The Gacha Architecture: When the Publisher Is Referee, Rule-Maker, and Payee

The twenty-first day of a banner phase is a deadline. After that mark, accumulated pulls do not disappear, but the chance to own the specific advertised character closes. For a player entering phase two of version 7.0 with premium currency already exhausted, the question is no longer "should I pull or not." It is a choice between wagering on a name already on the banner — Flins, Ineffa — and preserving a buffer for two characters mentioned only by name: Vesna and Vodyanitsa, expected to open phase one of version 7.1.

I follow this market from Berlin, where my daily work is transfer valuation and turning emotional decisions into verifiable probability ranges. When I look at a banner schedule, I see the exact structure I analyze in sports: a system designed, operated, announced, and profited from by a single entity. The publisher here does not merely sell content. They blow the whistle, write the rules, and collect the money from the very match they officiate.

Of the twenty-eight information points that make up this schedule story, twenty carry no source, only one cites an official publisher announcement, and three are the author's opinion. That is the first number I recorded, before touching any banner schedule. Numbers never lie — only the reader's heart turns them into lies.

LABEL IT CORRECTLY BEFORE ANALYZING

There is a professional mistake I encounter often in sports reporting: naming a thing incorrectly, then using the wrong analytical framework for it for the rest of the article. Here, the label that needs fixing is "esports." Genshin Impact is an open-world action role-playing game operated on a gacha model — players spend premium currency to pull for characters or weapons at random odds. Its correct label is gacha game, not esports.

That distinction is not a matter of wording. An esports discipline has a professional tournament circuit, a club system, a transfer market, competitive balance patches, and an ecosystem in which power is shared among publisher, organizer, teams, and broadcasters. A gacha title like Genshin Impact has no official tournament circuit, no clubs, no transfers, and its "versions" are Player-vs-Environment content drops, not balance patches. In other words, the entire nine-dimension esports framework — from rosters to transfers, from meta to tournaments — is inapplicable to most of the content under discussion.

So I will not manufacture false equivalents to fill a template. I will not call characters "players," I will not call banner phases "tournaments," and I will not assign Vesna or Vodyanitsa a "form curve" by career age. Names such as Vesna, Vodyanitsa, Flins, Ineffa, Skirk, Escoffier, Odette, Aino, Iansan, and Lan Yan are fictional in-game characters. Applying an injury-risk framework to them is a form of numerical sophistry, and a data monk is not permitted to falsify his own scripture.

What genuinely deserves analysis here is a revenue architecture. Each version splits into two phases, each phase runs about twenty-one days, and each phase has its own banner. This is a deliberately designed rhythm. It creates recurring, time-boxed spending windows that are short enough for players to feel pressure before they can accumulate again. When a system designer sets a rhythm like that, everything else — which character, strong or weak, worth it or not — becomes a variable dependent on that rhythm.

THE ARCHITECTURE OF THE MACHINE

To understand why a schedule carries so much weight, one must understand the three mechanisms beneath it.

The first mechanism is the guarantee threshold, or pity. On an event banner, every ninety pulls guarantee one five-star character. In other words, players always know the maximum cost ceiling for a single success, and that ceiling is published transparently. This is the key difference from a pure gamble. A gamble does not tell you how much you will lose. A system with a guarantee ceiling gives you a sense of control, while giving the publisher a predictable maximum price — and a predictable maximum price is one consumers are more willing to spend up to.

The second mechanism is the 50/50. The first five-star on an event banner has a fifty percent chance of landing on the advertised character and a fifty percent chance of landing on a standard-pool character. If you lose, the next five-star is guaranteed to be the advertised one. This is not a minor detail. It is a variance regulator. For the player, it turns the experience into a dramatic series of wins and losses. For the publisher, it turns average spending into a quantity higher than face value, because reaching the guarantee forces many players through both rounds.

The third mechanism is shared counter between banners of the same type. When same-category banners share one pity counter, switching from one banner to another is no longer starting from zero. This design lowers the marginal cost of switching segments. Behaviorally, when switching cost falls, switching frequency rises, and total spending over a long window tends to rise with it.

A pity architecture combined with a 50/50 mechanism is not a way of selling characters — it is a way of selling variance. The publisher does not price what you receive; they price the uncertainty of whether you receive it at all.

These three mechanisms operate inside a time rhythm. One phase runs about twenty-one days. That number is not random. It is long enough for an active player to accumulate part of the currency, and short enough that the rest of the spend falls beyond free accumulation. In other words, after every twenty-one days, players face a small shortfall — not too large to give up, not too small to avoid topping up. As a data person, I cannot help seeing an inverted "decay coefficient" here: a banner's perceived value does not decline over time; it spikes near the deadline and collapses to zero right after.

Outside the twenty-one-day rhythm lies another policy, and this is where I believe the highest analytical value sits. The rerun schedule — the schedule for bringing older characters back to limited banners — is not fixed. Some characters are absent for more than a year, while others return within a few versions. This lack of fixity, from a system-design standpoint, is a deliberate scarcity mechanism. When you do not know when the character you want will next appear, the cost of missing this round rises, and the opportunity cost of waiting rises too. This is the logic of every limited-time event, converted into a revenue structure.

