The 4% SSR rate never moves. Not at pull 50, not at pull 100, not at pull 199. This site’s own guide to soft pity and guaranteed pull mechanics makes the point plainly: nothing in the Rate Info or Recruit Info screens suggests probabilities escalate, so there is no formal soft pity in the base rates.
That is unusual, and it means the system works differently from how most players describe it. What Nikke offers instead is cleaner and considerably easier to plan around, once you run the numbers.
The Published Figures, and What They Rest On
Special Recruit sits at 4% SSR overall, with 2% allocated to the featured unit. Pilgrims and Over-Spec rate-ups are 1%. A single pull costs 300 Gems, a multi costs 3,000, and every Special Recruit pull earns one Gold Mileage Ticket. Two hundred tickets exchanges for the featured unit outright.
Everything below is derived from those numbers, which is worth stating explicitly, because the arithmetic is only as good as the disclosed rate. Gacha publishers disclose their rates because platform rules or particular jurisdictions require it. Disclosure and independent verification, though, are two different things, and gacha rates are published on the developer’s word alone.
The casino side of chance-based entertainment is considerably further ahead here, which surprises most players who assume the opposite. Slot and table game payout percentages are certified by independent test houses such as eCOGRA and iTech Labs before a title goes live, and the audited return-to-player figure is published per game. A player at a licensed casino can look up a verified number for the specific game in front of them. A gacha player cannot.
That verification layer travels with the licence rather than with the player’s location, which is why it reaches Utah casino players as well. Utah operates no state gaming regulator of its own, so the casinos serving Utah players are licensed in jurisdictions that do run testing regimes, and they inherit those RTP certification requirements along with the licence. Comparisons of Utah casinos rank operators substantially on that basis, weighing licensing body, audit status and published RTP alongside game selection and payout speed, according to this analysis. A Utah casino player checking a slot’s certified RTP has better information about their odds than a Nikke commander has about a banner.
For what follows, take 4% and 2% at face value. Just know that is an assumption rather than a certified figure.
What “No Soft Pity” Actually Means
Each pull is independent. The system has no memory. Sixty pulls without an SSR does not improve pull sixty-one, which follows a geometric distribution with a constant success probability.
The consequences are worth seeing as numbers. At 4%, the median number of pulls to your first SSR is 17 and the mean is 25. Those differ because the distribution is right-skewed: most runs resolve quickly, and a small share run very long and drag the average up.
Here is the part that gets overlooked. Because there is no hard pity on rarity itself, the tail on “any SSR” is genuinely uncapped. Roughly 1 player in 59 goes 100 pulls without a single SSR. About 1 in 3,514 goes 200 pulls with nothing. Those players are not doing anything wrong and the game is not broken. That is what a flat 4% produces at scale.
A caution while we are here: several third-party pity calculators model Nikke with soft pity beginning around pull 60 and a hard rarity pity at 80. The rate screens state no such thing. If a tool is giving you escalating odds, it is modelling a different game.
The Featured Unit Is a Separate Question
At 2%, the numbers stretch out considerably.
Median pulls to the featured unit is 35. The mean is 50. After 50 pulls you have a 63.6% chance of having her. After 100 pulls, 86.7%. Push all the way to 200 and 98.24% of players have pulled her naturally, which leaves roughly 1.76% arriving at the mileage exchange having never seen her drop.
That 1.76% is the entire population the pity system exists to protect. Everyone else gets there by luck.
What the Cap Is Actually Worth
This is the calculation that reframes the whole thing.
Expected pulls to acquire the featured unit, accounting for the 200-pull cap, is 49.1. At 300 Gems each, that is roughly 14,700 Gems against a worst case of 60,000.
So the expected cost is around a quarter of the maximum cost. The cap barely moves the typical experience, because the typical experience resolves somewhere near pull 35. What the cap does is delete the disaster scenario entirely.
That is insurance, not a discount. It is worth exactly as much as the tail it removes, which is small in probability and large in consequence.
Why This Is a Price Rather Than a Gamble
Put the mechanics together and the framing changes.
Two hundred pulls guarantees the unit. Gold Mileage Tickets do not expire and carry across banners, so accumulated progress is never lost to a banner ending. That combination means the maximum you can ever pay for a featured Nikke is fixed at 60,000 Gems, and the randomness only determines how far below that ceiling you land.
Framed that way, the question stops being “will I get her” and becomes “how much of the 60,000 will I actually spend.” The answer, on average, is about 14,700. The income calculator is more useful against a fixed ceiling than against an open-ended risk, because you are budgeting a known maximum rather than an unknown one.
The wishlist works on the same principle from the other direction, narrowing which SSRs the standard pool can hand you rather than changing how often it hands you one.
How to Budget It
Two rules follow from the arithmetic, and they run opposite to how most people plan.
Budget the ceiling, expect the mean. Set aside 60,000 Gems if a unit genuinely matters to you, then be pleasantly surprised when it costs 15,000. Planning around the average and hoping is how people end up 140 pulls in with nothing left.
And treat mileage as an asset rather than a consolation. Tickets carrying forward means partial progress toward one banner is real progress toward a future one, which makes stopping early a coherent decision rather than a wasted spend.
None of this depends on the rate being generous. It depends on the rate being flat and the ceiling being fixed, which are the two things Level Infinite has been consistent about. A flat rate with a hard exchange is a system you can actually do arithmetic on, which is more than most gacha economies allow.
