Genshin Impact's banner machine: a transfer market with no contracts

In the crossover week between Genshin Impact version 7.0 and 7.1, I reopened the tracking sheet I had built over three months. There were no player wages and no transfer fees in it. Only pull cadence, accumulated counts, and empty cells marking players who had drained their Primogems before the new banner went live. Looking at it, I realised: this is not simply an entertainment product. It is a labor market with no contracts, no agents, and no players in the sense we usually understand. An acquaintance asked me when I started following games. I answered exactly as I do about football: I do not follow games, I follow cash flow and the motives behind it. My Excel sheet is full of formulas, but the answer always lies outside the cell. CONTEXT: A MACHINE BUILT TO RUN WITHOUT PAUSE Genshin Impact is published by HoYoverse and runs on a gacha model. Players convert Primogems into wish pulls to obtain characters or weapons through a disclosed probability system. Each version splits into two phases of roughly 21 days, each tied to its own banner set. In phase two of version 7.0, most banners are reruns, re-offering older characters such as Flins and Ineffa. In phase one of version 7.1, two new characters, Vesna and Vodyanitsa, arrive at the same time; phase two returns to reruns. Information on the two new characters comes from the publisher's official announcement. The detailed schedule of the following banners was not yet finalised at the time of writing. If in football the summer window is a fair and winter is a reckoning, then here every 21-day cycle is a miniature market. Two markets per version, continuous, with no dead season. That is why I started recording its rhythm instead of merely reading the news. THE CORE: PITY ARCHITECTURE AS PRICE DESIGN The first thing to read is the pity architecture. On a limited banner, a five-star character is guaranteed within 90 pulls. But the 50/50 mechanic is what makes the story: the first five-star has a 50 percent chance of being the featured character and a 50 percent chance of a standard one. If you lose, the next five-star is guaranteed to be featured. Read as an analyst, this is a price design, not a mere luck mechanic. Ninety is the cost ceiling the publisher commits to. The 50/50 is variance inserted in the middle. The publisher sells players the feeling of an upper limit while keeping a wide enough band of fluctuation that revenue never goes flat. This is what transfer analysts call contract structure: who bears risk, who holds pricing power, and which clause makes people unable to walk away. People ask what I look at before a deal lands. I look at motive, not price. With banners, the motive sits elsewhere: loss aversion, or FOMO. And FOMO here is engineered, not accidental. The second element is the rerun policy with no fixed schedule. Some characters are absent for more than a year; others return within a few versions. On the surface, that is inconvenience. From a cash-flow view, it is a scarcity tool: when you do not know when a character you want will return, the opportunity cost of waiting rises sharply. In football, a player nearing contract expiry creates pressure on the club. Here, a banner nearing closure creates pressure on the player. Same logic, different table. The third element is shared pity across same-category banners. This mechanic lowers the marginal cost when players move from one banner to another within the same group. In other words, it reduces friction so cash does not get stuck at one gate. For the publisher, it smooths revenue between debut and rerun windows. The fourth element, the subtlest of all: Chronicled Wish. It is a separate banner type with its own rules, usually for older characters. Its existence shows the publisher has opened a dedicated revenue lane for deposited assets. Characters no longer rerun often can still be monetised again without disturbing the primary banner cadence. What stands out is that spending pressure peaks not in the rerun phase but in the debut phase. When two new characters, Vesna and Vodyanitsa, arrive together, players must split resources across two gates in the same window. Conversely, when phase two holds only reruns, pressure eases. On the surface, this is a content decision. From a cash-flow view, it is revenue architecture: concentrate pressure at one point, then let players breathe at the next. Assembling the four pieces, I see a machine operating on principles identical to a closed transfer market, except for one thing: there is only one seller, and that seller also writes the rules. HoYoverse is the publisher, the probability discloser, and the collector of money. In football, we wrestle with financial fair play and layers of bookkeeping camouflage. Here there is no camouflage at all. The rules are printed in public, and the one printing the rules is the one collecting the money. The pandemic did not kill the transfer market; it only stripped bare the rules we camouflaged with FFP. In the gacha market, that camouflage never existed. Placed side by side, the two systems share one core trait: both sell expectation. A club buys a player because it believes in his future output; a player pulls a banner because they believe in the character's utility in their current roster. The difference is verifiability. With a player, you have match data, metrics, and injury history. With a character, you have a probability table, but no guarantee about the actual experience. Information risk on the gacha side is far higher. That is why I always keep a separate column in my tracking sheet labelled blind spot. In it, I record what neither side wants to say out loud: the publisher does not want to admit revenue depends on impatience; the player does not want to admit they are buying a feeling, not power. THE CONTRARIAN ANGLE: A SCHEDULE IS NOT VALUE The orthodox story players have heard through this period is simple: save your Primogems for 7.1. It sounds reasonable. But it carries a category error. A schedule is not value. Knowing the banner date in advance does not tell you how strong the character is, how well they fit your roster, or whether they are worth the Primogems you are holding. Based on my own experience tracking banner cadence, I find most coverage of this event stops at the level of schedule. In the analysis I read, as many as 20 of 28 information points carry no source, only one cites an official source, and three are the author's personal opinion. Several character names and version numbers cannot be cross-checked against known game state. Rumour is the only thing in football never flagged offside, but for someone who has worked with Excel sheets and financial rules, missing citations is a professional offence. I do not republish rumours without sourcing, and I do not manufacture false equivalences to fill an analytical template. This is the biggest blind spot of the whole story: readers are being sold when, while they need whether. A schedule piece is only useful when paired with data on a character's strength and kit. Without that, it is a poster, not an analysis. One more counter-intuitive point: though called a market, this is not a labour market. Characters have no career age, no injuries, no re-signing, and cannot negotiate. All valuation activity sits on the buyer's side. The only genuinely volatile thing here is player spending behaviour, not character value. In other words, what is priced is not the character, but the player's impatience. OPEN ENDING If one thing is worth carrying from the gacha Excel sheet to a transfer-analysis desk, it is the habit of reading a deal through opportunity cost rather than list price. A 90-pull threshold does not express a character's value. It expresses the risk level the buyer is willing to bear, and the safety zone the publisher is willing to commit to. A smart player will not ask when the banner opens. They will ask: if I pour my Primogems here, what do I give up, and is what I give up worth it. In transfers, the right person at the wrong time is still wrong. In gacha, that holds too. The next thing worth tracking is not the banner date. It is how long the publisher can keep this pity architecture intact, as regulators in many markets tighten rules on probability transparency and the protection of younger players. A monetisation machine built on the fear of missing out will only truly be tested when the rules force it to be more transparent, not when another new character arrives.