90-Pull Pity and the 50/50: What Genshin Impact's Gacha Engine Teaches Sports Business

90-Pull Pity and the 50/50: What Genshin Impact's Gacha Engine Teaches Sports Business

In 2026, I sat in the third row of LoL Park in Seoul, at a match with no audience at all. No cheering. No banners. Only the clatter of keyboards and the quick breathing of the players. I wrote in my notebook: this is a stadium that is still alive, just with nobody watching it.

Six years later I sat next to a young friend in Gangnam, watching him count every pull on his phone screen. Pull eighty-nine came up empty. On pull ninety, a golden flash, a character appeared, and he let out a long breath as if he had just escaped danger. There was no grandstand there. But there was something more crowded than a grandstand: a system designed so that every single pull has someone waiting.

I have been an esports commentator long enough to recognize that I was looking at a revenue engine that runs far more smoothly than most of the models leagues currently use.

The subject here is Genshin Impact, an open-world role-playing game published by HoYoverse and operated on a gacha model. One thing must be said up front: this game has no professional tournament circuit, no clubs, and no transfer market in the esports sense. Its version cycle is a content-release cycle, not a competitive balance patch.

According to the compiled document I have, version 7.0 phase two was given to rerun banners, including Flins and Ineffa. Version 7.1 phase one is said to open two new characters at once, Vesna and Vodyanitsa; phase two returns to reruns, with Skirk and Escoffier named. Each version splits into two phases of roughly twenty-one days.

On the rules: players spend in-game currency to pull, and a guarantee system — pity — ensures a five-star character within ninety pulls. The first five-star has a fifty-fifty split between the featured character and a standard character; if a standard character appears, the next five-star is guaranteed to be the featured one. Pity is shared across banners of the same type. There is also a separate banner type for older characters, called Chronicled Wish, with its own rule set. The rerun schedule is not fixed: some characters are absent for more than a year, while others return within a few versions.

I should add something the document itself admits: of twenty-eight information points, most carry no source, and only one cites an official publisher announcement. Several character names and version numbers cannot be independently verified. That detail becomes part of the analysis below rather than a footnote added for form's sake.

The first point lies in the architecture of the guarantee. Placing a hard ceiling at ninety pulls serves one specific purpose: lowering the psychological barrier before a buyer steps in. When buyers know a ceiling exists, they are more willing to take the bet. But sitting inside that ceiling is a fifty-fifty mechanism that generates variance. Variance determines the average number of pulls a person ends up spending, and it does not lean toward the person pulling.

The guarantee makes buyers feel safe; the variance makes them pay more than they think.

The second point is shared pity across banners of the same type. On the surface this looks like a decent detail. Look closer and it lowers the marginal cost of switching from one banner to another, which means it raises the frequency of spending decisions. It is a mechanism for smoothing revenue across the new-character window and the rerun window.

The third point, and in my view the most valuable one: the rerun schedule is not fixed. For a sports league, the fixture list is a promise. Fans know the day, the hour, the opponent, and they buy tickets on that certainty. With gacha, uncertainty is the product. Nobody knows when the character they are waiting for will return, so nobody can plan; and because nobody can plan, people stockpile, worry, and pull the moment the window opens.

The fourth point is Chronicled Wish. A secondary revenue lane for older characters, enough to re-monetize their value without disturbing the rhythm of the main banners. This is the problem every commercial department of a league dreams about: selling memory back without having to change the fixture list. In the transfer era, people buy players, but they sell memories.

The fifth point concerns power. In this value chain, the publisher is at once the rule-maker, the operator, the announcement authority, and the beneficiary. There is no independent arbiter verifying the published rates. In esports that concentration of power also exists — the publisher decides patches, schedules, revenue shares — but at least there are clubs, players, professional associations and third parties as counterweights. In the gacha model, the counterweight is close to zero.

Based on my experience following matches across many seasons, I have seen plenty of ways to convert loyalty into revenue: jerseys, tickets, item bundles, broadcast rights, prize pools. None has been as direct and as repeatable as this model. Esports lives on sponsors, rights fees, in-game item revenue shares and prize money — all of it dependent on the calendar, on the presence of an audience, on the cultural rhythm around it. Gacha lives on direct, recurring spending with no intermediary. It is less vulnerable to schedule shocks — like the audience-free stretch I once sat through — but it exposes itself to regulation on randomized rates and the protection of underage players.

An empty stadium still carries the applause of a generation we have never met.

There is a story the community keeps telling itself: save up, wait for version 7.1, because a great adventure is waiting there. I understand the pull of that story. But looking at the data, the document confirms exactly one thing: the schedule. There is not a single line about strength, about kit, about the practical value of the upcoming characters. A schedule is not a value. We are mistaking the question of when for the question of why.

This is where I want to be blunt, because I once mispronounced a syllable and realized I had mispronounced an entire career: the source document is labeled esports, yet inside it there is no team, no player, no balance patch, no tournament. That is a classification error. We call anything with a crowd and no touchline esports, and then we build business strategy on a mistake about categories.

The risk worth worrying about here does not sit in the puller's wallet. It sits in the fact that most of the information carries no source, that some character names and version numbers cannot be cross-checked, while the tone is very confident. The most honest sentence in the whole document is the one admitting the exact banner schedule is still to be confirmed. A reader making a decision on that schedule is acting on a rumor written in a confident voice.

The generation of fans not yet born will read our spending logs and banner calendars the way we read an artifact. They will not ask which character was strong. They will ask what we built: a playground where presence is measured in money, or a place where memory is measured in time. I do not judge anyone who pulled. I only ask myself whether I am writing precisely enough that generation will not have to correct me.

This article is based on public information and a compiled analysis document, for reference only; it does not constitute investment or betting advice.