The Architecture of Ninety Pulls: A Revenue Rhythm Built to Need No Stadium

The third night of the launch week for update 7.0, I sat in a small apartment in Hamburg, a recording of an old match on my left screen so I could still hear the commentary, and a banner schedule for a completely different game on my right. No stands, no cheering, no faces in the locker room. Just a number blinking over and over: ninety pulls.

I have spent nine years following teams from training ground to arena corridor, recording what remains after the lights go out. But that night, for the first time, I realized there is another kind of rhythm running in parallel to the rhythm of football and esports. A rhythm that needs no one to cheer, no extra time, no explosive moment to sell tickets. It only needs time, patience, and a guarantee threshold written clearly in the terms.

I do not analyze matches; I remember each face when the match ends. But here there was no match to end. And that very absence is what made me sit down and start typing.

Context: a market mislabeled

Before any detail, I must state something it took me days to admit. The source document I am analyzing came from an article tagged "Esports." But reading closely, the entire content revolves around Genshin Impact's banner schedule and the pity mechanics of its gacha system. This is not esports. It is an open-world action role-playing game operated on a gacha model, published by HoYoverse.

This matters for a very specific reason: Genshin Impact has no official professional tournament circuit, no franchised league, no club ecosystem, no player-transfer market in the esports sense, and no competitive balance patch system as esports understands it. Its "versions" are PvE content drops, not competitive-balance patches.

In other words, the nine-dimension framework I normally use for League of Legends, Arena of Valor, or Teamfight Tactics is structurally inapplicable to most of this content. If I tried to force it — treating characters as "players" and banner phases as "tournaments" — I would manufacture a false equivalence, violating my own principle of transparent sourcing.

So why write about it? Because the mismatch itself tells a story. And because, as a training-ground observer, I believe understanding how a revenue model works matters no less than understanding how a team works.

The source document records 28 information points. Of these, 20 have no source, only one cites an official Genshin Impact announcement, and three are explicitly the author's opinion. Several named entities — Odette, Flins, Ineffa, Vesna, Vodyanitsa, and versions 7.0 and 7.1 — cannot be cross-verified against known game state. This is a worrying reliability signal, and I will return to it at the end.

But first, let us talk about structure. Let us talk about the number ninety.

The core: the architecture of a revenue machine

When I counted how many times a midfielder turned his head to check his shoulder before receiving the ball, I learned that behaviors repeated hundreds of times reveal more than explosive moments. The same applies here. The pity mechanic is not a dry technical detail. It is the spine of the entire system.

Specifically, the system guarantees a five-star character within a maximum of ninety pulls. On an event banner, the first five-star has a 50% chance of being the featured character and a 50% chance of being a standard-pool character. If the result falls into the standard pool, the next five-star is guaranteed to be the featured character. This is the mechanism the community calls "the 50/50."

The second notable point is that pity is shared across banner types of the same category. That is, the pulls a player has accumulated on one banner still count when they switch to another banner of the same type. This is a small technical detail with a large revenue consequence: it lowers the marginal cost of switching between banners, and therefore likely increases overall spending frequency.

Now place these three elements side by side: a soft guarantee at ninety pulls, the 50/50 mechanic, and shared pity. This is the pricing structure of the gacha model — not a contract, but in design logic it is equivalent to a dynamic price list.

The ninety-pull threshold creates a sense of accessibility. Players know that no matter how unlucky they are, they need not pull infinitely. The number ninety turns a probabilistic behavior into a calculable plan. That is basic psychology: people spend more readily when they believe there is a defined stopping point.

The 50/50 mechanic, meanwhile, creates high variance. A player might get the desired character on the first try, or might have to pull twice the expected number. This variance is the core revenue driver of gacha. If every player got the character at pull ninety exactly, revenue would be stable but low. If outcomes were fully random with no threshold, players would quit out of a sense of unfairness. The combination of a guarantee threshold and variance is precisely the equilibrium the publisher needs.

And shared pity is the mechanism that smooths revenue. It keeps players from feeling they have lost something when switching from a new-character banner to a rerun banner, or vice versa. The result is that cash flow is not interrupted between debut and rerun windows.

If I had to choose one sentence for this section, it would be: the entire gacha architecture is designed to balance the player's sense of fairness against the publisher's revenue variance. That is not an arbitrary system. It is a carefully calculated one, and every parameter serves a specific purpose.

Two phases, twenty-one days, and the breathing of a cycle

At St. Pauli, I learned that a training session has its own heartbeat. Players warm up slowly, ramp up intensity, peak in the middle, then cool down at the end. That structure repeats daily, and the repetition itself builds collective strength.

Genshin Impact's banner cycle operates on similar logic but at a larger scale. Each version is split into two phases of roughly twenty-one days, each with its own banner. This is a designed revenue rhythm, not a random schedule.

Twenty-one days is an interesting number. It is long enough for players to accumulate free in-game currency, creating a sense that opportunity is available if they are patient. But it is also short enough to create time pressure, making players anxious that if they miss this window, they will face an indefinite wait. The tension between these two feelings is the driver of spending behavior.

