The Invisible Referee Never Blows the Whistle: Gacha, Pity, and a Revenue Structure Esports Doesn't Have

In the evening in Busan, I reopened my tracking sheet and found a column of data that did not belong to football. It was the banner schedule for Genshin Impact, running from phase two of version 7.0 into phase one of version 7.1. In that sheet there was a threshold very familiar to anyone doing analysis: 90. The pity threshold guaranteeing a five-star character. But what made me stop was not the 90. It was the 50/50 split right in front of it.

This is the kind of number that makes me sit down longer than usual. A clear ceiling, a halved probability, and a guarantee mechanism behind it. Put together, those three things form a complete pricing structure presented as probability. I look at xG, then at the scoreline, and I learned not to trust either. The same applies here: I look at the rate, then at the pity threshold, and I learned not to trust either the way the publisher wants me to.

This piece does not aim to tell anyone whether to pull or not. It aims to dissect a revenue machine, place it beside the revenue machine of esports, and point out where the two are similar enough to be confused, and where they are different enough that they cannot be merged into one analytical frame.

Context: reading a gacha document with a data analyst's eye

My job is to read match data: xG, PPDA, advanced metrics, transfer valuations. When I moved to reading documents about Genshin Impact's banner schedule, I did not try to turn characters into players or banner phases into rounds. Doing so would violate the first principle of my trade: never build a false equivalence to fill a gap.

Genshin Impact is published by HoYoverse and operates on a gacha model. It has no official professional tournament circuit, no franchised club system, no transfer market in the esports sense, and no competitive-balance patch system in the sense I am used to. Its versions are PvE content drops, not competitive balance patches.

In other words, the nine-dimension framework I still use to analyze esports does not apply to most of this document. Most dimensions must be marked as insufficient information or not applicable. But there are four genuinely overlapping zones: publisher revenue, gacha regulation, scarcity dynamics, and source quality. Those are the places where I can say something of value without inventing a bridge.

On source quality, I have to be blunt from the start. Of the 28 information points the document provides, 20 carry no source. Only one cites an official source, an announcement from Genshin Impact itself. Three are the author's personal opinion. Several names such as Odette, Flins, Ineffa, Vesna, and Vodyanitsa, along with version numbers 7.0 and 7.1, cannot be cross-verified against the game state I know. This is a high risk that the content is auto-generated, speculative, or drawn from rumor.

The document itself concedes this at one point: the exact banner schedule is still to be confirmed. To me, that is a positive integrity signal and, at the same time, an admission that everything above it is provisional. I record both facts in my sheet before analyzing, because a number without source context is not data. It is a trap.

Version rhythm: a season with no matches

Each Genshin Impact version splits into two phases, each lasting roughly 21 days. Each phase has its own banner. This is a content rhythm, and at the same time a revenue rhythm.

When I look at the number 21 days, I see something different from a match calendar. A match calendar exists to allocate competitive resources along the time axis. The 21-day rhythm here exists to create recurring purchase windows. Players are not allowed to buy at any time. They are allowed to buy inside a time-boxed window, after which the window closes.

An empty stadium does not remove football; it only reveals variables we used to overlook. Here too there is an overlooked variable: time pressure. In the crowdless Bundesliga data I once collected, home win rate fell from 43 percent to 31 percent, while average goals per match rose from 2.7 to 3.1. The crowd is a variable. In gacha, the countdown timer is a variable with a similar function: it changes players' decision behavior, not by changing the rules, but by changing their perception of time.

Structurally, the currency-allocation pressure peaks in phase one of version 7.1, when two new characters, Vesna and Vodyanitsa, arrive at the same time, while phase two is reruns. This structure funnels spending decisions into a narrow window, turning phase one into a liquidity bottleneck for most ordinary players.

This is an observation about revenue architecture, not about character strength. The document provides no data on kits, on actual power, or on how necessary the characters are for a roster. It provides only a schedule. And a schedule answers the question "when," not the question "should I."

I noticed a small detail in the framing: the document describes the new version as an exciting adventure in Snezhnaya. That framing is promotional. In a gacha economy, opening a major content region typically coincides with a period of elevated spending pressure, because returning players come back, new players join, and the perceived value of banners rises with it. That is my hypothesis, not a fact from the document, and I mark it as low-confidence.

The pity architecture: price presented as probability

This is the part I want to spend the most time on, because it is the only part of the document with a structure clear enough to analyze as a mechanism.

The pity threshold is 90 pulls for a five-star character. On an event banner, the first five-star has a 50 percent chance of being the featured character and a 50 percent chance of a standard character. If the result falls on the standard branch, the next five-star is guaranteed to be the featured one. Finally, pity is shared across banner types of the same category.

