Metagame

A meaningful patch hands the community a fresh argument, and arguments are content. For the week after, every match is a little research expedition: is this still good, and what beats the new thing? That exploration is play, and it’s some of the cheapest play you’ll ever manufacture.

This is the kernel of truth in the “perfect imbalance” argument that’s floated around since the Extra Credits video popularized it in 2012: a slightly imbalanced game gives the community something to discover and argue about, and rotating the imbalance keeps that discovery renewable. I mostly buy it. But I’ve seen the failure mode up close, so I want to push on it. When balance churn becomes the only content, you’ve swapped a meta for a treadmill, and players can tell the difference. To the player who just lost six months of mastery on a champion or legend, that shake-up lands as a tax, whatever the patch notes call it. The patch cadence should disrupt the answer key, not the player’s investment in learning.

David Sirlin has the sharpest counter to balance orthodoxy I’ve read, and it’s worth sitting with:

He’s right, and most balance discourse blurs exactly this distinction. Fairness is the floor; interesting is the actual target. Data will tell you when a matchup is unfair (win rates, ban rates, presence), but it will never tell you when a meta is boring. Riot eventually formalized the fairness half into the Champion Balance Framework in 2019, with explicit win-rate and ban-rate thresholds across four audiences. That’s an honest public admission that “balanced” is audience-relative, and it made meta-stewardship accountable policy. It did not make the meta interesting, though. That part is still authorship, and I expect it always will be.

The familiarity is what makes the twist legible: a player can only feel the new pacing or the reworked weapon because they have a thousand games of “normal” to feel it against. Novelty with no baseline just reads as noise, and a baseline with no novelty is a solved game waiting for its community to drift away.

This is the sharper version of the treadmill caveat in META 01. Forget how much to change per patch; the real question is what the change is denominated in. Change the answer key (which comp, which build, which pick is currently correct) and players get to rediscover the game using skills they already own. That kind of change is renewable: reading a new meta is itself a skill, and the players who do it fastest experience that speed as expertise, so you can rotate the answer key forever and mastery keeps compounding. But change the alphabet, the core mechanics and foundations their mastery is written in, and you’ve devalued the accounts of exactly the players most invested in you. That spends trust you can’t easily mint back.

So the craft is really in the dosing. A twist should be big enough that the community has to argue about it, but small enough that last month’s mastery still pays rent. When a patch lands and the lab work starts (scrims, spreadsheets, creators racing to publish the new answer first), that’s the meta working as designed. If players’ first question is instead whether their main still exists, you’ve twisted the wrong layer.

One more piece of the discovery rhythm lives outside the patch cycle entirely: pick/ban is the community’s between-patch governance valve. Chapter 04 treats it as an instrument.

My position here is pretty simple: open your data and accept the consequences. The third-party ecosystem that grew on top of match history (op.gg, u.gg, the trackers) does more for player investment than any first-party feature could. It turned match history into the community’s shared memory, made mastery legible, and gave millions of players a public body of work, a resume made of games. No studio could have built all of that in-house, and honestly, no studio should try.

The consequences are real, and you don’t get to pretend otherwise. Draft-shaming and lobby dodging based on someone’s profile only exist because the data is open. Cookie-cutter builds spread faster, and the meta gets solved in days instead of months. I’ve heard all of these used as arguments for closing the data down, and I think they get the accounting wrong. The pathologies are the tax; player investment at ecosystem scale is the revenue. Design around the tax instead of forfeiting the revenue to avoid it. As far as I’m concerned, the players delivered the verdict on open data when they built their identities on top of it.

That match record is also the bridge to the second half of this chapter. The same artifact that feeds the strategic meta (tier lists, scouting, what’s good) also feeds the account: your history, your progress, your stuff. Which brings us to the metagame you build on purpose.

If the answer is no, you didn’t build a mastery system, you built a time clock, and you’re paying out wages in the shape of medals.

This sounds obvious, but almost nothing ships this way, because attendance is so much easier to measure and reward than skill. Daily quests, login streaks, and most battle pass tiers all pay attendance. Riot’s 2024 Honor postmortem said it outright about a system I worked on: it had drifted into rewarding attendance (chapter 06 owns that story). The drift pattern is always the same. You ship a system that means something, you attach rewards to it, and the path of least resistance for earning those rewards slowly becomes “show up a lot.” The signal dies quietly over years, and more than likely the players will notice before you do.