According to the document, the first phase of version 7.1 is said to introduce two new characters at once, Vesna and Vodyanitsa, while the second phase consists of reruns. If accurate, currency-allocation pressure peaks in phase one. Players wanting both new characters face a hard choice: pull both, choose one, or skip both to save for the future.

This is an observation about revenue architecture, not competition. But it is worth noting, because it shows the publisher understands how to create tension in consumption behavior. Placing two new characters side by side in one phase is no coincidence. It is a design decision.

I recall a training session with the St. Pauli U19s, when the fitness coach scheduled two heavy drills back-to-back without full recovery. He explained he wanted to see which players could manage their energy under pressure. Some were overloaded, but those who endured improved dramatically. I am not saying the game publisher is training players. But I am saying both use the same psychological principle.

A season without cheering, only the patter of boots on grass. In the gacha world, the gaps between banner phases operate the same way. That is when players calculate, regret, or prepare for the next round.

Reruns, scarcity, and a policy with no fixed schedule

One feature described in the document is the absence of a fixed rerun schedule. Some characters may be absent for over a year, while others return after only a few versions. This is a deliberate scarcity mechanism, the gacha equivalent of limited-time event design.

From a revenue standpoint, this is a rational choice. If every character had a clear rerun schedule, players could plan long-term and defer spending. But if the rerun schedule is unpredictable, players face fear of missing out, and that fear drives immediate pulls.

I have seen this mechanism in another form in sports. When a club announces it will not renew a key player's contract, fans rush to buy the jersey bearing his name before it is too late. That is not essential demand. It is a reaction to impending loss. Gacha publishers understand this well, and they design their systems to exploit that feeling.

The document also mentions a mechanism called Chronicled Wish, a separate banner type with its own rules, typically for older characters. Its existence suggests a secondary revenue lane for legacy characters, reducing pressure to rerun them on primary banners. This signals the system's sophistication: the publisher does not just sell new characters, but re-monetizes old ones in different ways.

Here I want to emphasize what the document calls a "revenue-smoothing effect." When you combine an unfixed rerun policy with shared pity, you get a system that produces spending spikes around unpredictable reruns while maintaining a stable baseline cash flow. This is a strategy adjacent to sports sales techniques, but at far greater scale and with no sports event needed to trigger it.

When the stadium is empty, I understand whom I am keeping rhythm for. In this case, the rhythm is kept for a self-operating revenue system that needs no audience.

Concentration of power: the publisher as rule-maker, operator, and announcer

One point in the document made me pause longest. It is the observation that the publisher is simultaneously the maker of gacha rules, the operator of the game, and the sole official source announcing those rules. In the source, only one data point cites an official Genshin Impact announcement. Every other pity rule is described without any independent verification body.

This is a remarkable concentration of power. In most esports ecosystems, even if the publisher wields great power, other stakeholders exist — teams, tournaments, media organizations, and sometimes regulators. But in this pure gacha model, the publisher controls nearly both supply and information.

What does this mean? It means there is no independent arbiter to verify that the announced probabilities are accurate. It means that if the publisher changes pity rules in an update, players can only learn through the publisher's own announcement. And it means the entire system operates on trust, not on verification.

This is where the esports analytical framework can genuinely contribute a perspective, even if only by analogy. In esports, the tension between the publisher as rule-maker and the publisher as commercial beneficiary is a familiar theme. But in gacha, that tension is even less controlled, because no competitive ecosystem generates external transparency pressure.

The document also notes that pity rules and probability disclosures echo mandatory gacha probability-transparency requirements common in several jurisdictions. This is important because it shows that even in a system of concentrated power, external forces — specifically consumer-protection and minor-protection regulations — exert some pressure.

But the document also concedes that the source content does not mention any specific regulator. This is a notable information gap. It means we are analyzing a rule system without full legal context. And that, once again, reminds me of the importance of source verification.

The contrarian angle: when "esports" is not esports

This is the section I want to spend the most time on, because it touches a broader problem of how we classify and consume information.

When I read the source, the first thing that struck me was not the content about banners or pity. It was the "Esports" label affixed to an article about a game with no esports ecosystem. This is a misclassification, and in my view it is more dangerous than it appears.

Why? Because misclassification corrupts every downstream analysis. If an automated system or an editor treats this content as esports, they will apply inappropriate frameworks, draw unsupported conclusions, and ultimately erode reader trust in the entire category.

But there is a contrarian angle here. This misclassification is not merely a technical error. It reflects a broader trend in game media: the boundary between "game" and "esports" is being deliberately blurred, because the "esports" label attracts more attention than "entertainment game."

I have seen this before. A mobile game with a simple leaderboard is advertised as "esports." An RPG with a PvP mode is called a "professional arena." Each time, part of the term's meaning is eroded.

And the irony is that when we mislabel content, we do not just ruin the analysis of that content. We also miss what it could genuinely teach us.

Because, setting the "esports" label aside, the source actually contains a valuable lesson in revenue design. The pity architecture, the 50/50 mechanic, the unfixed rerun policy, and Chronicled Wish — all are examples of how a publisher can build a self-operating revenue machine independent of any sports event, tournament, or season.

