Every pity system, in one table

“Pity” is one word doing at least four different jobs. In one game it is a ramp that quietly raises your odds from pull 74. In another it is a counter that hands you the unit outright at 200 and does nothing whatsoever before then. In a third it guarantees a rarity and then draws at random from twenty-six units, which is not the same promise at all. And in two of the games here it does not exist at all.

Every game explains its own system in its own words, on its own terms, and never next to anyone else’s. Here they are next to each other.

The table

Soft pity from is the pull where the per-pull rate starts climbing, listed only where the climb is separate from the cap. Hard pity is the pull that forces the top rarity. Spark is a separate counter that buys the unit outright regardless of what you have pulled. Guaranteed by is the pull at which the game finally owes you the specific unit you are chasing — which, as the column shows, is a much rarer promise than the other three. Rows are in the order of what a featured unit costs, and the table is regenerated from each game’s own Pull Planner model on every build.

Three things the column layout makes obvious

1. Hard pity and a guarantee are different promises

Look at Arknights and CookieRun: Kingdom. Both have a hard pity — 99 and 100 — and neither has a value in the Guaranteed by column. That is not an omission. Hitting hard pity in either game gives you a top-rarity unit, chosen from a pool. The unit you were actually pulling for is one member of that pool, and the game never owes it to you.

A game can have pity and still never guarantee you the unit. Two of the 31 here (Arknights and Chaos Zero Nightmare) have a pity mechanic that promises a rarity and nothing more specific — and two more have no pity at all, which lands them in the same column for a completely different reason.

The spark games make the opposite promise and make it plainly: Fate/Grand Order’s 330 and Blue Archive’s 200 are counters that end in the unit itself, whatever the pulls did on the way. A spark set high enough stops being a promise and becomes a formality — Seven Knights Idle Adventure’s sits at 1,000, past where almost every player has long since landed the unit by chance.

2. The advertised rate tells you almost nothing about the ramp

Several games here advertise the same 0.6% and they do not behave alike. Genshin Impact, Duet Night Abyss, Honkai: Star Rail and Zenless Zone Zero start climbing at 74 and cap at 90. Persona5: The Phantom X starts at 70 and caps at 80 — the same headline number, ten pulls cheaper at the ceiling. CookieRun: Crumble advertises 1.25% as well and never climbs at all, which is why the identical rate costs it 110 typical pulls against Star Rail’s 90. What you experience is set by the ramp, not the rate; soft pity explained walks through what the climb actually does to a pull.

3. “No pity” is a real design, not an oversight

Dragon Ball Z Dokkan Battle and Rise of Kingdoms have no pity on the pull. Every summon is independent, which means there is no number of pulls after which anything improves. That is exactly why they sit at the expensive end of what a featured unit costs, and why their unlucky cases run so much further than everyone else’s — see the unlucky tail. Rise of Kingdoms shows where that design ends up when the per-unit chance is tiny: 2,488 pulls on average, with nothing at any count to stop it.

What this table does not tell you

It compares mechanics, not value. A 200-pull spark is only generous relative to what a pull costs and how often the unit you want is on a banner — neither of which is in this table. It also describes the banner type each game leads with; several games run additional pools with their own rules. And a pity counter that carries between banners is worth considerably more than one that resets, a distinction the columns above deliberately do not try to compress into a number.