model incentivizes scale and growth, exactly what we see want to see from Superstars.

Blockchain gaming companies like Immutable and Mythical use games to “prove out the tech”. This strategy places games as a means rather than an end. But what exactly is the technology to prove and what are the gains to proving it?

Games as a technology loss leader are nothing new. Steam launched under Half-Life, and Gears of War functioned as an Unreal showcase. Fortunately, there wasn’t much loss as both products were runaway consumer success stories. Improbable recently founded three game studios to prove out its Spacial OS and Facebook acquired Oculus as a technology company, not a gaming company. In all cases, gaming is/was a trojan horse for developing a technology platform. Recall that Unreal is now not only an engine but increasingly a marketplace platform. Conversely, I argued in Game Companies Are Not Tech Companies:

Game platforms have, and will always be, valueless: there is far too much game specificity and velocity for meaningful platform features to scale. 

Do Immutable and Mythical represent an exception to the rule? On the surface, it appears so. Both firms are building NFT marketplaces with potential outside of games. Blockchain-based trading is full of friction from sky-high gas fees to hour-long processing times. If they can solve it for their own games, they can scale the solution to others. Marketplaces like Immutable X promise to end gas fees and improve transaction speed. The pitch to developers is clear: list your assets on our marketplace and reap the benefits. For Immutable and Mythical, a transaction tax provides a stable revenue share model. Grow the number of transactions on the platform and grow revenue. But the open nature of blockchain threatens the entire model from working.

Card owners in God’s Unchained are not obligated to list cards in Immutable’s marketplace. This what blockchain gaming firms position as “true ownership”.

In fact, players could list the card across dozens of marketplaces. Multi-homing or listing content across multiple platforms is consumer welfare-enhancing at the cost of platform power. For example, delivery services like Uber Eats, Foodora, and Wolt share many of the same restaurant listings. Restaurants maximize audience reach by listing on as many delivery services as possible. And restaurants face declining marginal cost curves for listing on another delivery service. Even end customers multi-home by installing multiple delivery apps on their phones. This forces delivery services into aggressive and deep discounting.

Fortunately, marketplaces have more design space than delivery services. In the long run, however, what is the ceiling for marketplace fees? Apple has taken a lot of heat at 30%. Valve never has to ponder the answer – items on Steam marketplace can only be on Steam marketplace. Public blockchain marketplaces, by definition, compete for content. As a result, blockchain empowers game makers against game platforms.

Ethereum-based blockchain games allow developers to insert blocks of code called “smart contracts”. No matter where the asset travels, the contract follows. The contract is a part of the asset itself. A developer might leverage this to attach a kickback on every transaction. It could read, “Every time this good transfers, {the developer} receives 30% of the selling price.” Such code allows developers to be marketplace listing agnostic. Developers only need to focus on producing new assets and increasing the value of old ones. This is good for blockchain players who are inadvertent shareholders of the games they play.

Dominant blockchain marketplaces will emerge despite this. Uber Eats is highly profitable and Deliveroo IPO’d at $4.5B. But at the margin, content has never more been king. And this means the gains to being a content holder have increased. Consider video streaming content where TV showrunners now ink nine-figure deals. If blockchain gaming takes off the competition for marketplaces will be fierce. On the other hand, it’s never been a better time to be a content holder.

If true, this is more evidence that revenue ties to ADMC. Let’s revisit the ADMC table:

At Fortnite’s launch, the rotating store consisted of eight items.

If we add all item prices over seven day period, the total monetization cap clocks in at $334. This results in an ADMC of $47.71 compared to Battle Passes $1.90. A whopping 2511% increase!

This has only grown — the store has added 22 new slots for a total of 30. Today, ADMC is north of $300.

Call of Duty: Warzone first employed a high item carry with 15 bundles available at any given time. However, content is not sold individually but only in 4-10 item bundles. This results in higher ADMC via a higher average transaction price.

And it’s Warzone, along with CoD Mobile, that’s transformed ATVI’s growth. To be sure, explosive player numbers are the growth driver, but higher revenue per player hasn’t hurt.

AVTI MAU by Segment

Fortnite has followed suit and regularly offers both in-game bundle SKUs as well as first-party bundle SKUs.

Direct stores are the unsung heroes of the Battle Pass model, but many features remain underdeveloped. Why are items rotated rather than continually added to the store? A large store means a large ADMC. I’ve heard two bad arguments for rotating rather than expanding stores: time pressure and UX costs. On Twitter, I explained my frustrations with the time pressure argument.

The reasons physical retailers run time-limited promotions are not applicable to games. Stores run price promotions to clear inventory or as a “loss leader” to drive store traffic. There’s no virtual stock in games and every time a player opens the game they also are opening the store. If time pressure is a net revenue positive, why not make the game itself available on a limited-time basis? Proponents of a direct store need to make a better case for such a big feature.

UX concerns from running a much larger storefront are reasonable. But there are lots of blueprints: e-commerce retailers, Netflix, Spotify, Roblox… These content discovery portals spend billions testing and refining scalable UX. Why not copy? To be sure, UX costs for large stores are higher than for rotating/small stores. But it’s not enough to outpace the gains in ADMC.

If direct stores account for 66% of revenue they deserve 66% of our attention. It’s clear we have a ways to go!

The previous model of battle pass (BP) focused on average daily monetization cap (ADMC) as the key lever in driving more monetization from BP. Special attention was paid to the role of tiers and we’ll continue to do so here.

One of the more interesting shortcomings of BP is the inner temporal nature of the pass. The pass is available not on demand but at fixed time intervals. If a player joins in the middle of a twelve week season they face radically different pass economics then someone who started at the beginning of the season.

Model Extension

To understand, let’s first describe the difference between a player’s [latex]\text{Current Level}[/latex] and where they’ll be by season end ([latex]\text{Forecasted Level}[/latex]). The difference between the two is a player’s current [latex]\text{Level Gap}[/latex].

[latex display='true']\text{Forecasted Level}_{t+x} - \text{Current Level}_t = \text{Level Gap}_t[/latex]

Also consider a player with a stable goal of the level they’d like to be by churn or season end. The difference between the [latex]\text{Target Level}[/latex] and the [latex]\text{Current Level}[/latex] is the [latex]\text{Level Deficit}[/latex].

[latex display='true']\text{Target Level} - \text{Level Gap}_t = \text{Level Deficit}_t[/latex]

Finally, we understand the cost of reaching the [latex]\text{Target Level}[/latex] as the price for the number of tiers needed to close the gap.

[latex display='true']\text{Level Deficit}_t \times \text{Tier Price} = \text{Target Level Cost}_t[/latex]

A player has autonomy over this. The player can elect to spend more labor, rather than capital, to lower the explicit cost. The player will select the optimal combination of labor and capital where the cost of another unit of labor is just equal to the cost of the tier.

Battle Pass Level Production Function

We might consider many players with variable accuracy around [latex]\text{Forecasted Level}[/latex]. If a player buys too many tiers, they’ll have leftover “progression”. This is progression they otherwise would would have been able to earn with labor rather than capital. It suggests that players should wait until season end to minimize the forecasting error and thus [latex]\text{Target Level Cost}[/latex].

One way developers address this is through tier bundles. Many games offer a BP + 25 tier bundle at a significant discount. If a player plans to allocate 100 tiers worth of seasonal XP effort, then the tier bundle provides no additional benefit. To account for this, developers offer a discount on a per tier basis for the bundle. This operates as a cheap insurance policy for players uncertain how accurate their forecast of effort is. But we can take it a step further: complete pass. Offer the entire BP + 100 tiers (or something much higher than the standard 25) at season start for a significant discount. It’s like offering more expensive insurance policies to help hedge uncertain players.

Time Effects

However, we know consumption now is more valuable than consumption later – net present value exists. Buying tiers at season start gives players earlier consumption then waiting 90 days until season end. In a cosmetic economy, we know negative externalities are likely to be present – if everyone has a high level item it’s not as valuable then if fewer have the same item. Buying tiers early means high tier cosmetics are less subject to negative externalities then at season end. The model suggests tier purchasing faces a U-shape sales curve throughout the season. Tiers are purchased early for the “exclusivity” period and then later to minimize the forecasting error.

Battle Pass Tier Purchasing (Intra-Season)

But let’s again return to the player who joins in the middle of a twelve week season. The BP value proposition is far worse for them as they have less time to earn the same amount of content. Furthermore, the LTV of newly acquired players tends to decline overtime while CPIs increase for these same players. The players are likely to be more price elastic compared to the golden cohort. BP has created another odd set of incentives. The solve is to price discriminate. This is good for both players and developers! We could imagine bundling an an increasing number of gifted tiers in the bundle. The number of tiers bundled increases with the recency of a player’s season start date. It’s like resting XP….for prices.

A dash more creativity might also imagine a bundled XP booster. This would allow late joiners to converge on the current level of players by season end with the same remaining effort. Different mechanic, same incentive.

Catalog Lock

Most BPs contain exclusive, time-limited content. If you’re not around to earn/purchase it during the season then you can never obtain it. On one hand, it provides a player-developer contract. Players are protected from having a cosmetic that shrinks in value from negative externalities. On the other hand, almost no other industry does this. Even the sneaker market doesn’t function like this due to a second hand market. In games, goods are generally non-transferable or doing so has at least large transactions cost (buying/selling an account). Cosmetics are a poor store of value.

Temporal time effects re-emerge due to the total player base for a game varying overtime. Fortnite’s player base has grown, but Epic prevented itself from selling season one BP content to a much larger total addressable market.

Fortnite MAU

BP back catalog content should be available for sale either in new passes or in the direct store. This maximizes the return for a huge collection of content. It might result in lower initial sales, but win in the long-run.

Keep the Pressure Coming

For gaming to continue to increase its growth rate we need to grow monetization. A great deal of games bet on massive scale, but the market can only support so many of those strategies. We will hit diminishing returns at some point. A better BP means a better industry. I’ll conclude with a reiteration from last time:

MTX design has evolved and there’s no reason to think BP won’t do the same. At the end of the day, it’s a mechanic not a destiny.

In 1931, American economist Harold Hotelling published the seminal paper The Economics of Exhaustible Resources. Harold described a problem many firms face: how much of a non-renewable resource should they sell at any given time? This problem is more apparent when thinking about managing an oil supply but just as relevant when considering how to manage match-3 levels.

For oil firms, since supply is always declining, price should move up at every [latex]t + 1[/latex] period, holding all else constant. As a result, the optimization question emerges: Do I sell my oil now or wait for a higher price later? To solve, we first need to understand net present value or the idea that value now is worth more then value later. We can model this as:

[latex display='true']V_{today} = \frac{V_{later}}{(1+r)^n}[/latex]

Where [latex]r[/latex] is the interest rate or a proxy for opportunity costs and [latex]n[/latex] is the number of periods. All we’re saying here is that $1,000, say, 7 years from now is only worth $710 dollars today at a 5% interest rate. This is because I could take that $710 dollars, invest it, and in 7 years I’d have $1,000 given the 5% interest rate.

[latex display='true']$710 = $1000 / (1 + 0.05)^7[/latex]

The oil firm needs to compare the rate at which the oil price increases against the interest rate. The more the interest rate rises, the more sense it makes to sell oil now and invest the money rather than waiting for the price of the oil to increase. As such, the cost of oil is strongly correlated with the interest rate increase over time.

The tools of the field help us solve tricky now or later dilemmas. Hotelling’s approach and intuition helped form what is called dynamic programming economics. More wrinkles have been added to the model to solve everything from cake-eating to managing how many fish should be caught.

Consider the release cadence of television show episodes. Should Netflix release all episodes in a given season all at once, in batches, or weekly? Perhaps even monthly? The problem is littered throughout gaming: should I use a limited-time direct store or an always growing catalog store? How should I distribute rewards throughout progression? Should I save content for a UA burst? And like the oil firm, we’re in search of the content rationing solution that maximizes [latex]LTV[/latex].

Match-3 Player Level Management

Consider how King should ration Candy Crush levels for new players. King is producing new levels at some constant rate [latex]p[/latex]. But we know that players are consuming levels at a greater rate than King can make them in any given time period [latex]t[/latex]. [latex display='true']c_t > p_t[/latex]

At some point, a given player will run out of levels, the speed of which is governed by the size of [latex]c_t - p_t[/latex]. We can pick some random constants to visualize this:

We might also think of a player as having a churn probability on any given day since install. This declines over time: the best predictor of a player retaining on D30 is if they retained on D29 (decreasing convex or [latex]\frac{\partial Pr(Churn)}{\partial DaysSinceInstall} < 0[/latex]). This suggests elder players are less difficulty “elastic” or respond less to increases in difficulty.

And yet, we might imagine that the probability of churn increases every day without new levels as well (increasing convex or [latex]\frac{\partial Pr(Churn\mid NoNewLevels)}{\partial DaysSinceInstall} > 0[/latex]). In the example above, the player ran out of levels on D205.

In the long-run, King could ramp up level production, but that would eventually hit diminishing returns and has its own labor costs. As a substitute, King’s main lever to control player progression speed is difficulty. By raising difficulty, King can shift the line further right, as it would take longer for the player to reach the no content or exhaustion point. King could even tune difficulty such that [latex]c_t = p_t[/latex] and the player would never run out of levels. On the other hand, if difficulty is too hard then churn will spike (low difficulty also causes this)! This suggests that King needs to solve an intertemporal difficulty optimization problem. That’s a mouthful, but all it means is that King needs to balance the spike in probability from not enough levels against the spike in probability from increasing difficulty too much. To do so, we first need to model the total marginal effect of no new levels for every day since install, [latex]d[/latex]. This is the counterfactual churn probability of when there were levels against the current probability of churn when there are no more levels.

[latex display='true']\text{Total Marginal Effect of No New Levels} = \Sigma\left[Pr(Churn\mid NoNewLevels)_d - Pr(Churn\mid NewLevels)_d\right][/latex]

Whereas increasing difficulty in delaying this from occurring has its own cost.

[latex display='true']\text{Total Marginal Effect of Increasing Difficulty} = \Sigma\left[Pr(Churn\mid Difficulty_{x+1})_d - Pr(Churn\mid Difficulty_x)_d\right][/latex]

The solution to this system of equations is what we might call the maximum sustainable difficulty. And the intuition is similar to what was first developed for the oil firm.

A bunch more stuff falls out of the model. For instance, King should significantly increase the difficulty of levels just near the exhaustion point, just as the oil firm would dramatically increase the price of the last bits of oil. There’s a minimal downside to doing so, as the player will be in a high-probability space soon enough. Furthermore, since King is always stockpiling levels, new players face an exhaustion point further away than players who downloaded the game at the first launch. It’s in King’s interest to ease up on the difficulty for early levels since the area underneath the churn curve expands as the exhaustion point moves rightward.

