The Curious Fall of Marvel Snap

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.

Why No One Cares About App Store & the EU Anymore

Somehow the United States beat the EU at a solution for App Store stangleholds: unlock developer free speech to ease off-platform payment flows and messaging (apps can communicate with users from an email address collected inside the app, and directly link to web payments pre-filled out). The California court ruling has Apple continue to oversee App Store distribution (addressing legitimate security grievances), while the off-platform payments function serves as a new “watchdog” group for App Store taxes.

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.

The Best Web3 Arguments (w/Yat Siu, Cofounder of Animoca Brands)

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

Three Things from 2025 Play Ventures Summit

---

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.

Economy Diagramming Using Figma & AI

Every game/economy diagram worth its salt contains a +Power ↔ +Action slide. It says something, but usually it’s not enough to understand what the hell is going on. Like traditional economic modeling, diagramming clarifies uncertainty by revealing a game’s internal structure or “wiring.” It forces a simplification of assumptions, and, like the effect of understanding the world from a Mercator projection map, the diagram itself shapes one’s understanding of reality. 

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.

7 Things Squad Busters Needs to Experiment with NOW

The game needs to go crazier. If game development aligns with an explore-exploit model, Squad Busters must continue exploring. That means covering as much exploration space as possible. Eric Seufert makes similar claims in UA creative testing, and those lessons apply similarly to games. The changes from 2.0 were big, but they need to be bigger.

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.