Compare this with esports. In esports, revenue comes from many sources: sponsorship, broadcast rights, in-game item sales, tournament revenue sharing, and player transfers. Each source depends on an external ecosystem — teams, tournaments, audiences, and a match calendar. If a tournament is canceled, if a team dissolves, if a star player retires, cash flow can be affected.

In gacha, none of that exists. The revenue model is self-sustaining. It needs no season. It needs no champion. It only needs players to keep pulling.

This is why I say misclassification is not merely a technical issue. It is a symptom of a deeper confusion about the nature of different revenue models in gaming. And if we cannot distinguish them, we cannot analyze them accurately.

A fan's pain needs no tactics to be heard. In this case, I would add that a player's joy needs no esports label to be understood.

The reliability problem: 20 of 28 data points have no source

I need a separate section for this, because it affects every other conclusion.

The source records that of 28 information points, 20 carry "Source: None." Only one cites an official source (a Genshin Impact announcement), and three are explicitly the author's opinion. Several named entities — Odette, Flins, Ineffa, Vesna, Vodyanitsa, and versions 7.0/7.1 — cannot be cross-verified against known game state.

This is a red flag. And it is not just a problem of this particular article. It is a problem of an entire content ecosystem.

I have spent nine years in an environment where sourcing is everything. When I write about a player moving from the Bremen academy to St. Pauli, I must verify his name, birth date, position, and shirt number. When I write that the German national team slowed to an average of 14 steps per minute at the 70th minute, I must rewatch the entire match to count. There is no room for guesswork in specific numbers.

But in game content, especially content about gacha games, this standard is often lowered. Articles are produced quickly, character names are guessed, and schedules are "leaked" from unverifiable sources. The result is a sea of information in which fact and rumor mingle, and readers have no way to tell them apart.

Notably, the source itself concedes uncertainty. One information point states that "the exact banner schedule is still to be confirmed." This is a positive reliability signal, but also an admission that the content is provisional.

Looking at the whole picture, I see a familiar pattern: promotional tone, largely unsourced information, and several details likely generated or rumor-based. This is what I call "traffic-filter" content — content designed to attract clicks rather than deliver accurate information.

And this is where I must be careful with myself. My nature is to forgive before judging. I understand that writers of this content may lack the resources or time to verify everything. But forgiveness does not mean concealment. I must say clearly: if you read an article about a banner schedule and it cites no official source, you are reading a hypothesis, not a fact.

And in a context where players' spending decisions rest on this information, the reliability problem is not merely academic. It has real consequences.

What is genuinely transferable

After setting aside everything inapplicable, I want to summarize what is genuinely valuable in this document.

First, the pity architecture is a clean, teachable example of revenue design. The ninety-pull threshold, the 50/50 mechanic, and shared pity form a system balancing fairness perception against revenue variance. This is an evergreen reference for anyone interested in consumer psychology.

Second, the unfixed rerun policy combined with Chronicled Wish is a useful illustration of scarcity-based revenue and legacy-character re-monetization. This is a strategy other publishers can learn from, and analysts can use as a comparison point.

Third, the source is a cautionary example of "traffic-filter" content — promotional tone combined with thin sourcing. This is a media-literacy lesson with immediate value.

And fourth, most important to me, is the misclassification. It reminds me that even in the work of a training-ground observer, correct classification is the first step of any analysis. If you do not know what you are watching, you will not know what you are looking for.

The first rhythm is not made with the feet, but with the ears. In this case, the first rhythm was realizing I was listening to a different song than I thought.

What to watch next

As I sit here in Hamburg on an annual-season evening, I wonder what happens next.

There are signals to watch. Official confirmation of version 7.1 banners from HoYoverse channels will confirm or refute the source's claims. Verification of entities such as Odette, Flins, Ineffa, Vesna, Vodyanitsa through official character reveal archives will validate or invalidate the article's reliability. And developments in gacha regulation — probability disclosure, loot-box policy — will alter the revenue model I have analyzed.

But one thing I know for certain. However the banner schedule changes, whether the characters are confirmed or not, the fundamental structure of the gacha model will remain. Guarantee threshold, variance, scarcity, and concentration of information power — these are the constants of this system.

The rhythm-keeper never stands at the center of the field. In this case, the rhythm-keeper stands at the edge of an ecosystem with no field. And perhaps, for that very reason, observation becomes harder — but also more important.

We watch the match, but we live in the silence between matches. And in the silence between banner phases, a revenue machine operates quietly, waiting for its next beat.

The question I leave you, the reader, is not whether to pull or not to pull. The question is: when you read an article about a game you love, are you reading an analysis, or merely a schedule presented as if it were fact? And if the latter, who benefits from that confusion?

I have no answer for you. I only have my notes. But sometimes a good question is worth more than a ready answer. And if you have read this far, perhaps you are also keeping your own rhythm — the rhythm of someone who does not want to be led by numbers designed to make you pull one more time.

On August 13, 2026, in Hamburg, I closed my notes and turned off the screen. Tomorrow, I will return to the training ground, listen to the patter of boots on grass, and try to remember that the true rhythm of any system — whether a football team or a revenue machine — always lies in the smallest details, the ones we can only hear when we know what we are listening for.