There’s much more room for expansion of the model, and I encourage data scientists to play a vital role in systems design.

Professional sports give us something to aspire too. Players are celebrated as heroes and children grow up wanting to become them. It’s no secret that the internet, and games in particular, have found even more ways to engross us in the world of sports. But the terms of that engrossment are not incidental, they’re crucial. NBA TopShot lets users “own” iconic moments. FIFA Ultimate Team (UT) has players collect star footballers. Fantasy sports gives betters big stakes based on outcomes. These platforms offer us an opportunity to insert ourselves closer to the action. French start-up Sorare fuses these aspects in a way we haven’t seen before; it’s the greatest challenger to UT and fantasy sports in years (sorry PES).

Ultimate Team

Some games seem to escape monetization punditry despite wild success (GTA:V!). EA’s Ultimate Team model goes relatively unstudied compared to the newest Supercell arrivals. Yet, Ultimate Team, and FIFA UT specifically (it’s in Madden as well as NHL), have clearly tapped into something:

These sort of numbers, overall growth rate, and sustainability make UT one of the most successful live service models of all time.

The UT game loop is eerily similar to many F2P collection RPGs: collect and upgrade rosters of “heroes” that increase in power level (called OVR or overall rating in UT).

By playing against a mix of CPU and real world opponents, players collect footballers who they can use to collect even more powerful footballers. One ingredient of the UT secret sauce is the lack of an end game. Payer churn is most likely to happen when content dries up but in Ultimate Team each yearly version of the game resets progression entirely.

Fantasy Sports

In parallel, we’ve seen an explosion of fantasy sports in DraftKings and FanDuel. Players must bet on key real-world sport stars performing well against a deep scoring system.

But generally, fantasy sports doesn’t offer between bet progression in the same way that UT lets players grow OVR overtime. DraftKings has entry tokens, a store of value, but that’s hardly progression. There is some persistence in that high scoring players receive a unique cosmetic and access to exclusive betting pools. When leagues span a season, persistence emerges, allowing players to effectively GM a team. This creates multitudes of implicit side-bets. If you think a player is not likely to come back from injury his market price should drop accordingly. If it doesn’t, you can exploit that knowledge to avoid owning an asset that doesn’t generate fantasy points. Studying prices via markets and making forecasts turns out to be a pretty fun game.

The Opportunity

Unsurprisingly, the total addressable market for sports is enormous. In North America alone, media rights, gate revenues, sponsorship, and merchandising represent around $75B in annual revenue. Sports games account for about $6.6B or 11% of the U.S. video game market. This likely translates to about $1-$2 spent on sports games for every $10 spent on “core” sports activity. If we add in collectibles (cards, memorabilia) and fantasy sports, we’re likely approaching $3-4 dollars for every $10 spent on “core” sports.

Who wouldn’t want a slice? Dominant leagues, however, maintain a moat: licensing. It’s been difficult to challenge FIFA since EA maintains a monopoly on simulation style football games with FIFA and many individual clubs. Konami and PES have faced bidding wars with EA and unsurprisingly the asset has gone its highest valued use (EA’s FIFA). But if you want a monopoly, you want it where the value is. I just think we’ve just misidentified the location. The “simulation” aspect of FIFA (i.e. control a player/team via joystick) is almost tangential to FIFA’s popularity. FIFA only became a monetization and engagement juggernaut after the release of UT. It started as paid DLC and long after FIFA’s initial release.

So where is the value if not in the simulation aspects of FIFA? Consider the explosion of FIFA Companion App downloads. The app lets owners buy and sell players from their mobile phones rather than a console. Based on store download rankings, it likely does north of 35M downloads per FIFA edition. Sensor Tower only provides the last 90 days of data, well after FIFA 20’s Oct 6th release, but we can still see how dominant it is in key countries.

Based on historical gamestat data, FIFA 21 has around 45M players at this point in its lifecycle. It’s highly unlikely all Companion App players own FIFA 21, but an even conservative attach rate would be more than 60%. However, if we restrict the funnel to only those who play UT, the attach rate increases! It’s not wild to imagine that 8 in 10 UT players use the Companion App. Players are spending more time engaging with the market then they are with moment-to-moment gameplay.

FIFA PS4 Player Count by Days Since Launch

As we know, studying prices via markets turns out to be a fun game. It’s no wonder Football Manager, an entire game based around transfers, sells over 1M copies a year. Beyond a game mechanic, auction style monetization tends to align MB = MC. This blog has long argued effective monetization depends on prices reflecting player’s willingness to pay. To boot, Sorare doesn’t use loot boxes for price discovery, an inferior solution to their current strategy. Instead, Sorare direct lists players on the auction house and collects the proceeds. Players and developers have more aligned interests as Sorare’s entire revenue stream depends on the growth of the platform. If Sorare can make the cards valuable then it makes Sorare valuable and users of the platform more valuable. All of this has given Sorare’s community overwhelming positive sediment despite a top team costing over $500.

Community Reaction to New MLS Cards

But beyond markets, other key aspects drive fantasy and UT:

• Stakes (winning or losing matters)
• Persistence (more powerful overtime, build identity)
• Skill Progression (better short-run choices overtime)

The inclusion of stakes is rather obvious, but it doesn’t just have to be money. A top ranked game in a shooter is plenty tense enough. Or top Magic the Gathering matches as comparable in tension to top chess matches. The tension comes from what a win or loss would mean. It matters. Fantasy sports has this in spades, while FIFA has an extremely competitive ranked system and a strong e-sports community.

Persistence separates fantasy sports from some live service games. Many fantasy sports run over a short time period (a weekend) whereas persistence allows players to carry over the state of a prior session into a new one. Vertical progression is a literal interpretation of this.

Vertical progression or power progression means players earn more powerful and downside-free items overtime. Unlike, say, League of Legends where players earn different characters but the characters are intended to be equally powerful (horizontal progression). In Ultimate team, earning power is achieved by increasing the OVR of the team. Each player is tuned against a wall of stat dials, like this Kevin De Bruyne card.

Persistence also allows for the possibility of investment. Players play matches because it earns them characters. Higher level characters mean higher win probability. More win probability means earning more high level characters. But that entire loop may not complete in a single session. Players cycle throughout the loop, picking-off where they last left (in given OVR, currency balance, etc) from session to session. The expectation of future output based on today’s inputs is the definition of investment. Players can expect a given unit of input to pay a stream of benefits over their engagement with the game.

Persistence also lets players build in-game identities unlike, say, Chatroulette. How people think of themselves is largely a product of other people’s collective consciousness of what they think that person. A constant username, rank, icon, etc… builds relationships, identity, and a collective consciousness. MMO EVE Online, for example, has an entire history book devoted to the drama of the in-game events and player-to-player betrayals.

Skill progression/learning refers to players getting better at key moment-to-moment action/choices. It’s distinct from horizontal progression because it doesn’t necessitate the player unlocking new “choices” or expanding their in-game decision space (i.e. do I play as this Operator or this Operator in R6:Siege?). In CS:GO, this might refer to more accurate aim. In fantasy sports, this might mean signing the right players (“Goalies are less likely to get injured, so that investment is less volatile. I’ll sign a goalie”). Skill progression/learning hits diminishing returns quickly, so few players constitute the top of the sill distribution. It’s why ranked scores for players stabilize after a few games and only increase after hundred of hours of time spent. Mastering short-run gameplay is incredibly difficult.

Sorare

Sorare has a market where prices can fluctuate based on buyer and seller demand. We also know Sorare has stakes since it’s based on the blockchain and players can take money “out” of the game. If a footballer doubles in price, the player also gets real life rich. (Editor’s note: I made it through 1500 words before mentioning Sorare is on the Blockchain. It’s the least interesting thing about the platform.)

Sorare’s vertical progression is delivered via built in levels to each card. From the website:

Every time a card is played in a game week, it accumulates experience (XP). The XP will help level up the card to a maximum of 20 levels in the space of 3 years.

A card with a higher level has a higher multiple on points. If a high level card does well, it scores a lot of points. Just fielding a player allows it to become more powerful. And it’s all monetized via auction house prices.

Skill based progression in Sorare is similar to fantasy sports: make correct predictions about footballer output. In Sorare, players put together “teams” of footballers to compete for points. If a footballer does well on the real life field, the card does well and thus the player does well. It’s unclear if players will get better at these sorts of predictions overtime, but I suspect few will. Unfortunately, this anchors progression to vertical growth. FIFA shines here as mastering controls and movement returns big gains in win probability. The process of mastery is also just so damn fun. That said, a series of 3rd party tools has emerged to help players excel at the skill based aspects of Sorare, so perhaps there’s more there than meets the eye.

Closing Arguments

UT and fantasy sports are unique ways to engross fans in the world of sports. Markets, stakes, persistence, and skill progression are unique mechanics only found in games. Yet, we haven’t seen a sports game combine them in the right proportions. Sorare can be the first to do so. FIFA and fantasy sports are primed for a challenger; there’s a lot of pie to be won. But Sorare still has a ways to go. Oddly enough, monetization is solved for (auction house), engagement is a work in progress and acquisition is completely unproven. That’s the opposite of almost all start-up roadmaps. As the below chart suggests, the market opportunity will likely involve Sorare shifting a bit more up and to the left.

It’s really acquisition that needs the biggest boost, 26k Twitter followers won’t cut it. As a16z has pointed out, it’s up to Sorare to get distribution, before UT/Fantasy sports gets innovation. Do so likely rests on:

• Launching a mobile app
• Significant UA
• Signing more clubs

Other innovators are starting smell the opportunity as well, NBA Top Shot is moving to a player card model with “game” like aspects. This represents another threat and could box Sorare into football rather than expanding the platform to NFL, NHL etc. Hopefully Sorare’s recent $50M Series A will be put to good use.

In Six common mistakes when moving to live-service games and free-to-play, Ben Cousins argues that cosmetic-only monetization is a mistake:

The games that make billions from cosmetic-only economies typically only succeed because of the sheer numbers of players. On a per-user basis they actually have very poor monetization, relative to games that use more aggressive methods. This is because for a multiplayer game that is built from the ground-up to be about dominating other players, the proportion of the audience who are interested in self-expression via cosmetics is rather small.

He’s right. Traditional HD developers choose cosmetics because there are no core design implications. Designers can layer cosmetics into nearly any game at any stage of production. But for it to “work,” the game needs to achieve massive scale (low ARPU), and even still, it’s risky. It’s no wonder that very few mobile games in the top 100 grossing use exclusively cosmetics-style monetization. But I also think Ben misidentifies the challenge of cosmetics in his piece. I agree; they are certainly not about self-expression (at least the successful ones). However, male-centric multiplayer games are about domination, which is why cosmetics are viable instead of unviable.

Former “Cam Girl” Aella best sums this up by describing what it takes to get tips on stream. She writes,

Men want a few things, and probably one of the biggest is winning a competition.

You see, you’re not just trying to get a guy to pay you – you’re trying to get a guy to pay you in front of a bunch of other guys. This is a super key. A man wants to feel attention from an attractive women on him, and this is made even more satisfying when it’s to the exclusion of those around him. He is showing off his power by buying your happiness.

High-level cosmetics in multiplayer games often signal domination whether or not the cosmetic is attached to skill. Apex Legends is particularly effective at this. Consider low and high-level banners:

The stat tracker element directly shows a player’s time commitment to the game. We can also see the more exotic colors and shapes communicating danger—sort of like how the poisonous dart frog communicates danger with color.

Finisher animations are particularly humiliating since both the player performing the finisher and the one being finished must watch. It’d probably look like this if you could ever animate domination in a single five-second animation.

None of this should be shocking; we are political beings.

While these might be good observations, we need a falsifiable hypothesis about cosmetics. And to an economist, this means a model! This allows us to make more predictive statements about how a new cosmetic will or will not sell.

The Model

Consider a basic model of cosmetic demand:

[latex display='true']D_i = (C_{ij} - \mu_j) P_j T_j[/latex]

Where:

There’s much to tease out of this model. However, let’s keep it simple for this post, and instead focus on the inflation problem or what’s suggested by [latex](C_{ij} - \mu_j)[/latex].

• [latex]D_i[/latex] is demand for the [latex]i[/latex]th cosmetic
• [latex](C_{ij} - \mu_j)[/latex] is the level of differentiation of the cosmetic from the average cosmetic in circulation at that given cosmetic vector [latex]j[/latex]. Just think of [latex]j[/latex] as any customizable “slot.” This benchmarks the cosmetic against its closest comparables in the same way you wouldn’t benchmark a goalie against a midfielder.
• [latex]P_j[/latex] is the prominence of the cosmetic vector or how “featured” the cosmetic is in-game (a watch versus an entire costume)
• [latex]T_j[/latex] is the amount of time the cosmetic vector is featured on screen, both for the owner and others

There are two elements: the cosmetics level of the game and player’s individual cosmetics level. Let’s imagine a player starting a game at time [latex]t_1[/latex] and with cosmetic [latex]C_1[/latex]. During this period, the player derives a certain amount of cosmetic utility, [latex]u_1[/latex]. After earning a level, a player may unlock a higher rarity cosmetic ([latex]C_2[/latex]) and choose to equip it. They move to [latex]u_2[/latex] at [latex]t_2[/latex] as a result.

Player Cosmetic Utility Choice Model: Getting a Better Cosmetic

Now, let’s consider what happens when a player unlocks another cosmetic ([latex]C_3[/latex]), but this time the cosmetic is of lower rarity. Because [latex]u_2 > u_3[/latex] the player does not equip it.

Player Cosmetic Utility Choice Model: Getting a Worse Cosmetic

Players are benchmarking a given cosmetic against the currently equipped cosmetic in the same slot. Overtime players earn/purchase cosmetics that give them higher and higher utility. This makes it harder and harder to sell cosmetics – each one must make the player better off then the one before it. Every player in the game is going through this journey, and as such the average cosmetic level rises, just as we modeled in [latex]\mu_j[/latex]. And when this rises differentiation falls and with it quantity demanded. That’s a long winded way of saying that cosmetic inflation hurts monetization.

Horizontal or Vertical Progression

Progression is a powerful toolkit, even for cosmetics. One way to combat the inflation problem is essentially printing more money. And by this, I mean introduce higher and higher rarity. Riot, for example, introduced Ultimate II skins (3250 RP).

