The 90-Pull Guarantee and the 50/50: Reading the Revenue Architecture Behind a Gacha Banner Schedule

I reopened my tracking board at two in the morning, Seoul time, and what kept me at the screen was not the name of any character about to appear, but the price architecture underneath the schedule. A content version of this action role-playing game splits into two phases, each running roughly twenty-one days. Phase two of version 7.0 belongs to two returning characters, Flins and Ineffa. Phase one of version 7.1 opens with two entirely new characters, Vesna and Vodyanitsa, before phase two hands the slot back to Skirk and Escoffier. The number worth re-reading sits elsewhere: ninety pulls. Under this system, a player is guaranteed a five-star character within ninety pulls. The first five-star on an event banner splits the odds evenly between the featured character and a standard one; miss, and the next five-star is guaranteed to be the featured character. One easily missed detail: the pity counter is shared across banners of the same category, so the marginal cost of jumping between banners is far lower than players feel it to be. I follow transfer markets not to catch news, but to catch patterns. The pattern here is uncomfortably familiar. A banner cycle behaves much like a miniature transfer window: there is a deadline, there is a fixed quota, some arrive and some must wait. The difference is that the seller is also the rule-writer, the rule-publisher and the cash collector, a concentration of power that most competitive gaming ecosystems simply do not have. My years of watching transfer windows taught me one thing: when a single entity both prices the asset and publishes the information, the value of that information falls by exactly the degree of its monopoly. The two-phase rhythm, roughly three weeks per phase, produces a repeating spending cycle. Players accumulate premium currency over that window and then face an allocation decision. Allocation pressure peaks in phase one of version 7.1, when two new characters arrive together, while phase two offers only reruns. An important footnote: this is an observation about revenue architecture, not about character strength, because the source supplies not a single line of data on kits or statlines. One easily overlooked detail is that the rerun calendar is not fixed at all. Some characters disappear for more than a year; others return after just a few versions. The Chronicled Wish mechanic exists as a separate lane for older characters, opening another revenue path for names that have exhausted their promotional cycle. This is deliberate scarcity design, and it only works if players cannot predict when a rerun lands. Place the three mechanisms side by side and the price architecture becomes legible. The ninety-pull guarantee is a soft floor: it keeps the outcome feeling within reach. The 50/50 creates variance, and most revenue in this model comes from the player who just missed, not the one who got lucky. Sharing the pity counter across same-category banners lowers friction when a player switches targets, which raises spending frequency rather than the value of any single purchase. Laid end to end, the chain reveals a self-contained revenue engine: the publisher controls character supply, release timing, probability rules and the announcement channel. No third party stands in the middle to verify anything. The value chain of an esports tournament, by contrast, must split revenue across several layers including sponsorship, broadcast rights, prize pools and in-game item revenue sharing, which exposes it to schedule shocks, reputational crises and shocks from outside the industry entirely. The gacha model is nearly immune to calendar shocks, but exposes itself to a different risk: regulatory change. The most notable data in the source is not the schedule at all, but source quality. Of twenty-eight flagged information points, twenty carry no source whatsoever; exactly one cites an official publisher announcement, and three are the writer's own opinion. Names such as Odette, Flins, Ineffa, Vesna and Vodyanitsa, along with version numbers 7.0 and 7.1, cannot be cross-checked against the known game state. The source itself concedes that the exact banner schedule is still awaiting confirmation. The scoreline is a liar; data is the only witness I trust, and here the witness has not been fully named. A crisis is only an uncleaned dataset. Here, the uncleaned dataset is a domain mislabel. The source is filed under esports, yet the entire content concerns banner schedules and probability mechanics in a single-player and co-op action role-playing game. That game has no official professional circuit, no franchised league system, no player transfer market in the competitive sense, and its updates are content drops rather than competitive balance patches. Turning characters into players and banner phases into tournament rounds is a category error, not an analogy. What passes for meta here is also different in kind. In competition, meta is the set of most effective tactics, established by win rates and pick-ban rates. In this model, meta is merely the community's spending priority order. One is measured in match results, the other in pull counts. Applying a single analytical frame to both is self-deception dressed in jargon. The biggest risk to a reader lies in acting on a schedule that cannot be verified. Twenty of twenty-eight data points carry no source. If this were a transfer report submitted to a market desk, it would be sent back within the hour. The signal to track over the coming month is an official announcement for phase one of version 7.1, because a single character name appearing on an official channel is enough to confirm or refute the whole calendar above. The next test is whether the names Odette, Vesna and Vodyanitsa appear in official materials at all; that is the reliability gauge for this source. The remaining variable is regulation on probability transparency and player protection in major markets, since that is the only force capable of breaking the price architecture just described. When new data arrives, I will publish an update and correct myself on this page. Before the first pull is made, the number has already whispered the outcome.