The deeper design lineage here is worth naming, because every retention system in the industry descends from it. John Hopson’s “Behavioral Game Design” essay laid it out in 2001: “Each contingency is an arrangement of time, activity, and reward, and there are an infinite number of ways these elements can be combined to produce the pattern of activity you want from your players.” Every daily quest and streak mechanic shipped since is downstream of that sentence. These tools really do work, which is exactly why you have to decide what they’re for. Arrange time, activity, and reward to deepen mastery and you get players who are proud of their accounts. Arrange them to maximize sessions and you get players who are tired of your game and can’t quite articulate why.

They price your items against everything else in your game, against other games, against the broader market. And they do it instantly, whether the cost is money or time. You don’t get to set perceived value; you can only honor it or violate it.

There are two corollaries I’ve carried through every economy I’ve worked on. First: underpricing destroys value just as reliably as overpricing does. Make the prestige item too cheap or too easy and it stops being prestige, because players read effortless acquisition as a statement that the thing wasn’t worth anything. Scarcity, effort-to-reward, the long collection arc: these aren’t dark arts, they’re how worth gets communicated in a virtual economy. Keep the collecting pointed at cosmetics rather than power and the “gotta catch ’em all” drive is one of the healthiest engines you can build.

Second: scarcity needs an in-world reason. If something is rare, the rarity should be justified (seasonal, narrative, earned), not manufactured purely to spike engagement. Artificial scarcity with no story behind it reads as exactly what it is: manipulation. Players will forgive a limited item that means something, but they won’t forgive a countdown timer bolted onto a JPEG.

The Hextech tradeoff (agency for cost) is one psychology, and it doesn’t stop at loot. It’s the same contract in ranked entry costs and the same contract in commerce. The player gives up choice, stakes, or time in exchange for something they value, and the deal holds as long as it’s consensual and legible. The moment the terms get hidden, any of these systems curdles.

The improvement first: determinism beats gambling. The pass lineage runs from Dota 2’s International Compendium in 2013 through Fortnite’s battle pass in late 2017 to total ubiquity by 2020, and it largely replaced the loot box as the default monetization. The regulatory reckoning after Battlefront II accelerated the swap: Belgium and the Netherlands ruled paid loot boxes gambling in 2018, and EA told UK Parliament that loot boxes were “surprise mechanics... quite ethical and quite fun” in 2019, a line that did more damage to industry credibility than any regulator managed. With a pass, you know what you’re buying, and that’s genuinely better. For the receipts on where I stood: OpenCritic, which I co-founded, announced in October 2017 (mid-Battlefront II) that it would flag games with loot boxes, and shipped the warning label in early 2019.

Now the catch: the deadline replaced the dice. The pass traded variable-ratio reward psychology for deadline-plus-sunk-cost psychology: the dice are gone, and now you’re racing a clock. Bungie’s own director admitted on the record in 2020 that Seasons had too much FOMO (chapter 07 sits with that quote), and the 2023–25 fatigue discourse was the industry catching up to the confession. Meanwhile the anti-FOMO position proved itself commercially: Halo Infinite’s non-expiring pass in 2021, then Helldivers 2’s Warbonds and Marvel Rivals’ pass in 2024, selling “there’s no need to worry about not having enough time” as a feature. As it turns out, urgency isn’t actually load-bearing.

I wrote a heading in my book years ago that I’ll stand behind without softening: FOMO sucks. The full position has three parts. Scarcity needs an in-world justification, not a manufactured one. There should be second-chance routes, where a missed limited item can be obtained later at greater cost or challenge, so its value holds without the punishment. And then there’s the line I keep coming back to:

I wrote that sentence about loot-box-era mechanics, and it applies without modification to the deadline era. If your retention curve depends on players being afraid of what they’ll lose by leaving, that’s not a metagame, that’s a hostage situation with cosmetics.

I was leading Experience Design at Respawn when the Season 22 battle pass shipped in 2024 (two real-money-only passes per season), and sixteen days later the change had been fully reversed, because players priced it instantly, exactly as META 05 predicts. I’m not going to narrate the inside of that here; chapter 07 runs the tape.