Alternatively, the horizontal way to attack this problem is to add cosmetic vectors. Dota 2 has mastered this with things like chat lines and ping cosmetics as customizable vectors. Of course, this too will face challenges when players collect more items in the given cosmetic vector.

Cosmetics is a hard road to follow, and inflation is always chasing developers. It’s essential to think carefully about mitigation strategies and how to grow cosmetic vectors over time as you would any other element of a live service.

Monetization’s modern paradigm is defined by a direct store and battle pass (BP). After years (and ongoing) criticism of loot boxes, Fortnite re-wrote the rulebook in a way that seems to make both developers and players happy. However, it’s important to consider that at sufficient scale any monetization scheme looks like a winner. It’s unclear if Fortnite is a winner because of the pass or despite it. For instance, the collapse of Clash Royale’s monetization can be partly traced to the introduction of its own pass.

Reports of the death of loot boxes have been greatly exaggerated as well. Of top 10 grossing free-to-play games, 8 sold loot boxes in some form. Of the top 10 premium games, 4 did. It’s going to take more work to dislodge the loot box paradigm.

And it’s understandable. If a developer isn’t going to reach Fortnite scale, battle pass isn’t a sufficient monetization solution. But it doesn’t have to be this way. We’re in the early innings of BP; by breaking down the model we can pivot the pass from an engagement to monetization driver.

The Model

We can understand BP spend depth by benchmarking it against an [latex]\text{average daily monetization cap}[/latex]. The core challenge with the pass is the relatively small spend cap. To start, consider a pass with a fixed entry cost, [latex]fc[/latex], and [latex]N[/latex] tiers at a given price [latex]y_i[/latex]. [latex display='true']{\text{total monetization cap}} = \sum_{i=1}^N y_i + fc[/latex]

In part, the maximum spend is limited by the cadence of the pass. [latex]d[/latex] is the pass length, in days. Dividing by this gives us the average daily monetization cap (ADMC).

[latex display='true']\text{average daily monetization cap} = \frac{\sum_{i=1}^N y_i + fc}{d}[/latex]

We now have a model to compare different passes. Let’s examine Fortnite, Valorent, Call of Duty Warzone and Dota 2.

Game	ADMC	Total Cap	Entry Cost	Tiers	Tier Cost	Pass Length
Fortnite	$1.90	$160	$10	100	$1.50	84 Days*
Valorant	$1.50	$85	$10	50	$1.50	56 Days
Warzone	$2.86	$160	$10	100	$1.50	56 Days
Dota 2	$6.39	$1,010	$10	2,000	$0.50	158 Days
			100 HC = $1			*12 weeks avg.

Each game has a widely different approach. Some go for more tiers and longer length while others go for less tiers and shorter season length. While not accounted for above, it is also important to adjust for pass frequency over a year long period. Dota 2 has a 200% increase in ADMC over Fortnite, but Valve only releases the pass once a year. Warzone, on the other hand, maintains a 100 tier season over an 8 week span consistently. This would average 5-6 passes a year. Nonetheless, battle pass presents far less spend depth over other monetization vectors. In a loot box system, the spend cap is the average price price to unlock all content. In a direct store, the cap is correlated to the total sum of prices divided by the rotation cadence. In almost all games this puts ADMC north of ~$20.

However, we’ve still failed to account for opportunity costs to paint a complete profit maximizing picture.

The Pivot

Marginal Pricing

BP tier pricing is plagued by inconsistency. Most games use exponential time to complete curves, with each level taking more XP (and therefore time) to complete then the prior one. Here’s Fortnite chapter one, season four as an example:

And unlike RPGs, XP earn rate does not generally increase overtime. Yet, tier price is constant at $1.50 despite increasing XP requirements and constant earn rates. This is a rather odd proposition as it suggests that the usual $1.50 tier price forgoes a differing amount time depending on the tier number. By dividing by average time to earn we can understand the cost of an hour forgone per tier number. For example, tier 100 may take 6.5 hours to earn, and with a price of $1.50 this means the player pays $0.23 per hour if they choose to purchase the tier ($1.50/6.5hr to earn). Tier 1 only takes 12 minutes, suggesting a hourly price of $120 ($1.50/0.125hr to earn)!

This incentivises players to withhold tier spending until the end of the season.
More bang per tier buck. Yet players who make it to the end of the season are the most price inelastic! Why not vary tier price to maintain a constant benefit? Here’s what this’d look like assuming $1.50 buys 6,000 XP instead of a fixed tier:

Tier 1 would cost $0.10 since it forgoes very little time while tier 99 would be $9.00 since it forgoes over 6 hours.

I chose 6,000 XP to illustrate, but any amount can be chosen as the constant. At around 7,000 XP per $1.50, marginal tier pricing produces a greater spend depth then constant tier pricing (99 tiers at $1.50 = $148.50 spend cap).

This also combats the sunk cost problem. If a player is in the middle of a long tier, the price to complete the tier remains constant in hard currency. Purchasing a tier doesn’t always boost a player to the same “spot” in the next tier. In that case, it’s easy to feel like the early XP earned in the long tier is “wasted” if the player is considering purchasing the tier outright.

Setting the optimal XP/tier price is an exercise in price elasticity. If we set a tier price to say, $10,000, almost no player would buy tier. On the other hand, a price of $0.01 would produce very little revenue. Therefore, there exists a revenue maximizing amount of XP at a given price. While we may not know this exact number, it’s likely to be higher then current prices may suggest.

Tier UX

Moving to marginal tier pricing puts much more pressure on the UX wrapper. Frustratingly, despite exponential time to complete, nearly every pass displays the distance between levels as constant. All of these levels in Apex look just as easy/hard as one another:

There are two solves for this: (1) alter tier distance (increase spacing between levels consistent with XP required) and/or (2) hard level labeling.

First used by King’s groundbreaking experimentation team, hard level labeling simply labels a label as hard (as long as it actually is). This encourages booster spend by a significant margin, as it tells players it’s optimal spend in the level. This has spread to other match-3 titles and now we find super hard levels.

Simply labeling a BP tier as hard should produce the same effect. This becomes a bit trickier for exponential curves as all the hard levels are backloaded. Shifting to an s-curve design avoids this as well as being a more equitable distribution of rewards.

More Tiers or Less Days

A simple way to increase ADMC is increase the cadence of the pass. Instead of 12 weeks to complete, refresh the pass after 8 weeks. In Fortnite’s case this would increase ADMC by ~50% ($1.90 to $2.86). We can model the marginal effects of changing the pass refresh rate:

Adding more tiers produces a similar effect but in reverse. Instead of 100 tiers, why not Dota’s 2,000? Cost and smart use of content plays a strong role.

Alter Cost Profile

Shortening cadence or adding tiers runs into supply side problems. There ain’t no such thing as free content. Furthermore, it’s wise to consider the vector with the highest return per price of content: does adding an Epic outfit to the direct store produce more marginal revenue then adding it to the BP? But even before answering that question we need to consider the cost of producing items all together. Despite Epic employing hundred of employees and engaging in outsourcing, ~12% of tiers in the pass are “costless”. This means they consist of either currency or boosters rather than distinct items. 

Having cheap items altogether is another avenue. Apex has mastered this. Consider the stat tracker: it’s simply a tally of a particular stat. Or loading screens: an old piece of 2D art. This type of content is extremely cheap to produce. The foundational conversations of a game economy should include content vector specifications.

Valve has employed clever use of changing colors on cosmetics and Fortnite adopted something similar in their new crystal levels.

Where To

MTX design has evolved and there’s no reason to think BP won’t do the same. At the end of the day, it’s a mechanic not a destiny. In future posts, I’ll expand on pivoting the pass even more. This is just an appetizer.

Modern live-service games have self-segmented in genres: match-3, 4x, collection RPG, battle royale, etc. We know these genres evolve and start to incorporate new mechanics. Over time, these mechanics become standard genre fare. For example, invest-n-express titles like Gardenscapes are an outgrowth of the match-3 genre, adding collection mechanics to the core match base. In HD, we’ve seen innovations like Apex Legends’ revive mechanic modified in Warzone’s Gulag – players fight for revival in a 1v1 mosh pit. But how could we better understand why game genres change rather than simply observing them change? I argue Thomas Kuhn can help.

From the Stanford Encyclopedia of Philosophy:

Thomas Samuel Kuhn (1922–1996) is one of the most influential philosophers of science of the twentieth century, perhaps the most influential. His 1962 book The Structure of Scientific Revolutions is one of the most cited academic books of all time. Kuhn’s contribution to the philosophy of science marked not only a break with several key positivist doctrines, but also inaugurated a new style of philosophy of science that brought it closer to the history of science. 

Kuhn’s task in the Structure of Scientific Revolutions is to explain the process of scientific change. He proposes that at a given time, in a given field, there exists a “paradigm.” Paradigms serve two functions: a framework for asking questions and the tools to solve them. It’s helpful to think of genres in the same way: they are frameworks for answering key questions to engage a particular Bartle-type. MMOs have different mechanics and frameworks to engage MMO players, whereas match-3 games have their mechanics to engage match-3 players. But that’s a levels explanation, not a change explanation. We still need to explain why paradigms would shift or change.

Paradigms shift when the current paradigm cannot sufficiently account for key emerging questions. The complete and utter destruction of the RTS genre to MOBAs serves as a case study. Consider the RTS genre to answer the following question: what mechanics do I use to create deep and long-session strategic gameplay?

Company of Heroes and Starcraft certainly had answers, but MOBAs challenged those answers. Are complex multi-unit controls necessary to create that sort of gameplay? MOBAs told us that, no, multi-unit controls are not required. Can teamwork create more profound strategic experiences ever? Yes, says MOBAs.

It’s too perfect a parable when we further consider that MOBAs were an extension of RTS to begin with. The original DOTA or Defense of the Ancients was a Warcraft III mod, and this is the exact type of model drift Kuhn describes.

Next, Kuhn would predict, comes a model crisis. As MOBAs started to siphon players from RTS, the genre had a chance to strike back and incorporate these elements. From a Kuhnian perspective, failure to do so would result in its death, yet incorporation was attempted and resoundingly rejected. Dawn of War III tried to add strong hero units with deep in-session progression. Relic misunderstood the questions MOBAs were trying to answer and copied the wrong elements. Perhaps more importantly, RTS never had a full free-to-play answer; Starcraft was too little too late. Kuhn would call the period of conflict “revolutionary science” and the MOBA victory a paradigm shift.

Dawn of War II and DOTA 2 Steam PSU

However, it’s not always the case that incorporation will fail. Hearthstone rapidly created Battlegrounds to repel the threat of Auto-Battlers. It seems to have worked, but the lack of monetization in Battlegrounds leads to year-over-year revenue declines.

The Kuhnian model suggests framing a given game under a “game theory.” What is the game trying to answer in the genre it’s fighting in? What are the existing answers (if any)? Why is it worth answering? It’s instrumental for teams to consider these questions. Generally, pitches I’ve seen describe the game’s vision rather than the game as a back and forth conversation with market participants.

Tech has had a better understanding of the paradigm process. The near frictionless mass distribution of software means the paradigm process moves at light speed. Clubhouse generated huge mindshare, now everyone from Slack to Linkedin is trying their hand at audio. We saw the same movement with “stories,” an original Snapshot feature copied by Instagram and Twitter. Time will tell if their incorporation process answers the same questions Clubhouse has.

In the 60+ years since The Structure of Scientific Revolutions, philosophers of science have challenged a good deal of Kuhn’s original ideas. Despite this, Kuhn provides a rich model for why some game genres live and others die.

The “metaverse” discussion seems more about cultural “in-group” signaling then a thorough exploration of an idea. It’s frustrating. Rather than talking about “what the metaverse will look like” instead we should examine the forecast of a “metaverse” altogether.

Most commonly, I’ve heard it referred to the idea that the future of games is a single shared 3D world with a variety of experiences driven by user-generated content. Roblox is most often associated with the term. Yet, what evidence is there that games are converging on a single shared 3D world?

Roblox and Fortnite are the best and biggest data points. Roblox is a successful game-as-a-platform unlike LittleBigPlanet or Halo’s the Forge. Those titles and ones before them are best thought of as mod platforms, wherein players are creating extensions of the core experiences rather then entirely new ones. On the other hand, Roblox allows for vastly different experiences, from FPS to RPG. Roblox also allows creators to earn cash in a straightforward and consistent way, uncommon for mod platforms. Roblox gets the “game-with-a-game” billing rather then marketplace because since it’s held together by a thin layer of: a launcher, avatar system and unified virtual currency (Robux).

Roblox has undoubtedly been a massive hit with over 40M DAU, but it’s only captured a specific age demo. As players mature, they fall off the platform. This isn’t the direction you’d like LTV curves to take. According to their S-1 filing, 67% of users are under the age of 16 while 14% of Roblox’s users are over 25 years old. Roblox’s inability move up the age ladder is largely attributed to its childish cosmetics and unpolished playgrounds. 

The more devolution Roblox gives to developers to solve these problems the less it means to be a “Roblox game” since there’s less in common between each title. Devolve enough and you’re just Xbox Live. We’re starting to see evidence of this as developers can override the avatar system.

But poking and prodding the definition of metaverse reveals a series of questions. Why isn’t something like Xbox Live a metaverse (persistent avatar, friends list, social hangout)?

• What are the player benefits of a 3D shared world rather than a 2D menu?
• Why did PS Home fail?
• Is Second Life a metaverse? 
• If so, why did Second Life fail?
• What observable things will happen in the next 2, 5, and 10 years that suggest we’re on track for a metaverse?
  • How do these things differ from more UGC game platforms?

The metaverse is a forecast. Proponents should be more forthcoming in laying down testable hypotheses.

Prima facie, PS+ faces an uphill battle – developers receive near $0 for offering up their games. Additionally, any users brought in on PS+ incur additional server costs for the developer. Should developers really be so willing to throw their titles at the service?

The service charges players ~$5 monthly to receive two random games. These are full entitlements, not subscriptions. A player has an entitlement to them as long as they remain on the service. The billed benefits of the program are:

(1) the additional sales generated from other platforms during the PS+ period
(2) the “momentum” generated after the PS+ period to additional sales on Playstation
(3) MTX revenue from additional players

I wanted to examine expected PS+ player volume as well as sales uplift afterwords. Examining cross-platform effects are possible if the game is cross listed on Steam, but I’m lazy. And I have written enough code. So I’ve posted the spaghetti code for others to do so (and spot my errors!).