Those three lines describe a complete pricing structure. Let me break it down.

First, the ceiling of 90 creates a sense of safety. Players know the maximum cost of one character. This is an important psychological function: it turns a random event into a budgetable expense. The publisher accepts giving up the long tail of the distribution in exchange for players being more willing to spend.

Second, the 50/50 rate reintroduces variance. If there were only a 90 ceiling and no 50/50, the expected cost would be nearly fixed, and revenue would become as steady as a subscription fee. The 50/50 restores variance. Those who win the 50/50 pay less; those who lose pay double. Total revenue is pushed up by the very uncertainty the publisher just reduced at the ceiling layer.

Third, the guarantee mechanism after a loss keeps players in place. Someone who just lost the 50/50 and knows the next one is guaranteed has a very strong incentive to top up within the same window. This is a behavioral anchor, not a probability anchor.

Fourth, shared pity across banners of the same category lowers the marginal cost of switching between options. Players are not penalized for changing their minds between a new banner and a rerun banner. In cash-flow terms, this smooths revenue across both window types and likely increases overall spending frequency. I mark this as a medium-confidence inference, because the document describes the mechanism but offers no behavioral data to verify it.

Put those four layers together and you have a design that is very elegant technically. It balances perceived accessibility and revenue variance: the ceiling is low enough that players dare to enter, the halved probability is high enough that they spend more than they planned, and the guarantee is firm enough to keep them around after a loss.

This is where I have to be careful in exactly the way I am with xG. An elegant structure does not equal a correct conclusion about behavior. I have the mechanism; I do not have behavioral data. A high correlation between design and revenue is a correlation, not causation. If I cannot explain the intermediate mechanism, I am only allowed to record it as an observation, not a conclusion.

Reruns and scarcity: a transfer window with no contracts

The document notes an important feature: there is no fixed rerun schedule. Some characters are absent for more than a year, others return after only a few versions. Players waiting for Skirk or Escoffier may wait a very long time, while other players are served faster.

In esports and football, the transfer window has a published opening and closing date. That transparency is part of the rules of the game. Here, the lack of transparency about the rerun schedule is not an operational shortfall. It is design.

I once spent three years tracking the youth transfer market and writing about the price bubble for players who had not yet played 50 top-flight matches. What do that market and the rerun mechanism here have in common? Both operate on expectation, not on verified value. In the transfer market, clubs pay for potential. Here, players pay for the ability to own something with no known return date.

The essential difference: in transfers, the club and the player are two parties with bargaining power, and the contract specifies a term. In gacha, there is no contract. Players do not sign anything. They only buy a ticket into a window that opens without notice and closes without notice.

Scarcity here does not come from a shortage of supply. It comes from the publisher controlling the timing of supply. That is a pure FOMO mechanism, and it works because it strikes at perceived opportunity cost, not at price.

People call surprise reruns luck. I call them a pre-solved equation. It is just that the equation is not published for players to see.

Chronicled Wish: a second revenue lane

The document mentions a separate mechanism called Chronicled Wish, operating under its own rules and typically for older characters. I read this detail as a signal of a multi-layered revenue architecture.

With only a main banner, the publisher has to choose between rerunning an old character and launching a new one on the same time budget. Each rerun occupies a window. Each window given to an old character is a window not given to a new one.

Chronicled Wish solves that by opening a separate lane. Old characters are re-monetized without taking up the main banner's rhythm. This is something I observe across many digital content business models: when an old product still holds value but cannot stand on its own, a secondary channel is created to sell it.

The confidence of this inference is low, because the document states only the mechanism's existence, not operational data. But it is worth recording because it shows the publisher managing a product portfolio the way a content company does, not the way a tournament organizer does.

The publisher as both referee and beneficiary

This is the point I find most worth discussing when placed beside esports.

In a mature esports ecosystem, the publisher sets the rules, but third parties participate: tournament organizers, teams, player associations, broadcasters, auditors. Power is distributed, though unevenly.

In this document, the publisher sets the gacha rules, publishes the schedule, operates the game, and is the sole party collecting money from that system. There is no independent referee checking whether the published rates match the operating rates. There is no third party verifying that the pity threshold works as described.

The only information point with an official source in the entire document is an announcement from Genshin Impact itself. In other words, the rule-maker is also the rule-publisher, and also the beneficiary of the rules.

I am not saying this to assign blame. I am saying it to put it into the right analytical frame. The concentration of power here is higher than in most esports ecosystems, because in esports there remain parties with different interests and incentives to monitor one another. In this model, the incentive for internal oversight is nearly zero.