The counterweight to that scarcity mechanism is a secondary revenue lane. Chronicled Wish is a separate banner type with its own rules, typically for older characters. Its existence implies something architecturally important: the publisher can re-monetize characters that have gone dormant commercially without disrupting the cadence of main banners. This is a secondary lane for legacy assets, allowing old revenue not to be forgotten while keeping main banners the site of peak spending pressure.

Put it all together and you get a self-running machine. You have a fixed time rhythm creating spending windows. You have a guarantee ceiling making cost acceptable. You have a variance regulator raising average spend above face value. You have a shared counter reducing switching friction. You have a scarcity policy creating action pressure. And you have a secondary lane to re-monetize legacy assets. No link in this chain depends on an external cultural or sporting event. This machine runs on its own calendar.

This is why I want to place it beside the esports ecosystem for comparison. An esports ecosystem earns from sponsorship, broadcast rights, in-game item revenue sharing, and partly from prize pools. It depends on a chain of intermediaries: publisher, organizer, teams, broadcasters. That chain creates durability but also bottlenecks and coordination risk. By contrast, the gacha model is a direct loop between publisher and spender, with no intermediary. It depends less on the cultural calendar, and is therefore less exposed to calendar shocks. But it exposes itself to a different risk: legal risk. Every gacha structure sits close to the gray zone of probability-transparency and consumer-protection regulation, especially for underage players.

Here I want to pause on a power-concentration observation. In this system, the publisher is simultaneously the game operator, the gacha rule-maker, the probability discloser, and the direct beneficiary. No independent referee verifies those disclosures, and there is no arbitration mechanism to adjudicate disputes. Every publicly stated rule is published by the same party that issues it. That is a higher degree of power concentration than in most esports ecosystems, where even if the publisher remains powerful, organizers and other stakeholders still have a voice.

At this point, my usual line holds in another sense. A transfer is not buying a person; it is buying a probability distribution. With banners, pulling works the same way: it is not buying a character, it is buying a probability distribution designed by the seller. And when the seller designs, operates, and announces the distribution, the buyer has only one correct defense: understanding exactly what is being bought.

THE COUNTERINTUITIVE ANGLE

There is a strong temptation in reading schedule-style articles: turning the schedule into value. When the official announcement for version 7.1 states that phase one debuts two new characters at once, the community immediately reads it as "7.1 is worth saving for." I believe that is a logical leap that is not permitted.

A schedule is not power. A phase with two new characters says only one thing: allocation pressure in that phase is higher than usual. It says nothing about how strong those characters are, which compositions they fit, or whether they open any new playstyle. The source analysis provides not a single line of data on strength or kit. It provides only the schedule. And a schedule without accompanying strength data is a schedule, not a recommendation.

I recognize this bias because I have committed it. In my role as a transfer valuer, I have had to say no many times to names booming after a short tournament. A star who flares across six matches at a major can look more attractive than a striker scoring steadily across three seasons. But when you build a model on a sufficiently large dataset, the right choice is often the boring one. Short-term dazzle is a variable, not an indicator. With gacha, the same phenomenon plays out: a simultaneous debut of two new characters generates high dazzle, but until strength data appears, there is nothing to value.

There is a false label that needs correcting here, and it relates directly to information quality. All twenty-eight information points of this schedule story reveal a familiar structure: little official information, high community expectation, and most forward-looking claims without verifiable sources. Every crisis is unlabeled data — and here, what is unlabeled is not a crisis, but a simple fact: claims about future versions are provisional.

The source analysis itself concedes this when it states that the exact banner schedule is still to be confirmed. To a data person, that is a credible signal. A document willing to say "I am not sure" is worth more than a document asserting certainty without sources. But it also means every conclusion drawn from that document must be labeled provisional.

I do not believe in intuition — I believe in the decay coefficient of intuition. And the decay coefficient of dazzle in this case is running very fast: it peaks on announcement day and falls back to baseline the moment the twenty-one-day phase ends. Anyone treating that dazzle as a value indicator is reading an emotional curve and calling it a data curve.

WHAT TO WATCH

There are three signals I will track, and all are publicly observable.

First is the official announcement of version 7.1 banners from the publisher's channel. This is the only signal that can confirm or refute the entire circulating schedule. Second is entity verification. When a character appears in official materials, the source document's reliability is established; when it disappears, that document downgrades itself. Third is regulatory developments on probability transparency and loot-box-related policy. Any change at that layer directly alters the revenue model analyzed here.

What I want to leave is not a conclusion, but a question. If a revenue system can make millions of people decide based on a calendar rather than on data, is calling it a "market" accurate enough — or are we calling an emotional curve by the name of a price curve? Some matches end when the referee blows the whistle — and some only begin when the data speaks.