The Dataset

gamestat uses the Playstation API to estimate player counts. This is particularly apt for PS+ analysis as it excludes players who “claim” a game without playing it at least once. The dataset only includes “played entitlements”. We’ll cross-reference this against a list of PS+ PS4 games. Occasionally there are regional differences but I selected the North America list for simplicity and personal bias. Remasters or collections are attributed to the base titles and are not parceled out in API data.

The Growth of PS+

The subscription has grown overtime, consistently attracting more and more downloads for the average title. Fall Guys didn’t just outperform prior titles, it obliterated past comparisons.

Hall of PS+ Fame

PS+ Played Entitlements Per Title

In 2019, PS+ exploded with the median title having 2.46M downloads compared to 2018’s 665k. This only increased in 2020. Some of this is probably fueled by Sony dropping the number of monthly titles from 6 to 2 in February 2018 (thereby increasing download concentration) as well as picking titles with stronger Metacritic scores.

Additionally, the service moved away from indie darlings towards titles with more publisher heft. Past year’s Call of Duties have found a second life in the program. This success has only attracted more “old” AAA titles to the program. We also find titles previously released on other platforms choosing to launch on Playstation via PS+ such as Grow Home and Super Meat Boy.

Weeks After First Playstation Launch to PS+ Inclusion

But heighted downloads are not enough, developers need revenue upside especially when considering Sony limits the ability of PS+ launched titles to appear on other platforms. We need to consider increases in PC sales and post PS+ uplift in PlayStation sales.

Is There Post-PS+ Uplift?

At quick glance, Fall Guys shows a shape decline in entitlements after the PS+ program ends. That makes sense when a title goes from $0 to $20 for a large swath of players. However, there is a “handover” month in September that looks highly correlated with the prior month’s PS+ numbers.

Fall Guys ‘Played’ Entitlements

On the other hand, the effect is only predictive of sales in the waning days of the PS+ period wherein downloads numbers are smallest. Additionally, the lift is short-lived (~2-3 weeks after the PS+ period ends).

If we examine all titles, we find a PS+ “hump”: entitlements increase during the PS+ period and then decline.

Monthly New Entitlements Per Title: Before, During & After PS+

For a clearer picture, let’s connect entitlements 1 month before and after PS+ for those games with a non-PS+ launch period (i.e. excluding Fall Guys). We should see upward sloping curves if there’s a post-PS+ lift and indeed that is the case.

Monthly New Entitlements Per Title: 1 Month Before & After PS+

Even after all those free entitlements, there’s a still a good deal of players who purchase the game.

The mean uplift was +277K entitlements when comparing the pre-PS+ month and post-PS+ month. However, there’s a whopping standard deviation of +286K entitlements, with a median uplift of 160k+ entitlements.

Not all PS+ uplifts are created equal it appears. The biggest winners were already big franchises, further riding off of brand awareness and possibly some networking effects with so many newly added players. NBA 2K 2016 was the top earner, with over +1M entitlements compared to it’s pre-PS+ haul.

Adding It All Up

What’s not accounted for are players that downloaded the game during PS+ and only first played after the PS+ period. I’m going to guess and say that accounts for 20% of the entitlements.

We also need to consider that many games pair price sales during and after the PS+ period. I’d reckon a $30 average selling price of which $20 goes to the publisher after platform margins are accounted for. For every 100k units, this works out to $1.6M net marginal revenue before severs costs are subtracted or MTX is added. PC uplift isn’t accounted for and neither is the opportunity cost of having to be PlayStation exclusive for a period of time.

The best I can tell, PS+ revenue uplift isn’t lifechanging (know your OCOR!) and the scale of the uplift is correlated with how well you were doing anyways. That said, I can’t find a decrease in post-PS+ sales and that’s good for developers nervous about giving away their game for free.

Rushing development feeds into narratives around greedy firms. “If only they didn’t want so much money!” Much of this banter is comprised of cheap shots devoid of making real claims about what CDPR should have done. Should the game have been delayed an additional 4 months? 6 months? And if so, why? If the board really didn’t understand the scope of the bugs then the question turns to the organizational design CDPR. What organizational breakdowns led the lack of information the board had about the bugs in game. Were QA leaders not empowered to speak up or not trusted?

These are much tougher questions to answer. After all, as Pixar is fond of saying, “[Games] don’t get finished, they just get released”. The key question is when to release. There will always be bugs and there will always be new features to add. Ultimately, release timing is a cost/benefit decision. Relative to the additional development cost what increase in sales would we expect from a delay? Do we have ever higher margins from a 4 or 6 month delay? To be clear, Cyberpunk is already outselling all other CDPR games, hardley “one of the most visible disasters in the history of video games“. What further increase would analysts expect with what additional delay?

I love the way Ran writes https://t.co/3XUxfsXjDL

— Growf (@pinchedforgrowf) December 22, 2020

The article’s writer, Ran Mo, is a Lead Product Manager at EA according to his Linkedin while the quote Tweeter is a fellow product manager. Dive a bit deeper and the retweets are all from fellow VCs or product people (guess there’s something to this).

This isn’t exactly uncommon. The Deconstructor of Fun podcast is hosted by three Product Managers and guests frequently come from a similar ilk.  A scroll through the last 20 DoF blog boosts a breakdown dominated by PMs.

DoF Posts by Job Discipline

Games are at the cleavage of art and science, so why are PMs the only ones with something to say about it? The alternative voices we do have, Eric Seufert (UA), Alexandre Macmillan (Analytics), and Javier Barnes (Design), only take a couple of sentences of digestion to realize the dramatically different way they frame and discuss problems. Their pieces tend to have more backbone or a strong theory that underlies an empirical observation. I’m a fan of this approach.

PMs are driven by their social caste, mainly moving up it. Networking is crucial to this, an insight that seems to go over the head of analysts and designers (at our own peril). The PM hierarchy is reflected in the “up or out mentality” re-enforced at tech and gaming firms. A scroll through PM Linkedin and you’ll see the following ladder:

None of these motivations discredit, in any way, the strength of the ideas expressed by PMs. Or the fact they actually take the time to express them. But it does help explain why they can feel hollow at times, trying to fit a socio-political mold rather than a genuine expression. This is reflected in how many game PMs will depart for higher paying tech PM jobs in the Valley or to fellow gaming firms for title bumps. And there’s nothing necessarily wrong with that.

 If I had a plea, it would be for all game disciplines to write vigorously. Write everything you know to be true and let’s hash it out. The game craft is too important to be dominated by one discipline. We should all be thinking hard about these problems.

It’s the incentives, stupid.

Players want to unlock content and the most efficient way to do so is to maximize how many FPS games control progression speed: SPM or score per minute. Score is usually a formula composed of objectives and kills. The key is that it’s uncapped: there’s not a fixed amount of XP up for grabs in a given match or time played (this would be a better design). The formula implies that the more “action” in a given minute of gameplay then the more score per given unit of time and the faster a player will progress. Small maps excel at encouraging this – there’s a short amount of time before you bump into an enemy or objective.

FPS players like small maps because they function as costless XP boosts. Nuketown will be making its 5th appearance in the CoD title with Cold War.

A Quick Refresher

Pricing is challenging to get right. Ideally, we’d like to select a price that maximizes profit while keeping everything else constant.

In the above example, we consider a single price [latex]P[/latex] but we can expand this to consider the set of prices or price set that define a game or service [latex]P_s:{{P_1, P_2…P_x}}[/latex].Again, we want to pick the set of prices that maximize revenue.

Many goods or services operate under a single fixed price. For instance, a book might cost $18—everyone who values the book $18 and above purchases it. [latex]\text{revenue} = n \times P[/latex] where [latex]n[/latex] is the number of readers who value it $18 and above and PP is price. Simple enough, right?

Let’s add consumer surplus to the story. In the example below, this is the shaded yellow area. Some readers value the book at $30 and thus are the most profitable in the transaction ($30 – $18 = $12 economic profit). The other readers made out, but perhaps not as well.

Imagine, however, that we could charge two different prices to two different segments of readers. Say one reader valued the book at $18 and another at $30. The trick is to ensure that the good is non-tradable; otherwise, the customer who faces a lower price could resell to the higher-priced customer and pocket the difference.

While we could raise the book’s price to $30, we’d lose out on the $18 reader. The problem of price discrimination, the one described above, is fundamental to understanding entertainment business models.

Our Case

The consumer surplus model struggles to capture utility. After paying a fixed price for the book, the reader chooses to consume it (is anyone surprised?). The reader continues to accrue utility from this consumption—the enjoyment of reading the book outweighs other activities. At some point (but not always), the accumulated utility outweighs the initial cost.

Standard Utility Curve (Reading a Book)

It’s important to note: the reader is not “in the hole” when we see red in the above graph. The book is a sunk cost; the reader should only consume it better when it’s better than engaging in other activities. But if we extend this graph to include motimeime, the curve kinks immensely.

Utility Curves: Books

There are incredibly steep diminishing returns to reading a book a second time. It’s boring. Deep diminishing returns explain why book rentals (libraries) and reselling are so popular. Why pay a fixed price for what is usually a single-use item? Steep diminishing returns ring true for movies as well. Of all the films you’ve watched, what percent have you seen a second or third time? 1%? 5%? Subscriptions and rentals make sense for this form of entertainment. But what if, instead of flattening, the utility curve grew? Enter gaming.

Utility Curves: Games & Books/Movies

Games have a unique resistance to diminishing returns. As described in Part III:

The genius of PvP (Player v Player) environments is how they necessitate the emergence of a meta-game. PvP environments resemble game theory models where it has been shown strategies evolve in an evolutionary process. In mathematics, Player vs. Environment (PvE) resembles optimization where strategies are static – one and done. Each balance change reshuffles Equilibrium in PvP environments; the search for dominant strategies in an ever-shifting equilibrium is the game itself.

The strategic evolutionary process is a near-limitless piece of content to consume.

MB = MC

The efficiency of any monetization or pricing system is the degree to which it can correlate marginal cost (MC) to marginal benefit (MB). In the above examples, we fixed the price. Fixing the price makes sense, given that the utility curves flattened out. But the more the curve refuses to flatten, the more decorrelated MC to MB becomes, the longer it continues.

The above leads us to the emergence of DLC and MTX. Players were playing PvP titles for hundreds, if not thousands, of hours, and MC failed to catch up. DLC map packs like those in Call of Duty and Battlefield helped MC catch up (and grow MB!) in fixed intervals, but the correlation was still weak, and time persisted.

MTX solved for the explosion in the marginal benefit that multiplayer games were providing. Unlimited or greatly exaggerated spend caps allowed players to spend closer to their MB curves than they were previously able to do so.

Utility/Price Curve: MTX & F2P Games

Software as a service (SaaS) can generate similar growing utility but only charges a fixed price at recurring intervals. Again, this suggests that subscriptions might make sense for games. Subscriptions are better than fixed prices in correlating MB and MC, given that SaaS generates recurring homogeneous LTU returns. However, games generate a more heterogeneous LTU (lifetime utility) than many SaaS products.

We can model the heterogeneity as such:

Considering the total Lifetime Utility generated by a standard good or game, we add up an individual’s LTU from the lowest to the highest LTU. If everyone valued a standard good the same, LTU would be linear. If a few players valued a game at a relatively extreme LTU, you would see a bowed curve – the high LTUs skyrocket total LTU upon addition. Look familiar? A bowed LTU curve correlates with observed LTVs in F2P games.

Does Tech leave too much LTU on the Table?

But that’s not to say all non-video games have a linear Total LTU curve. Some users value Zoom more than others. As Zoom usage increases, LTU also does, but the price does not. Zoom fails to capture a great deal of LTU from high-usage customers. MTX theory could offer a hand.

Zoom does offer tiered pricing for organizations, with a higher price charged for additional features, but this doesn’t capture the high LTU users within an organization. Perhaps Zoom should get into the cosmetics business—backgrounds were a fantastic opportunity never capitalized on.

Even at the organizational offering level, Zoom could create an additional tier that offers other features in addition to the high-usage tier. Multi-track cloud recording, for instance, generates incredible value for Podcasters but costs nothing further.

Standard goods and games’ dramatically different value propositions necessitate other monetization schemes. Different content economics make applying the monetization of tech companies less applicable to gaming. Instead, SaaS firms have something to learn from games.

Gaming Companies Are Not Tech Companies!

So maybe game companies aren’t tech companies. As much as game companies seem to borrow from tech firms, tech firms run a bigger deficit in borrowing from games. If Netflix, a subscription service with over 10,000 movies and TV shows, has its biggest competitor in a single game, Fortnite, then perhaps there’s more for tech to learn from games. How games deal with the Content Problem is their defining characteristic. Of all forms of entertainment, games present the most compelling answer to the problem.

The Content Problem

The fundamental axiom of economics is unlimited needs and wants, and only limited means to fulfill them. The parallel for entertainment might consider that core content demand nearly always outstrips supply. For example, large swaths of Game of Thrones and Harry Potter fans are underserved by a couple of books, movies, and TV seasons. Executives try to fill the void with licensing: Harry Potter backpacks, Game of Thrones beer, etc. But filling the core content demands is impossible: it takes more than 1 hour to produce 1 hour of Game of Thrones, while the same is not valid for games.

Consider the following:

The content consumed in a game like Overwatch or Clash Royale is the pursuit of strategy equilibrium and/or mastery of mechanics. A new unit in Clash Royale changes how players organize their decks, even if they don’t use the unit directly (they must counter it). The new units provide hundreds of new hours of content to consume relative to the near 1 man-week of labor to produce the new unit. Therefore, the marginal content output of a given member of the 16 people (!) Clash Royale team is astronomical.

Furthermore:

The genius of PvP (Player v Player) environments is how they necessitate the emergence of a meta-game. PvP environments resemble game theory models where it has been shown strategies evolve in an evolutionary process. In mathematics, Player vs. Environment (PvE) resembles optimization where strategies are static – one and done. Each balance change reshuffles Equilibrium in PvP environments; the search for dominant strategies in an ever-shifting equilibrium is the game itself.

The marginal product of labor for a given game developer completely outclasses a given producer on a movie or TV show by the medium, not the individual. Unlike games with infinitely replicable assets, films and TV face fixed constraints: Emilia Clarke or David Benioff can only be in a single place at a given time. They must also eat, sleep, and socialize (sigh). Meanwhile, Captain Price faces no such constraints. There’s no more Game of Thrones to consume after the last episode cuts to black, while there’s always another hour of Fortnite to play. How can Netflix and others adapt to the reality of these media?

