Drop Rates

You are not cursed. You are — of everybody.

Oscilloscope screen: a long row of small dim blue dots running left to right across a dark grid, and one bright amber dot blazing at the far right end.

Two numbers settle every argument about a grind: the chance, and how many times you have already tried. Put yours in once.

— of hunters at your run count have nothing to show either

—

—runs is a coin flip — half of everyone has it by here

—runs before nine hunters in ten are done

—runs is the number on the wiki

—you have already spent, at the minutes you gave

The curve everybody is on

Run count along the bottom, the share of hunters who have it by then up the side. The line never reaches the top, which is the whole problem: no number of runs makes it certain, it only makes the leftovers rarer. Your position is marked.

Hunters done	Runs	Hours	Added since above

—

The number on the wiki is not the wait

A one-in-a-hundred drop gets written down as “about a hundred runs.” That figure is an average, and an average is dragged upward by the handful of people still going at run nine hundred. The number that splits the room in half is a different, smaller one — and the gap between them is fixed, whatever the rate.

Reading	Runs	Done by then

—

The game is not keeping score

There is no counter filling up behind the scenes, and nothing you have already spent is banked. Every row below asks the same question — what are my odds over the next stretch — from a different amount of history, and every row gives back the same answer to the last decimal. That is not a rounding coincidence; it is what an independent roll means.

Runs behind you	Next stretch	How that compares to a fresh start

—

So how unlucky is unlucky

Going the other way: pick how rare you want to be, and this is the run count that earns it. The last column is how long that is at the minutes you put in, which is usually the figure that ends the argument.

Still empty-handed	Takes	Hours	Rarer than

—

What a pity counter actually buys

A hard pity hands it to you the moment you reach some run count, however the rolls went. It is usually sold as generosity. What it really does is cut the tail off — and the amount it takes off the average turns out to be exactly the share of people who ever touch it, which is a much smaller number than it sounds.

Guaranteed at	Average runs	Reach it	Worst case

—

Every rate on one ladder

The rate is the only thing the grind knows about. Halve it and every answer on this page doubles — the coin flip, the nine-in-ten, the hours, all of it — give or take the rounding up to a whole run. So a drop described as “twice as rare” is not roughly twice the trouble. It is twice, and it stays twice however deep the ladder goes.

Rate	Coin flip	Nine in ten	Hours to the flip

—

Where these numbers come from

One assumption, stated once. Every run is an independent roll at the same chance. That is how ordinary loot tables and ordinary gacha banners work, and it is the assumption behind every figure here. Games that quietly raise your odds as you go, or that draw without replacement from a pool, break it — and when they do, the maker almost always says so, because it sounds good.
The coin-flip run count is not half the wiki number. It is about 69.3% of it, and the exact figure is how many rolls it takes for the misses to multiply down to one half. The fraction is the same whatever the rate, which is why the ladder at the bottom keeps its shape all the way down.
Nothing here is a simulation. Every row is arithmetic on your two numbers. The one place where a closed form is easy to get wrong is the pity table, so that formula was checked against four hundred thousand simulated hunts before it went on the page; the two agreed to the first decimal.
The share of hunters still empty is also your percentile. If the page says 30%, then three in ten people who have put in exactly your run count are in the same position. It does not mean you have a 30% chance of anything from here — your chance from here is the same as everybody’s, which is the point of the third section.
Hours are your own figure, not ours. The minutes-per-run box is whatever you put in it, including travel, resets and loading. A run that takes four minutes on paper and eight in practice doubles every hour on this page, and there is no way for a web page to know which one you meant.
The curve is drawn, not sketched. Each point on it is the exact share done by that run count, and the marked point is your run count. The flat top is real: going from nine in ten to ninety-nine in a hundred costs about as many runs again as everything before it.
Rates deeper than the list are the same arithmetic. The dropdown stops at one in four thousand because that is where the interesting part of the curve leaves the screen, not because anything changes. Read the ladder and keep doubling.