Regulatory risk: when probability becomes a public matter

The document describes probability mechanics in detail but cites no regulator. That is a notable gap, because probability mechanics in gacha are subject to regulation in many markets.

Common requirements include rate disclosure, transparency of the guarantee mechanism, and protections for minor players. This document inadvertently describes exactly what those frameworks target: the 50/50 rate, the 90 threshold, the guarantee, shared pity. It only lacks the regulatory citation.

As an analyst, I assess regulatory risk as medium over the medium term. The reason is not that this mechanism violates anything, but that the gacha model in general sits close to debates about loot boxes and consumer protection. A regulatory change in one large market can force mechanism adjustments globally.

This leads to a valuable comparison with esports. Esports depends on match calendars, sponsorship contracts, broadcast rights, and sporting-cultural events. Gacha depends on direct, recurring, in-game spending. The gacha machine is less vulnerable to schedule shocks than esports, but it exposes itself more to changes in rules on probability and consumer protection. Two risks of different natures cannot be measured with the same ruler.

The contrarian angle: the "esports" label and the correlation trap

This is the part I want to give to those reading this as content people.

The original document was labeled "Esports." That label is wrong. Genshin Impact has no official professional circuit, no franchise league, no club ecosystem, no transfer market in the esports sense. Its versions are PvE content drops, not competitive balance patches.

This wrong label is more dangerous than it looks. If an analyst accepts it without checking, they will start looking for metrics that do not exist: win rate, pick-ban rate, roster strength, player form. Finding none, they will invent them by assigning roles to fictional characters. And so a false analysis is born from a false label.

Three years, two World Cups, one question: is data born to understand football or to hide it? That question applies here intact. Is banner-schedule data born to help players understand when to prepare, or to hide the fact that there is no data at all on a character's true value?

At a more macro level, there is a very easy correlation trap. The publisher announces the banner schedule. The parent company's stock rises. A quick conclusion would be: a good banner schedule lifts the stock. But what is the intermediate mechanism? How does the banner schedule affect cash flow, with what delay, through which channel? If I cannot answer, I am only allowed to say two things happened together, not that one caused the other.

Morocco does not need to hold the ball much; they need to hold it in the right place. The same applies here. The publisher does not need players to spend a lot each time; they need players to spend at the right time. And that "right time" is defined by them.

The source-reliability problem: something I cannot skip

I have to return to this once more, because it affects every conclusion above.

Twenty of twenty-eight information points have no source. Many character names and version numbers cannot be verified. Only one point cites an official source. Three are personal opinion.

For someone who reads data for a living, this is a high risk level. When I analyze a match, I never assert a tactical trend after a single game. I require cross-verification over at least two seasons. I learned that after Euro 2026, when I wanted to write immediately about a new winger archetype based on Lamine Yamal's data at a short tournament, and was told to wait for the following La Liga season to verify. I was frustrated, but I complied, and I recognized the value of precedent.

Applying that principle here: a document where 71 percent of information points lack a source does not qualify as the basis for any conclusion about the game's future. It qualifies only as a case study in how promotional content is produced.

This is what I find most interesting from a media standpoint. The document has a promotional tone, an appealing narrative frame, and numbers that look very specific, yet a very thin verification foundation. This kind of content exists to keep readers inside the hype cycle before a new version, not to provide actionable information.

I am not writing this to criticize the document. I am writing it to place it correctly in my content taxonomy: traffic-filter content, not reference content.

What I take away

Putting it all together, the picture that emerges is not a game with a tournament, but a revenue machine that runs itself. The 21-day rhythm creates windows. The 90 threshold creates a sense of safety. The 50/50 rate creates variance. The guarantee retains players. Shared pity smooths cash flow. An unfixed rerun schedule creates scarcity. The secondary banner lane re-monetizes an old portfolio. And the publisher sits in all four seats: setting the rules, publishing the rules, operating the rules, collecting money from the rules.

That is a very effective structure, and also a very concentrated one. Compared with esports, it depends less on third parties and is less exposed to schedule shocks, but it is more exposed to regulatory change and more exposed to transparency problems.

I entered the trade for the numbers, but I stayed for the stories the numbers do not tell. The story here is a story about design: how a system is built so that every 21-day window has a reason for people to step in.

For me, the signal to track in the next cycle is not who will be rerun. It is whether regulators begin requiring independent audits of probability mechanisms. If that happens, the entire pity architecture will have to be presented differently, and only then will we have real data to analyze instead of just a schedule.

For now, I keep my tracking sheet. The left column holds the published thresholds. The right column holds the unverified parts. The gap between those two columns is where I work.