The most straightforward strategy is a content arms race. Netflix spends over $17B annually on original content while scooping up oodles of back catalog content. Of course, viewers must be interested in this content to be “effective,” and the recommendation engine plays a vital role. But the last episode of Stranger Things shows that the recommendation engine cannot fill the void while operating on the same indifference curve. The “more bodies” strategy to solving the content problem is expensive to execute and, as we’ll see in part 4, struggles to achieve Marginal Cost = Marginal Benefit.

Reality TV is a response: it needs fewer writers, editors, and CG to produce a given hour of content. Shows like The Amazing Race, Big Brother, and Survivor can do 20+ seasons of 22+ episodes, while Game of Thrones struggles with seven seasons of 10 episodes despite having so many more crew members. Netflix’s speed of investment here is breathtaking. However, the addressable audience is more limited in scope than traditional dramas. Netflix needs a bolder evolution to combat games: TV-as-a-service.

The forgotten genre of soap opera TV provides a near-perfect blueprint. Soap operas are year-round, weekly, serialized television shows for those unfamiliar. The unrelenting pace has resulted in popular series like General Hospital having 14,000+ episodes over 57 years. Netflix needs to invest heavily in moving performances to a similar format: year-round production with weekly releases. There’s always another piece of content to consume right around the corner, while the back-catalog for a given show continually expands for newcomers. In many ways, this mirrors match-3 level production. The number one reason why players churn from match-3 is a lack of new levels, and a glance at community pages confirms this.

King mitigates content churn by using an increasing difficulty curve, so it takes players longer to reach content exhaustion as they progress through levels. Another strategy is also possible, however: branching narratives. Reality TV does not have such an option.

Increasingly, Hollywood is shooting movies back-to-back. It’s cheaper to continue production rather than stop and go. Why not do something similar to produce more content? In this case, multiple perspectives in a given series are shot simultaneously. The Lord of the Rings production operated similarly with two production crews but with a singular end product. Game of Thrones operated with two film crews, but the end product was a single episode. Why not dedicate an hour to each perspective rather than splice the two in a single episode? A multi-perspective production multiplies a 10-episode season to 30 while holding down cost. Netflix can’t solve the Content Problem, but it can mitigate it.

Interestingly, YouTube has solved most of this problem via a two-sided marketplace. The smattering of volume helps the supply-side problem, even if a particular creator has a finite number of videos (remember, you can still play a given game for unlimited time without “running out” of content). YouTube has encouraged users to subscribe to many creators, accounting for a regular release cadence.

Diminishing returns for linear content are incredibly steep; few users will watch a film or movie more than once. Increasing and prolonging a viewer’s LTV is most elastic with more content, a costly proposition. To compete with games, TV, and movies, technology and business models need far more supplies. If technology and business models can change innovative products rather than be a vehicle for them, now is never a better time to explore changes in storytelling.

Part IV

Part I

Ran Mo retraces the incredible fast follow of the auto-chess genre. Before you could blink an eye, what was a mod in Dota 2 (which itself was a mod) had spun off into three incarnations. The first was from the mod maker, Drodo Studios’ plainly labeled Auto Chess, the second was Valve’s Underlords, and the third was Riot’s Team Fight Tactics. It’s not clear any of them have done “well.” Indeed, none of them are making any real money as there’s little to purchase amongst any of them.

Valve’s Underlords PSU Declines

But if there is a leader, it’s Riot’s Team Fight Tactics (TFT) by content cadence (from a player count perspective, it’s unclear if TFT is doing something meaningful). How do we explain the last entrant who is now leading the pack? Ran seems to argue for the power of the platform:

“Underneath the hood of League of Legends is a deep set of systems and tools: messaging and voice chat systems, social systems, matchmaking, internal development tools, event systems, cloud-based server infrastructure, post-game data analytics, and so on. These tools enhanced play experiences and supported the continuous release of contents, events, and updates.”

Ran further explains, “We’re likely entering a new wave of consolidation today, with new moats centering on live-service ecosystems […].”

I could not disagree more. Not simply on Riot’s platform power, but also in the broader value of game platforms altogether. Game platforms have, and will always be, valueless: there is too much game specificity and velocity for meaningful platform features to scale. “[…] messaging and voice chat systems, social systems, matchmaking, internal development, I could not disagree more. Not simply on Riot’s platform power, but also on the broader value of game platforms altogether. Game platforms have, and will always be, valueless: there is too much game specificity and velocity for meaningful platform features to scale. “[…] messaging and voice chat systems, social systems, matchmaking, internal development tools”? While standard and required for live service, they are not “deep.” If these features form a moat, it’s the equivalent of a kiddie pool defending Versailles. The ever-growing number of PC game launchers, all of whom do these exact things, suggests as much. The primary value of game “platforms” is discovery and perhaps security.

The deeper platform features of Steam (Marketplace and Workshop) have failed to see widespread adoption. Other platforms haven’t even built anything beyond a friends list and chat. None of them can justify a 30% cut, and for many games, “multi-tenancy” or listing on multiple platforms is the clear revenue-maximizing strategy. Of the top 20 Steam games, only 35% of non-Valve games use Marketplace or Workshop. Furthermore, 59% are multi-tenant or are on other PC launchers. Expect this to increase as Steam degrades as the sole game discovery launcher. Activision realized this long ago when they shifted Call of Duty to Blizzard.net and away from Steam. Blizzard titles anchor the launcher with eyeballs but do little feature work. It’s 2020, and the Blizzard region still locks many titles, meaning switching from NA to EU servers entails losing all progression for some titles.

Other firms take game platforms to eat at specific parts of the game “supply” chain. In addition to a given piece of technology working for games, these firms hope the tech will also scale outside of games. This is an executive fetish that rarely, if ever, pans out. Machine Zone (MZ) is the latest failed entrant into “but we’re really a tech company”. Cognant ‘s internal ad-buying platform, Cognant, and its cloud platform, Satori, went up in flames. MZ ended up selling to AppLovin at 10% of its peak value. The jury’s out if Improbable, the latest “but we’re really a tech company,” can dig out of $85M losses. A more considerable hope is Unity, but even then, only 8% of the 716 customers to spend $100,000 or more have been non-gaming firms—games as a technology, an executive fetish that rarely, if ever, pans out. Epic’s PR team was out in full force, tracking The Mandalorian using Unreal, but this is a minuscule share of Epic’s revenue. Games have similar challenges to tech firms, but tech firms rarely have identical challenges to games (all squares are rectangles, but not all rectangles are squares). A common theme of this blog, reiterated in this series, is that games are a distinct and unique medium. The faster we embrace this, the quicker the industry can evolve.

While game companies may never be tech companies, there may be a small place for game technology companies. Unreal and Unity are real and here to stay—they solve complex and prevalent problems game makers face. The total addressable market for game tech solutions grows if the games industry grows. Is there room for more game technology companies beyond Unreal and Unity? The previously mentioned Improbable seems to think so, but little additional evidence exists.

There’s more to be said for a game as a platform, but the evidence is still scant. Publishers haven’t been able to retain the value of extensions within a given title (Auto Chess splitting from DOTA 2) or outside of it (MOBAs splitting from Warcraft III and Heroes of the Storm). The simple ability for players to build content within games hasn’t precisely accelerated either. If anything, modding has become more challenging, not easier, since its ascent in the early 90s. Can we think of a single major 2020 release to embrace mods? Much was made of Halo 33’sForge, but we’ve seen few copycats since its release in 2007. If we consider a corollary to Bill Gates’ definition of a platform, it might go something like this: “A game is a game platform when the playtime of platform content eclipses the main game.”In this view, there hasn’t been a modern game as a platform. A softer version might consider a game a platform when “he playtime of additional modes eclipses the main game.”Despite a much lower bar, it’s not clear that Fortnite, with its concerts and social spaces, has reached it either.

We’re left with Roblox, a trustworthy game platform. Roblox provides valuable tools that make it easy for developers to create compelling games and facilitate discovery. But the jury is out on whether to consider this the future of games, especially with its elder sibling, Manticore, going nowhere fast. 

RPGs vary from CCGs, which vary from FPSes. FPSes might need to solve 64-player servers, CCGs may need transfer markets, and RPGs need deep customization systems. This specificity of these challenges shrinks the total addressable market for the tech solutions that game firms have devised. Central platform features haven’t adapted to a particular game design (and the speed at which it changes). As a result, we’ve rarely seen game companies make a successful transition to a tech firm.

Part III

Ran Mo and Joseph Kim (@jokim1) argue that games should be viewed from a Silicon Valley perspective. The usual three appear: moats, networking effects, and platforms. Moreover, while I thought it had died out after the launch of Halo 3, an inferiority complex continues to exist amongst game makers. Games never seem to get the mainstream or broader tech circle legitimacy that many think they deserve. Despite operating in the Valley and major tech hubs, game companies do not reach the crazy evaluations of FAANG (ATVI has a market cap of $63B, while Facebook sits at $215B). Furthermore, public intellectuals like Tyler Cowen or Ben Thompson rarely discuss games as a part of their commentary (not with Matthew Ball, but maybe he drinks too much Kool-Aid. Even internally, game firms seem to be chasing subscriptions on the heels of Netflix and Spotify.

The fundamental problem comes from not respecting or understanding games as a distinct field. The fundamental problem comes from not respecting or understanding games as a different and unique medium separate from linear content. Kim’s piece wants to make this point in many ways, but does not go for the kill by the end of the article.

I’m writing some posts to address and expound on this. In two parts (networking I’m writing a two-sided series to address and expound on this. In two parts (networking effects & platform power), I’ll examine why game companies fall short of traditional tech companies. Then, with another two parts, I’ll address what tech companies have to learn from games (the content problem & MC = MB).

Networking Effects

Networking effects describe the positive externalities from the n+1 user to a service. Every time someone joins Facebook, the benefit increases as others can interact with that person. Gyms, for instance, work in the opposite direction: every member who joins a gym occupies a fixed amount of capital and decreases the value to each other member.

Fawning over networking effects comes from the path dependency inherent in the model: more users lead to more users. What does VC not want, sis, self-perpetuating growth? However, as Margolis and Liebowitz argued during the Microsoft case:

Although these simple numerical and algebraic examples appear both logically sound and structurally uncontroversial, these examples entail severe restrictions. The logic underlying path dependence is seductive but incomplete. […] Given that the theoretical claim that can be made for path dependency should be understood as only a demonstration of possibility, the case for path dependence becomes an empirical one.

It is not that networking effects are not real, but they are not as powerful as they were first made out to be. After all, networking effects could not save Friendster or Myspace; as we will see, they mean little for particular games.

At a particular scale, diminishing returns decay positive network effects to zero. The 2nd user who joined Facebook was far more beneficial to the 1st user than the 100th million users who joined. While not directly comparable, Google’s Chief Economist, Hal Varian, makes this point regarding the predictive accuracy of models with additional data.

We can create a similar arbitrary model for a particular game: a diminishing network effect value for the marginal user that inevitably results in an asymptotic total network effect value.

Not all games face the same curve. A game like Hearthstone, with 1v1 play, has far less to gain from an additional user than League of Legends, which has 5v5 and many ranked segments. More users reduce matchmaking times, potential latency, and thickening skill distribution (higher P, you will be matched against similar skills). The fewer segments (modes, ranks, etc.), the less powerful the networking effects are, and the quicker the marginal value curve depresses. Cross-play doubled down on this by removing platform segmentation. Nevertheless, even for League, the networking effect power is infinitesimally small at scale; the significant gains are eaten up with a relatively low user count.

Games do not have the “sticky” elements of network effects. Synchronous consumption is another example. In any real-time game, a network effect is delivered when you play with friends. At the same time, something like Instagram Stories is consumed whenever the user pleases. It is unclear that having a friend play the same game I do is beneficial unless we play simultaneously.

Schelling’s Nobel Prize-winning work on tipping is a more apt model for describing many games today. Consider a group of players all playing the same game. These players are partially driven to play this game to be “in the know” or a part of the pop-culture conversation. Each of these players has a given threshold for defecting to a new game based on the share of public discussion consumed by the game. Players defect if the game declines as a share of the conversation; they leave as the game goes from 100% of the public conversion to 99%. More will leave at 99% and more still at 98%, continuing a downward spiral into a new equilibrium.

Of course, the inverse is true as well. Some might join a game if its share of public conversation goes from 0% to 1%, and even more would participate if it went from 20% to 30%. A new game release can set off this “tipping” chain reaction in players. Look no further than the migration of Fortnite players into Warzone. The power of tipping is as strong as “public conversation” is in player motivation. Unfortunately for developers, this means instability in the long-run capitalization of viral game hits.

I have avoided addressing two-sided marketplaces as they will more neatly fit into the next part: platform power (or lack thereof).

Part II

Eric Suffert acutely describes the dangers of extending payback windows. At every t+1 the accuracy of LTV declines while the variance in cohort profitability increases. LTV, however, is not an exogenous variable and clever design can incentivize players into revealing their long-run time horizons within a game.

Consider the design of a many subscriptions: you can pay a lower annual fee or a higher month-to-month fee. If you’re uncertain about the subscription, then the month-to-month is more economical while if you have more certainty then the annual fee makes more sense. The choice is a huge predictor of retention: annual users are far more likely to retain then month-to-month users. The mere inclusion of this annual/month-to-month choice gives users the opportunity self-segment into more predictable cohorts. Why can’t we use the same mechanics in game design to create more predictable LTVs?

Consider two possible goods for purchase via gems in Clash of Clans: a builder or gold. The builder increases the long-run growth rate of gold while the gold itself is a temporary boost in short-run capital stock. In layman’s terms: spending 100 gems on a builder might net you 200 gold today and 1,000 gold by D30 while spending 100 gems directly on gold may only yield 700 Gold today and 0 gold by D30. The builder is an annuity that pays dividends every period, the longer a player’s time horizon the more valuable the annuity.

Players who expect to have a long time horizon in a given title have an enormous incentive to purchase “investment” goods or goods that pay dividends overtime (battle passes similar to some degree). Not doing so results in a increasing opportunity cost penalty every period due to lost compounding growth.

F2P has experimented with direct daily annuities of hard currency. They offer players a discount over the standard IAP packs but must pay upfront to receive a daily allowance. Instead of a 30-day pass, why not ramp to a quarterly or bi-annual pass? Doing so would make LTVs more predictable early in given player’s lifecycle.

