The problem pity timers solve
With a fixed 1 percent chance, a meaningful minority of players will go several hundred attempts without success. About 5 percent need more than 300. For those players the experience can be discouraging, and the worst of it is unbounded: there is no number of attempts after which success is certain.
Pity timers, also called bad luck protection, address this by tracking consecutive misses and changing the outcome once the count gets high enough. They trade a small amount of designer revenue or item scarcity for a limit on how bad things can get.
Hard pity: a guaranteed ceiling
The simplest version is a hard guarantee. The counter increases by one on every miss, and when it reaches a set number the next attempt is certain to succeed. The counter then resets.
Take a 1 percent base chance with a guarantee at attempt 90. Several things follow:
- The worst case is 90 attempts. No one goes longer.
- The chance of reaching the guarantee is the chance of 89 straight misses, 0.99^89, which is about 41 percent.
- The average number of attempts falls from 100 to about 59.5. This comes from summing the chance of still being unsuccessful before each attempt: (1 - 0.99^90) / 0.01, which gives roughly 59.5.
So the protection removes the long tail and lowers the mean by about 40 percent, even though the per-attempt rate has not changed. The base chance still applies, and about 59 percent of players succeed by luck before the guarantee fires.
Soft pity: a rising chance
Some games prefer a gradual approach. Up to a threshold, the chance is the base rate. Beyond it, the chance rises with each further miss, quickly or slowly, until it reaches certainty or a high cap. This feels less like a cliff and more like a ramp.
Soft pity pulls the average down further than a hard cap at the same endpoint would, because many players succeed in the ramp zone before reaching the limit. The exact effect depends on the starting point and the steepness, which is a design choice. A related technique, where the chance rises after every failure from the start, is covered in the article on pseudo random distribution.
What it does to average cost
It is tempting to think “guaranteed” means “cheap”, but be careful with the average. Using the example above:
| System | Average attempts | Worst case |
|---|---|---|
| Fixed 1 percent | 100 | unlimited |
| Hard pity at 90 | about 59.5 | 90 |
The protected average is lower, but a designer who adds pity will often lower the base rate at the same time, so the overall cost to the player may be unchanged or even higher than before. Judging a system by the presence of a pity counter alone, without the base rate, is incomplete. When a published figure appears, check whether it is the base chance or the effective average, as discussed in the article on loot box odds.
Details that change how pity behaves
- Does the counter reset? Some counters reset after any reward in the tier, others only after the specific target item.
- Is it per banner, per account or per pool? A counter that carries over across pools is more valuable than one that resets whenever the pool changes.
- Does the guarantee cover the exact item or just the tier? A guaranteed rare slot may still draw a different rare than the one you wanted.
- Does it persist between sessions? Counters stored on the server generally do; some local systems may not.
These details matter more than the headline number, because they decide how often a long dry spell actually gets protected.
Pity and the gambler fallacy
A pity timer is the one situation where “I have missed many times, so a hit is closer” is true. The reason is that the system is built to remember. In a fixed-rate system with no counter, past misses do not affect the next attempt, and believing otherwise is the mistake explained in the article on the gambler fallacy. The right question is not “have I been unlucky?” but “does this game keep a counter at all?”
The cost side: what is being protected
From a design viewpoint, protection serves two goals. It keeps rare items from feeling out of reach, and it limits the very worst player experience. It also makes spending more predictable. This ties into broader choices about monetisation design, since a visible cap lets players estimate the maximum they might invest before committing.
From the player’s side, a cap does not remove uncertainty. Between the first attempt and the guarantee there is still wide variation, and the worst case may be more than you wish to spend or play. Treat the cap as an upper bound on effort, and the average as a typical figure, with real runs scattered around it.
A simple way to evaluate a pity system
- Find the base rate and the cap (or the point where the chance starts rising).
- Find out whether the counter resets and what it covers.
- Estimate the average and worst-case effort using the base rate and the cap.
- Compare against what the same item would cost without protection, where known.
None of this predicts what will happen on your next run. It only shows the shape of the risk.