The harm is described in Akerlof’s Nobel Prize winning paper on car lemons:

Akerlof’s original example of the purchase of a used car noted that the potential buyer of a used car cannot easily ascertain the true value of the vehicle. Therefore, they may be willing to pay no more than an average price, which they perceive as somewhere between a bargain price and a premium price. Adopting such a stance may at first appear to offer the buyer some degree of financial protection from the risk of buying a lemon. Akerlof pointed out, however, that this stance actually favors the seller, since receiving an average price for a lemon would still be more than the seller could get if the buyer had the knowledge that the car was a lemon. Ironically, the lemons problem creates a disadvantage for the seller of a premium vehicle, since the potential buyer’s asymmetric information, and the resulting fear of getting stuck with a lemon, means that they are not willing to offer a premium price for a vehicle of superior value.

If users start to have an expectation that a given game ad does not truthly describe the game in question then they’ll be less likely to click. This means higher UA costs even if your firm does not engage in these type of ads.

A similar problem is starting to creep up in PvP games. Developers have started to be confronted with the uncomfortable reality that PvP is a zero-sum game: for someone to win, someone else has to lose. And of course, when players cannot progress they churn. Supercell has heavily introduced bots in Clash Royale as a response.

Of course, players cannot identify if they are indeed playing a bot or a real player. This steals one of the compelling aspects, if not the compelling aspect, of PvP away from players: outsmarting another individual. But like the lemon problem above, if this trend continues then players will start to question if they’ve dominated a real opponent even if the particular game doesn’t use bots. Games that use unmarked bots start an industry expectation that diminishes the experience for all.

There’s a good moral rule here to help us and it’s from an 18th century philosopher called Immanuel Kant. Kant advocates for something called the Categorical Imperative. This claims that if we were to universalize a given action and it would result in a “contradiction” then that action is immoral. Consider lying: if everyone were to lie then the world would not function, therefore lying is immoral. Or consider being lazy: if everyone were lazy then nothing would get done. It’s a no-holds-barred approach to consider moral action, but thankfully we’ll use it a much more narrow scope.

If unmarked bots were to be universalized, PvP games would be irrevocably harmed. If misleading ads were to be universalized, then players would stop clicking on them all together. Both of these situations violate the Categorical Imperative and align with outcomes that benefit the entire industry. The Categorical Imperative makes for a simple and rule based approach to consider the “greyer” parts of developer action.

There’s a compelling aspect to achieving group oriented goals: being apart of something larger than yourself. Lots of F2P developers harp on the importance of social features. Yet the social experience in many games is abysmal. Lots of teammates or clanmates don’t seem interested in participating instead preferring to “free-ride”, putting forward little effort but getting the fruits of the team reward. Mancur Olson’s foundational work, The Logic of Collective Action, describes how this problem manifests in the public sphere (sometimes literally in the case of electric scooters). Game designers have a much easier time aligning individual and clan incentives than public officials yet they sometimes miss easy wins. How can we make the clan experience better then it might otherwise be?

In Clash Royale, clans advance a boat against rival clans. Advancing the boat depends on individual clanmates playing games everyday (and winning). The more clanmates play consistently, the more the boat advances and the better the rewards the clan will receive. But for many clanmates playing everyday requires a great of effort, why not let others earn the rewards for you?

The problem is severe in Battlefield where “PTFO” or “Play the Fucking Objective” is standard nomenclature. Players often won’t engage in activities that benefit the team (capturing flags), instead preferring to pursue their own objectives (generally: shoot players as fast as possible).

A given player faces two potential payoff schedules when considering to allocate effort to the clan. There’s the expected payoff with no effort (the probability that the clan/team will win if the given player did nothing) as well the probability that the clan will win if the player puts forth effort. We can model this as such:

[latex display='true']\mathit{expected\ payoff\ from\ effort_i} = P(\mathit{winning} \mid \mathit{effort_i}) \ * R[/latex]

where

[latex display='true']P(\mathit{winning} \mid \mathit{effort_i})[/latex]

is the probability of winning the clan event given give the effort of a given player or rather the additive probability of this given player participating.

While R is the reward from winning.

Of course if [latex]P(\mathit{winning} \mid \mathit{effort_i}) = P(\mathit{winning})[/latex] or the given player cannot sufficiently contribute to the probability of the clan winning then there’s zero incentive for them to put forth effort. Why bother?

This problem exacerbates as team size grows: the efficacy of a given player varies inversely with the number of teammates. This makes intuitive sense: in Battlefield, a player in 2 versus 2 match has a greater impact on the outcome then a player in a 32 versus 32 player match. The incentive to free-ride rises as the number of teammates or clanmates rises. Weakness hides in numbers.

We’ve also ignored the game-theory dynamics of this problem for simplicity, but it’s worth mentioning. If I know my other teammates are not going to put forth effort, why should I? This leads to Nash equilibriums where clans have almost no activity.

How can we overcome the free-rider problem and ensure that all teammates put forth effort? The highest cost-benefit feature is simply better monitoring tools. In many clan or team based games, clan leaders face asymmetric information: they simply can’t identify the players that do not put forth effort. A simple measure of activity (last login) or games played in the last week goes a long way to kicking out free-riders. We might also consider a joint-production function. In Battlefield or Clash Royale each player would receive a score based on their effort or contribution to team advancement, if the team wins they receive a multiplier on this score. Such a system would have two benefits: it would more closely align individual effort with individual outcome (reap what you sow), and it would increase the benefit for high performing clan members to engage in monitoring. For example, a high performing member might have $20 in contributions with a 2x multiplier or $40 for winning compared to a low performing member with $5 in contributions and therefore $10 for winning. In real terms, the high performing member has an even greater incentive to encourage low performers to put forth effort.

There’s a lot to be said for social shaming as well. While it hasn’t been effective for zero effort participants, there’s evidence it might help players on the margin. A push notification demonstrating that your clans needs you or perhaps better yet, a system where your clanmates can send you push notifications is a compelling way to push players into action.

Perhaps the greatest miss I see is not in clan monitoring (kicking out free-riders), but in self-selection to begin with. Clans are generally pareto efficient for players meaning that there’s zero cost and only benefit to joining one. Players then generally look for near max-size clans as they maximize the clan’s probability of winning a reward and thus the players. Reducing search costs by recommending (or restricting) clans based on device language, location, and some measure of progression maturity makes all players better off.

It’s hard for social monetization opportunities to take-off if team based activities suck. We still have a long way to go to fix top of the funnel problems. Afterall, teamwork makes the dreamwork.

The supply-side economics of subscriptions are fairly straightforward: get massive scale and distribute costs. Netflix has over 160M subscribers meaning a $100M production only costs each user $0.63. There’s zero marginal cost (unlike Spotify), so the more subscribers Netflix has the less a piece of content costs on a per user basis. At $11 per user per month, Netflix can spend over $1.7B on content each month and break even. In fact, as recently as 2019, Netflix does spend close to break even. In the long-run, Netflix has a incredible loop: content brings users, these users lower per user content costs, this accelerates more content spend, which brings more users. Rinse, wash, repeat until diminishing returns take hold.

But remember this wasn’t always the case. Content costs are fixed: the crew of Mindhunter doesn’t care how many subscribers Netflix has. They face costs unrelated to if 100M or 10M people watch the show. As early as 2010, Netflix only had 20M subscribers, suggesting a per user cost of $5 for a $100M production. There’s a gaping hole between $5 per user per $100M production and $0.96 per user per $100M production under their current 160M subscribers. This underscores the massive upfront investment needed for streaming to work on a supply-side basis.

It’s this exact reason Microsoft needs Playstation users for Game Pass to at least have a chance at buying down per user costs. It also explains their aggressive sub & console bundling as well as rock bottom introductory pricing.

Microsoft has disclosed Xbox Live has 90M MAU while Game Pass has 10M MAU. At $10 per user per month, Microsoft has can spend as much as $100M in content per month and break even. Development costs vary widely, but let’s start with assuming a conservative $70M per game. From a break even perspective, that’s around 17 games a year and consistent with Xbox Game Studios current production. That may sound like a lot, but the Xbox One has had over 2,500 games released in its lifecycle or about 350 games a year. Supporting this sort of volume is going to require a lot more users. A lot more.

Let’s have a little more fun with the cost side to get a better handle. Consider a simulation assuming anywhere from $30-$70M per game and anywhere from 100-350 games per year. We’ll randomly sample numbers in both ranges, multiplying them together to arrive at a per year content cost (ex: our simulation might pull $40M a game and 125 games for a yearly content cost of $5B. The density chart below represents 100,000 such pulls).

Simulating Content Costs of Game Pass

Based on this, Microsoft would most likely spend ~$7B per year to maintain similar volume. At $120 per user per year, this means 53M subscribers or about 5x the current amount to break even. At a 60% profit margin the number of subscribers jumps to 93M or more then the current Xbox Live MAU. Microsoft needs more users for this to work and they know it. Of course, perhaps this volume is insane and Microsoft wants much fewer but higher quality hits. In this case, however, the entire cost-savings model of subscriptions to the consumer declines. Our gaming subscription paradox re-emerges: subscriptions make increasing sense to the consumer wherein they would instead pay for many distinct pieces of content. Like music. Or T.V.

In many ways this explains the expansion of Game Pass to PC, Project xCloud, and the new Xbox console subscription bundle. Subscribers are capped by those with a console and Microsoft wants to start to unshackle that cap. Netflix recognized this long ago with Roku, Smart TVs, and apps on every platform or hardware they could find. The ability to sell Game Pass on iOS or Android explodes the addressable market, but is dependent on Azure wrangling the explosive unit costs of streaming. Hello Sony and the Playstation user base.

Microsoft clearly understands the supply-side economics of making streaming work, but over and over again misses the fundamental difference in how players consume a game as opposed to a T.V. show. I’m not quite ready to short Microsoft, but critical questions remain in executing on this strategy.

Amazon has announced Luna, yet another stab at game streaming and subscriptions. It seems like the failures of Apple Arcade, Stadia, Gamefly, and our oft-forgotten OnLive have not been effective deterrents. Not to be outdone, Rovio’s Hatch continues to drain money every quarter. Perhaps this is what Bezos warned shareholders about when discussing new “multibillion-dollar failures.”

And despite overwhelming failures, media pundits like Matthew Ball prop up skin-deep arguments in favor of game streaming and subscriptions. Instead of discussing why game streaming and subs might work, let’s talk about why they haven’t worked.

It’s essential to understand the brother/sister relationship between streaming and subs. Subscriptions unlock zero-marginal cost content consumption. Once you’ve paid Netflix or Spotify $10, there’s $0 additional explicit cost to consume another movie or song. However, there are transaction costs. In the “before times,” customers had to mail DVDs back to Netflix to receive the next DVD in their queue. This effectively limited the amount of content customers could consume in a given month. If mailing took 3 days in transit, on a 1 DVD-at-a-time plan, a customer could only consume 10 DVDs in a 30-day month. This assumes the customer turned around DVDs instantly.

Non-steaming subs like Game Pass and EA Play exist in a weird middle, solving some but not all of these issues. Games are distributed digitally, but not instantly. A game like Call of Duty: Modern Warfare can take 3-4 hours to download on a 135 Mbps connection. SSD’s aren’t cheering at the prospect of 200GB games either. But once streaming takes off for games, these problems are solved. No storage is needed!

The incredible rise of free-to-play and GaaS (Games as a Service) renders subs and streaming essentially valueless to the player. Players do not consume games like TV or music, which should have been made evident in the last decade. Players are playing fewer titles in a given year, but are playing fewer titles for longer periods of time. Games solve the content problem in a way that other mediums simply can’t.

The content consumed in games like Overwatch or Clash Royale is the pursuit of strategic equilibrium and/or mastery of mechanics. A new unit in Clash Royale, for instance, can change how players organize their decks, even if they don’t use the unit directly (they must counter it). This can provide hundreds of new hours of content to consume relative to the nearly one man-week of labor to produce a new unit. Therefore, the content output of a given member of the 16-person (!) Clash Royale team is astronomical. Compare this to the thousands of crew members and weeks necessary to produce even a single one-hour episode of Game of Thrones. Supply can’t keep pace with demand in the world of TV and movies. Netflix makes sense in this view because, after bingeing on 9 seasons of The Office, customers can immediately rip into 7 seasons of Star Trek: The Next Generation. It’s another reason why back catalogs are so much more important to Netflix than they are to something like Game Pass. If players are only investing in 3-4 new games a year, then the transaction cost reduction streaming is extremely minimal; the benefit it provides is in high unit consumption TV and music.

It’s a similar story for the failure of gaming subs. If players consume only 3-4 titles a year, subscriptions don’t make economic sense. Not to mention, these 3-4 new titles are increasingly becoming free. In the West alone, League of Legends, Fortnite, Apex, Warzone, and Rocket League dominate playtime. And let’s not forget the entirely F2P ecosystem of mobile. The march to F2P in the West will continue as long as MTX revenues grow and box revenues shrink. There isn’t a whole lot to save by signing up for a $100 a year sub and streaming service when Fortnite doesn’t cost a dime.

Game streaming and subs don’t solve billion-dollar problems for the player. In the absence of doing so, subs and streaming will continue to flounder.

Apple is increasingly under fire what’s claimed to be unfair practices in the App Store. The criticism takes three forms: (a) Apple’s 30% fee is much too high relative to cost (b) the rules are arbitrary and stifle competition and (c) the App Store as the exclusive avenue to install apps on iOS is unjust. That’s a lot to chew, so let’s focus on (a) for this post.

Eric Seufert’s criminally underrated podcast talks about this very topic but phrases the question as “Does Apple earn it’s 30%?” Various App Store benefits like payment handling and preventing fraud are discussed, but I wanted to scream by far and away the biggest way Apple “earns” it’s 30%: acquiring a mega fuck-ton users to iOS. Nearly the entire value of a platform to a developer is how many people it can reach. No one is rushing to get on Epic Game Store (EGS) or develop for Stadia because there are so few users. 12% or even a 0% fee is irrelevant: [88% * 0 users] = $0 versus [70% * more then zero users] = more then zero. This is an extreme example, but users are by far and away the most important ingredient of any platform. After all, developers can list in multiple stores with maintaining the listing or code for the particular platform as the only cost to do so. The fact that so few are willing to take on these small costs tells us a great deal about the user bases of Stadia and EGS.

But platforms fees also increase the LTV of the devices who run on the platform, if and only if, the firm internalizes those platform fees. For instance, the added platform LTV of an Android phone in China is less than in the United States because Chinese users install non-Google owned stores. In this case, platform fee revenue doesn’t accrue to Google.

If LTV of the iPhone X is say $1000, Apple should spend up to $1000 in UA to acquire a user. However, with platform fees the LTV grows. Sensor Tower estimates that iPhone users spend over $79 per year on apps. On a lifetime basis, that’s probably north of $250 (3+ years of ownership). That ups the iPhone’s LTV to $1250 and unlocks more UA budget for Apple to acquire more users. These marginal users benefit all developers. This is how Apple (or almost any platform holder) gives developers a rebate on the 30%. I’m not sure what the true fee is but it must be lower than 30%.

“We want to be more data driven” or “We want to create a stronger data culture” are common organizational refrains. Supposedly, having more data or data playing a larger role in the decision making process is profitable. It’s weird because I haven’t seen any research to suggest this is the case. In firms like Facebook, it’s obvious as more data improves ad personalization and thus revenue. But this is data as a engineering project rather then a tool in the decision making process. Firms want to make better decisions with data. This is a misidentification of the value chain. Data isn’t that helpful if it’s not packaged with empiricism, an epistemological way of acquiring knowledge.

To even get off the ground analyzing data, we need theory of measurement. What should we track, given limited engineering resources and raising storage costs? Claiming we should track, say payments and logins, at the exclusion of audio volume, implies a cost-benefit value ranking. Why are payment and login more value to track? The theory is that understanding payments and logins will unlock more insight then volume as volume plays a less significant role in the app. Is this true? Hopefully, the institution has the intuition or previously collected to knowledge to make an educated guess. Firms have discovered that refining this knowledge can make their bets more likely to succeed. As it so happens the West has created the best knowledge refinement process in the history of humankind: the scientific method.

Data or more broadly, empiricism, is a key part of the scientific method as it expands the sample size of a test beyond antidotal evidence. Doing this at scale, as well as the methodology of running true experiments, A/B tests, means that knowledge is more valid (less likely the result of antidotal evidence) and stable. Firms can now learn.

Arguing to be data-driven or informed misplaces the value in the supply chain. We need to more explicit in this endeavor – it’s not about data, it’s about science.

Performance based marketing makes intuitive sense; of course you want to optimize ROI on spend that compose 30% or more of a your firm’s expenditures. But here’s the kicker: if it makes sense for firms why not individuals? Shouldn’t we be tracking the impact of our output?

Are co-workers actually reading your analysis, concept art or brand pitches? Imagine if you knew how long internal onlookers spent reading or viewing: how might that change future output? For instance, we might be able to find the optimal length of a memo or detail of a UX mock-up. With more widespread read receipts (on email, calendar invites, PowerPoints) I’d think we’d learn discovery of information within firms is rather low because search costs are so high. I’m routinely shocked by low view counts on important internal Google docs while on the other hand observing the willingness of participants to take a firm stance on the conclusions. Many firms haven’t invested in strong intrawebs to create easy ways to access content or for creators to push them to relevant parties.

This means information travels via internal networks via word of mouth. With a given game publisher distributed over many countries and time zones, this simply doesn’t scale well. Easy consumer discovery, regular “pushing” of content, and tools assist rather than inhibit creation are paramount to spreading the gains from Haykiean localised knowledge.

Bloomberg is reporting that Epic is seeking to raise $500M to $1B in new funding. Tim Sweeney, Epic CEO and provocateur extraordinaire, owns 60% while Tencent owns the other 40%.

Why would Epic need more funding as it continues to rake in Fortnite money? Tim has bigger ambitions, let’s figure out what they are. As we’ve seen with Roblox, there’s incredible (and growing) money to made in user generated game content. Furthermore, consider the games spawned from mods: Counter Strike, the MOBA genre, Team Fortress, Auto Battlers, H1Z1 (which becomes PUBG which becomes Fortnite)… Imagine if you were able to take even 10% of the lifetime revenue of those titles. Despite Tim’s criticism of platform holders, Epic takes a cut of many games that run on Unreal as part of the licensing agreement. This is reportedly how Fortnite pivoted to BR: Epic got an enormous cheque when PUBG was blowing up.

The pieces have been coalescing:

• Content creation tool in Unreal – which now scales to mobile
• Direct to consumer distribution – Epic game store – which now scales to mobile
• User Acquisition engine and revenue bedrock – Fortnite

But I think this play would blur more of the line between gaming and social networking. The biggest evidence comes from Epic’s acquisition of Houseparty, an app that essentially lets users join a Zoom call and play games together. How else do you explain that acquisition? It ties into their already created, but lesser known, Fortnite Party Hub app and recently announced Party Royale.​

The recent foray into non-gaming content via Fortnite concerts makes more sense in this light as well (or they’re trying to keep the Fortnite in the mainstream).

This helps explain hiring as well; in less then a year Epic has setup offices in practically every major gaming hub: Seattle, San Francisco, Stockholm, Montreal, Helsinki and Berlin. Recent jobs ads focus on the social.

Competition is spinning up in the meantime. Manticore games has already launched an alpha, and it’s hard to imagine Roblox being content with only owning the 10-14 yr old demo. It’s important to note this is nothing like a ‘metaverse’ as each game on the platform would retain a distinct identity, whereas a metaverse is closer to the failed PS Home or Second Life. Trying to squeeze hundreds of experiences onto one game presents a variety of complications and very little in the way of benefits.

Maybe when it’s all said and done, Fortnite let’s Tim become the social platform holder Zuckerberg always wanted to be at Facebook.

Economists like Tyler Cowen or Brad DeLong are too self-respecting to study reality shows. Fret not, this economist has no such self-respect.

Previously, we examined the economics of the reality show genre but just as interesting are the economics of the a particular reality show’s design.

To Hot Too Handle introduces of the most interesting examinations of communal property dynamics: a group prize is reduced when individuals act in their short-term private interest.

At a more practical level, the show gathers ten attractive 20 somethings into a villa in Mexico for three weeks. Cameras are littered around the villa with the exception of bathrooms (to be replaced with mics). The contestants are only informed of the rules once the cameras start rolling; if they masturbate, kiss, or engage in any sort of sexual activity the prize pool of $100,000 is reduced. It’s unclear to contestants how “expensive” each activity is or how the prize pool will be divided or won. Shockingly, interviews with the show’s producers reveal they didn’t have the rules or the costs figured out until they happened. While there’s no traditional contestant elimination process, producers will ask contestants to leave if they’re not invested in the “process”. Supposedly, the show wants to teach these singles how to form emotional rather then physical connections.

The spectacle for viewers is how hard it is for these contestants to keep in their pants – of the original $100,000, over $40,000 is lost. Seems like a lot, right? How could they give up so much money?

Well… It’s really not that much. On the face of it $100,000/10 = $10,000 per contestant. The tax situation matters greatly – U.S. contestants or those with residency in the U.S. will probably pay about 50% of that $10,000 in taxes. Interestingly, if the show took place in the U.S. rather than Mexico, all contestants would be subject to U.S. taxes. It appears to the case that the Brits and Canadians don’t face game show taxes.

On a expected payout basis, the costs are far less then they might appear:

• $3,000 for a kiss is only $300 on a per contestant basis. Only $150 after taxes.
• $6,000 for oral sex = $600 gross, $300 after taxes.
• $20,000 for sex = $2,000 gross, $1,000 after taxes.

The show filmed for 3 weeks, at a max payout of $10,000 this is a yearly salary of $173k. Not bad, but many of the contestants already out gross that. Francesca Fargo is estimated to have a net worth of over $500k alone. Almost all of the contestants make money off their likeness or brand. Like Francesca, they model, sell clothing or act. Thus, building an Instagram following is directly connected to their revenue stream. Breaking the rules can help the contestants build that brand – losing out on $300 now could be much more in brand awareness later. Those without brands seemed to leave early or not attempt anything “interesting” – see Madison – a late arriver who never coupled up.

But the rules weren’t clear on splitting the prize and contestants could have been under the impression only 1 or 2 would win. Under an expected value model the payout is the same: $10,000 ($100,000 *10% chance of winning). However, if you feel as though you’re a weak contestant you might estimate yourself at less than 10% probability to win. I think this was the case for sorority girl Hailey who broke the rules a mere two episodes in and had no interest in continuing.

I think there’s room for improvement in the show’s design. It was rather strange to reveal to contestants who the rule-breakers were so early in the show. This introduced social shaming as retaliation for rule breakers, speculation and investigation makes for far more drama. If the show was about temptation, why not focus more on the money or relationships? Maybe contestants can choose to eliminate their show’s squeeze – money AND sex as tests of genuine connection. Discounting seems like a great lever for drama injection – this week sex is 50% off! Adding new contestants didn’t seem to work, everyone had coupled up by the time they got there. Subtraction or an elimination is lot more fun.

Well, here’s to a solid season two. Hopefully, the show remains tongue in cheek. But not literally – that would be a rule violation.

In case you haven’t been paying attention, Netflix just re-discovered the reality/reality-doc genre. In 2 weeks:

• Tiger King
• To Hot Too Handle (international cast)
• The Circle (4 international versions already released)
• Love is Blind

I chuckled a bit when when Netflix remarked that they saw Fortnite as their biggest threat, but after 6 weeks in quarantine, they may have been on to something. Reality television is an awkward, but fun shot back at digital Travis Scott concerts.

The economics of reality shows have always favored cost. Casts cost $0, 1 or 2 filming locations, short shoot time (sub 6 weeks), and strong vitality potential that front loads views (no one is watching the back catalogue of Survivor). But why now, 7 years since House of Cards launched?

I think this is a simple growth parable: they’ve hit diminishing returns on drama, and this is the next highest expected marginal value item on the menu. In other words, they’ve picked the low hanging, high value fruit already. Widespread international distribution expands the watercolor effect and achieves economy of scale. Shows with live elements like Idol and Got Talent are forced in country specific renditions which limit audience reach. Subtitle support for international versions of shows like The Circle helps it reach English audiences and vice-versa. Living in Europe, you’d be surprised of the appetite for English content across non-native speaking countries.

They’re innovating reality game design as well (or at least the 3rd party studios are). Technology play a huge role in all shows, almost always as a form of communication (or lack thereof). To Hot Too Handle introduces of the most interesting examinations of communal property dynamics: a group prize is reduced when individuals act in their short-term private interest. The Circle asks how people make decisions on limited information – contestants can only communicate over text.

If the last last years have been golden age of TV dramas, maybe the next 10 will be the golden age of reality TV.

Our Surfer is ‘risk loving’

The increase in price for a  given ‘bottle draw’ would equal the expected payoff of a lower priced ‘bottle draw’ that randomly refuses to give a bottle. This is an interesting solution to player frustrations in gacha (“I didn’t get anything of value when I opened a pack!”).

Anyone care to reckon which model one would perform better: Higher draw price but gives money back or lower draw price but sometimes doesn’t give anything?

Markowitz is not 5′ 9′ Milton Friedman and he resists your attempts to call him that.

Previously, I wrote about ads as a way to monetize non-payers, but there’s more to the ad exchange and what I’ll coin as ‘portfolio pumping’. It’s like portfolio theory, but not really.

These terms reference two growing phenomenon in F2P games. King is at the forefront of portfolio pumping, in which a given firm pushes a player from game to game within the firm’s portfolio.

Match-3  games in the portfolio and get $$$, ???.  Image credit: Eric Seufert

Unlike portfolio pumping, ad exchanges push players to another firm’s games. Companies like Scopely are more fond of ad exchanges.

Walking Dead ad in Yahtzee, interesting given Scopely has competing Walking Dead game.

Frequently, the ads being served are for competitor games. Why would a company show ads for its competitors? In addition, why would firms want players to move from one game in their portfolio to another? I argue the underlying explanation is Pareto Efficiency which is just a fancy term for trade.

Ads for competitor games only make sense to the ad-server if

and to the advertiser if

It tends to be the case that a given company will engage in both ad buying and selling. The outcome of these ad exchanges are migrations of players to the games in which they have the highest LTV; the initial allocation doesn’t matter. This process takes place in high-speed auctions where firms are constantly in the search for the maximizing the equations outlined above. The decision rule for portfolio pumping is similar, but we add some special conditions, mainly the probability of simultaneous play.

Where,
is the probability of playing both games simultaneously. We add up both of the LTVs in this case.
is the remaining LTV in the old game for the ith player, while nLTV is the LTV for the new game for the ith player.

This must be bigger than for profitability.

Of course, there are ways to play with this. Wooga tried altering portfolio game prompts during a player’s lifespan but found no effect.¹ King continues to portfolio pump but dropped ads in Candy Crush Saga.

It’s a goddamn gorgeous process that should litter econ textbooks like lighthouses and lemons.

Will Luton argues on the dangers and solutions to F2P inflation over at gameindustry.biz.

While there are some missteps in the opening of the article, Will makes a powerful and elegant point:

…a sale can only be considered profitable if the net revenue from the start of the sale until resource equilibrium, and so demand, is restored is more than if the sale hadn’t been run. For well run sales in games with well balanced economies this should always be true.

Sales flood the economy with resources via shifts along the demand curve. Holding all else equal, this is modeled as a move from P1 to P2.

The tricky part, not found in the textbook model, is time. Unlike say, refrigerators, a durable good, virtual currency is a consumable good. This means we expect repeat purchases, similar to say, gasoline. Sales in this sense pull revenue forward by changing purchase ‘schedules’ more so then a durable good. The sales are only profitable if the sale sinks resources players would have never sunk otherwise (net positive sink). In games, this is achieved this achieved via live ops. This model explains how Supercell runs their games; it’s no coincidence that Clash Royal is the first Supercell game to have sales and real live ops while their other titles have little of either. Introducing one without the other keeps net sink flat in the long run by shifting intertemporal time preferences rather than increasing the size of the ‘sink pie’ so to speak.

Progression is another confounding variable. Holding all else constant, a given item is worth less for each additional level a user is at. This is simply an artifact of rising difficulty (in the form of stronger enemies, more experience to level up etc). As a result, sales make late game players in different while making early game players better off.

Reproduced from the original article. Left: player progression on a game with without sales. Right: a game with sales.

The insight Will offers is that sometimes this is an advantage by changing the progression path of newer players to a higher equilibrium then current late game players previously had.  This allows new players to ‘catch-up’. This sounds a lot like the Solow model. Yes, that Solow model ². I don’t think Will models this correctly, however, as each player is not on a discrete curve as his graph on the left depicts. Even without inflation, the graph on the right is an accurate picture of a given game economy.

Consider two possible goods that could be put on sale (and thus inflated) from Clash of Clans: a builder or gold. The builder is a dramatically better purchase because it allows for more output per unit of gold or elixir (increase in technology). This shifts the growth rate of a given player up. On the other hand, the gold is a small one-time increase in capital stock that won’t scale with the game. For designers, this offers the chance to use sales as strategic instruments to alter the metagame. By offering Clash of Clan players discounts on a builder, players converge and then exceed the GDP of elder players. A sale of gold, however, merely ‘jumps’ the GDP of players without changing the long-run growth rate. This means designers can either jump the point along which new players are on the progression curve or they alter the new player curve entirely.

Sales jump players along the curve (A -> B), while a sale with ‘technology’ creates an entirely new curve. Also, I have an S pen. Watch out art team.

Back again

Known as ‘Destroyer of Toy Islands’ on Server 54G, Robert Lucas knows no mercy.

Unfortunately, this can deter some investment by changing inflation expectations. If players know a given dollar will have increased purchasing power later on, why make the investment now? Indeed, a 30+ paper written by Game of War players and subsequent boycotts attest to the negative side effects of perpetually trying to catch players up.

Careful consideration and analysis can make sales a valuable gameplay tool as much as they are a business one.

A paradigm forever changed, one man carries a dying tradition.

Steve Levitt, the last price theory samurai, and John List, future nobel prize winner, have published a paper on free to play economics.

In a textbook neoclassical experiment, Levitt alters the quantity of Candy Crush hard currency at a given price point. While economists generally think of price variation as the way of deriving demand curves, quantity variations are just as legitimate a tool.

Despite a sample size of over 15 million and a wide range of quantity convexity (80% variation across variants), all quantity discounting schemes produced similar revenue. Levitt concludes by commenting,

“…varying quantity discounts across an extremely wide range had almost no profit impact in the short term.”

The interesting and little explored result indicates that,

…almost all of the impact of the price changes was among those already making a purchase; radical price reductions induced almost no new customers to buy…

This suggests free to play games are made up of two groups of users: purchasers and non-purchasers. This means the decision of becoming a customer is exogenous, there is no ability to convert non-customers to customers  i.e. this is decided outside of the game.  Put another way, non-customers are perfectly price inelastic and customers are perfectly price elastic. Indeed, industry research collaborate this.⁴

Interesting, but is it actionable?

Were this to hold, it suggests a number of results. The first is that product manager’s ability to monetize non-customers (99%~ of users) will not come from IAP, but rather other forms. This may help explain why F2P ad revenue and incentivized video continues to show YoY growth.^{5 6}
Furthermore, product managers should consider experiments exploring the maxima point of ad frequency. Given that there’s a trade-off between retention and ad-frequency there exists an optimal ad frequency point.

With little chance of non-customers converting to customers, product managers should worry less about increased ad frequency turning off potential customers.

The final result suggests the ROI of trying to raise the LTV of customers exceeds that of trying to raise the new customer creation rate. Product managers should develop roadmaps in accordance.

Seufert argues for a number of institutional policies to surround analysts with within an organization. Frequently, analytics and data are as much about the appearance of sophistication as they are actual value adds. This need not be the case. The confusion arises over where the value of data lies. Perhaps ironically, data’s value doesn’t lie in the data, but rather in the data analyst.

In most organizations, analytics reports to product teams, a mistake, Eric argues. Often product managers face the principal – agent problem: their incentives and the companies do not align. Product managers want to successfully manage products and will present the narrative they are doing so. This is inefficient for companies who often wish to assess the true performance and trajectory of a portfolio. When an analyst’s career path depend on a product manager their narratives will often match. With organizational independence from product teams, analyst’s incentives align closer to the companies, providing more objective analysis.

Not just an accountability watchdog, real analyst value revolves around the ability to drive product roadmaps. At it’s highest order, analytics is a forward looking discipline, not a backward looking one. By experimenting and studying human behavior, analysts find levers that pull certain responses.  This creates opportunities to exploit these levers. Do currency pinches increase monetization? Are new gotcha characters or new levels driving revenue? Should we invest more in reducing load times or UI changes? Using theory driven empirical investigation analysts can move companies towards better outcomes than competitors. If organizations don’t allow analysts to pursue these questions, they’ll become cheerleaders for product teams. On the other hand, if first order information (RR, ARPU) is not accessible or automated, analysts will forever be running the hamster wheel of reporting. This is one of the more overlooked points Eric argues for.

I think this suggests a dual mandate of analysts: (1) accountability of features and (2) what features are worth developing. This creates a natural tension of not only playing the role of watchdog to product managers but partners as well. It is the duty of good analysts to navigate this relationship successfully.

“$10.00 LTV am I right?”

You’ve soft launched your game, done a UA push, and a string of hope appears. Against all odds, a dominant cohorted ARPU curve emerges! Is this this an anomaly or have you caught a whale?

The first way to examine this is to perform cointegration tests between the cohorted ARPU curves, testing for stastistical significance. It may be true the difference in the curves are real, but that doesn’t answer if you’ve caught a whale.

In 1905, Michael Lorenz developed a method for measuring relative inequality between nations known as the Lorenz curve.

Just keep saying what % of the population owns what % of the wealth and it’ll make sense.

The F2P application is to define wealth as revenue (either on a daily or game level) and players as the population in the context of free to play games. By measuring how bent inwards a cohorted Lorenz curve is relative to other cohorted Lorenz curves we can measure the ‘whali-ness’™ of different cohorts. Even better is how this reduces to a single metric – the gini coefficient. A gini coefficient of zero indicates a perfectly equal distribution of income, 10% of the population owns 10% of the wealth, 20% of the population owns 20% of the wealth and so on and so forth. A gini coefficient of 1 is the exact opposite – a single person owns 100% of the wealth.

This translates to what % of players are responsible what % of the revenue. Measuring gini coefficients across games rather than cohorts gives more insight into how a particular game monetizes – whether it’d be whale, dolphin, or minow driven.

Actionable insights might include how effective introducing ads could be. A high gini coefficient (very few players are responsible for revenue) might mean there’s a more fertile base to monetize on.

The main insight, however, is further understanding. It’s clear that success can come about in drastically different ways in free to play games, the gini coefficient is simple way to measure that.

Here we see the content problem in its natural habitat

F2P is as much of a design choice as it is a business choice. Given this, F2P has its own set of design challenges among  which is the content problem.

Developers will only continue making additional content until the benefits are greater then the costs. This is specified when

expected marginal revenue from content > development costt + opportunity costt

where

development costt is the cumulative cost by time of release (t)

but if

User Acquisition Rate (UAR) < Churn Rate (CR)

there’s a shrinking pool of buyers which only widens at t+1. This is the essence of the content problem: how do we create content fast enough to curtail churn and while minimizing development costs?

The genius of PvP (Player v Player) environments is how they necessitate the emergence of a meta-game. In mathematics, Player vs Environment (PvE) resembles the field of optimization where strategies are static – one and done. PvP environments, however, resemble game theory models where it has been shown strategies evolve in an evolutionary process. This means equilibrium in PvP environments is constantly being reshuffled with each balance change; the search for dominant strategies in an ever shifting equilibrium is the game itself.

It’s been 4 years since the launch of Clash of Clans and there continue to be oodles of strategy videos. Supercell is constantly debuffing and buffing different units which makes some strategies more successful than others and by trial and error players expose this.

Is it a paradox to watch mobile strategy videos on a desktop?

The push for PvP environments has seen the emergence of ‘Systems Design’ and the demise of a Level Designers. With few exceptions, linear and deliberate gameplay has gone the way of Spaghetti Westerns.

On the other hand, a different type of PvE has found ways to combat the content problem. For example, Trials Frontier adopted meaningful level mastery with a touch of PvP. This is achieved via quests that revisit locations, stars, leaderboards, mission rewards, and gameplay that rewards depth (back/front flips can improve my times!). That said, PvE shares a smaller piece of the pie than it once did. This trend will only continue as F2P marches into the console and PC arena.

Predicting the average cumulative spending behavior or Lifetime Value (LTV) for players is incredibly valuable. Being able to do so helps figure out what to spend on User Acquisition (UA). If a cohort of players has an LTV of $1.90 and took $1 to acquire then we’ve made money! This helps evaluate how effective particular channels of advertising are as we’d expect different cohorts of players to have different values. Someone acquired via Facebook may be worth more then some acquired via Adcolony.

But wait there’s more!

My argument in this post is that LTV has great deal of value outside of marketing. In fact, LTV might have parts more valuable then the whole. How to predict LTV can adopt numerous approaches and each approach has associated benefits. Remember, there doesn’t have to be just one LTV model!

Consider four requirements we’d want out of an LTV model:

1. Accuracy

LTV predicted should be the LTV realized. Figuring out upward and downward bias in your coefficients is important here. This gives insight into the maximum or the minimum  to spend on UA depending on the direction you suspect your coefficients are biased towards.⁷

2. Portability

Creating models is labor intensive and even more so when doing so for multiple games. There are particular LTV models that sweep this aside called Pareto/Negative Binomial Distribution Models (NBD). Since they’re based only on the # of transactions as well as transaction recency they don’t require game specific information. This means you can apply them anywhere!

3. Interpretability

This one’s big and perhaps the most overlooked. Consider the Linear * Survival Analysis model approach to LTV. The first part is to predict when a particular player will churn. By including variables like rank, frustration rate (attempts on particular level), or social engagement we gain insight in what’s retaining players. This type of information is incredibly valuable.

4. Scalability

If it’s F2P then there are going to hundreds of thousands to millions of players (you hope). I’ve seen some LTV approaches that would take eons of time to apply to a player pool of this size, our LTV should scale easily.

So how do the different approaches stack against one another?

Parteo/NBD is great, but it’s hard to incorporate a spend feature (it just predicts # of transactions).¹³ A small standard deviation in transaction value gives this model a great deal of value and something to benchmark against. This model also makes sense if data science labor is few and far in between.

ARPDAU * Retention is probably the approach you’re using; it’s a great starter LTV. If marketing/player behavior becomes more important, the gains to scale from a approach beyond this start to make more sense.

Wooga + Excel just doesn’t scale which kills its viability, but it’s conceptually useful to understand.

Linear * Survival Analysis  gives a great deal of interpretability that also sub-predicts customer churn time. This means testing whether the purchase of a particular item or mode increases churn time is done within the model. The interpretability of linear models also means it’s easy to see different LTV values for variables like country or device.

There are many, many different approaches beyond what’s been laid out here. Don’t settle on using just one model, each has costs and benefits that shouldn’t be ignored.

Friedman and Mario are about the same height.

In hindsight, one of Friedman’s great predictions is the Eurozone crisis. Despite being a massive champion for flexible exchange rates, Friedman never advocated for a common European currency.

Europe exemplifies a situation unfavourable to a common currency. It is composed of separate nations, speaking different languages, with different customs, and having citizens feeling far greater loyalty and attachment to their own country than to a common market or to the idea of Europe.

— Milton Friedman, The Times, November 19, 1997

The Greek financial crisis exemplifies many of the problems Friedman points out.

In an economic recession, central banks devalue the domestic currency to return the country to full employment. When economies are similar, recessions move together; if there’s a crisis in Texas, it’s probable there’s one in Washington. This makes central bank policy more effective because capital won’t escape from ‘recessed’ areas to ones with higher returns – there are none. It can be much harder to accomplish this in Europe, where every country’s economy is dramatically different, and institutional policy fluctuates widely.

What the hell does this have to do with game design?

Why do multiple currencies exist in games? Why not just have one type of currency rather than four or five?

The answer is segmentation and capital flight.

Once again, Supercell has provided us with a beautiful example, Clash of Clans (CoC). In CoC, some items cost gold, others and others elixir. After a quick scan, you’ll notice only the defensive items (cannons, archer towers, walls) cost gold, and only the offensive items (troops, barracks, spells). Why might this be the case? Segmenting these items gives designers greater control over the economy and minimizes the potential for ‘contagion’ effects. Consider a world in which Clash of Clans only contained gold. Players might prefer attacking rather than defending, encapsulating the idea of capital going to its highest return. If this were the case,e the game could become unbalanced as all players attack and none spend gold to upgrade their base defense. By segmenting base defense into the elixir, you remove any opportunity cost from expenditures for base defense. This is similar to giving your relatives a gift card; instead of spending it on whatever they fancy, they must now spend it on whatever is from the gift card store. A domestic currency is much like a gift card to that country’s ‘store’ just as an elixir is a gift card to only CoC’s offense ‘store.’

If Supercell finds players are not creating challenging defenses, increasing the rate of gold production is straightforward without worrying that money will be spent on offense. They can also do this by lowering the cost of items priced in gold. Supercell has toyed more with this strategy in their other title, Boom Beach.

The rules for when segmentation is worthwhile emerge from reading Mundell’s famous paper ¹⁴ backward.

Segmentation in games, like in real-world economies, gives game designers and central bankers more control.

One of the favorite topics of experimental economists concerns risk aversion and auction theory, risk aversion due to its ability to challenge the neoclassical paradigm (i.e., mainstream economics), and auction theory because it uses fancy mathematics. The first groundbreaking economic experiments employed auctions in lab settings to see if participants diverged from rational behavior. A series of experiments run by Cox, Smith, and Robinson¹⁵ appeared to show participants were not doing what we’d expect them to do if they were rational agents (i.e., getting the lowest price). The suggestion being participants were acting as risk-averse agents rather than risk-neutral agents. However, the critical insight came from a challenge in how these experiments were run from a 1992 AER article called Theory and Misbehavior in First Price Auctions.¹⁶

The author, Glenn Harrison, argued that the costs of engaging in non-optimal behavior were minimal. In other words, being dumb didn’t cost participants much, and being smart didn’t earn participants a great deal either. Glenn argued this casts doubt on the suggestion that participants were engaging in non-optimal behavior. Still, instead, participants weighed the expected mental effort of being innovative and concluded it wasn’t worth the foregone increase in income.

Each deviation from zero (the optimal bid) costs the participant little.

Glenn argued that what researchers need to do was bend the payoff curve, i.e., increase the reward for being smart. This way, researchers can see if their testing behavior is accurate.

What does this mean for A/B testing in my game?

Developers often turn to A/B testing to test even the most minute items; frustration emerges when results are inconclusive. For example, Supercell might test whether players prefer reward schemes X or Y in Boom Beach by way of sessions played. An A/B test that presents each scheme after a battle could be inconclusive. This is because each reward has a small outcome on player progression. That is an insight, but if we’re interested in whether A or B is better, it will make sense to ‘bend the payoff curve.’ That means we’d offer A + 5 or B + 5 to exaggerate the effects of the different reward schemes.

Think of it as amplifying two lights on each side of a room to see where flies gravitate toward. If the lights were dim, the effect on the flies would be smaller than otherwise.

See? Just like an A/B test.

While not always appropriate, bending the payoff curve is another tool developer scientists should consider when designing experiments.