The problem with true randomness
If a game uses a plain random roll for each event, with a fixed chance and no memory, long streaks are normal. A 25% chance to dodge will fail four times in a row about 32% of the time (0.75⁴, roughly 0.32), and a 10% chance will occasionally go dozens of tries with no result. Players often read those streaks as bugs or as unfair design.
Designers know this, and many games add a layer that makes outcomes feel more even than a plain roll would. These are often called streak breakers. They are not a single standard technique but a family of adjustments, and a game may use one without saying so.
Two common approaches
Shuffle bags
Instead of rolling independently each time, the game fills a virtual bag with outcomes in the desired proportions, shuffles it and draws without replacement. When the bag is empty, it refills.
Illustration: a bag holds 10 items, 1 rare and 9 common. Each refill produces exactly one rare in every 10 draws, in a random position. Over any full bag, the rate is exactly 10%, the same as an independent 10% roll would give on average.
What changes is the worst case. The rarest wait possible is a rare at the start of one bag followed by a rare at the end of the next, which means 18 commons in a row, and never more. With an independent 10% roll, the chance of 18 misses in a row is 0.9¹⁸, about 15%, and longer runs are possible with no upper limit. The principle is the one behind drawing without replacement in card games: each draw changes what is left.
Rising chance after each failure
Here the game starts with a low chance and raises it every time the event fails, then resets after a success.
Illustration: start at 5% and add 5 percentage points after every failure. The chances on attempts 1, 2, 3 and so on are 5%, 10%, 15%, and by attempt 20 the chance is 100%, so a success is guaranteed by then. The average rate over time is higher than 5%, because the rising chance makes early successes more likely than the starting value suggests. Designers using this approach tune the starting value and step so that the long-run average matches the figure they want to present.
Two points about this method:
- Success on the first attempt is still just 5%, so streaks of a few failures still occur.
- Very long streaks are impossible, because the chance climbs steadily toward certainty.
Where these systems appear
The same ideas show up in several areas of game design:
- Combat. Some games adjust critical hit or evade chances after long runs without a result, so that critical hits arrive at a steadier pace than a plain roll would give.
- Loot. Guaranteed-after-N rules and growing chances after repeated misses are widespread, and the article on pity timers covers them in detail.
- Hit chances. Some games are reported to calculate hits in ways that favour the player when the stated chance is high, which is a different kind of adjustment from a streak counter.
- Procedural content. Level generators sometimes avoid placing the same room type twice in a row or guarantee a certain item appears within a set number of rooms.
Which games do this, and exactly how, is often undocumented, and details can change between versions. It is better to say that such systems exist than to assume a particular title uses them.
What changes mathematically
A plain roll has a property called independence: the result of one roll tells you nothing about the next. A streak breaker removes that property. After a long run of misses, the next roll is no longer the same as the first, so the events become dependent events.
That has several effects:
- The belief that a result is “due” is partly right inside such a system, which is the opposite of the standard gambler’s fallacy. The difference is that the system was built to do it.
- The long-run average can stay the same while the spread of waiting times narrows. Fewer extremely lucky or unlucky players exist.
- Calculations that assume independent rolls, such as the 1 − 0.99¹⁰⁰ style of reasoning for drop rates, become approximate.
- A stated percentage may describe the average rate rather than the chance on each individual roll.
Reasons designers do it, and the downsides
The usual motivations are a better player experience and fewer complaints about unfair luck. Evening out outcomes can make a game feel more reliable and reduce the risk that someone quits after a long unlucky stretch.
There are downsides:
- Reduced surprise. Outcomes become easier to predict, which some players find less exciting.
- Exploitability. If players learn the pattern, they may time their actions to take advantage of it, for example by using a certain ability only after a long run of failures.
- Transparency. A hidden system can mislead players about their chances, particularly when real money is involved in paid random rewards. Disclosure of how odds work helps, and published loot box odds are one example of this kind of transparency.
A note on gambling
These adjustments are design choices in games with fixed rules. They should not be read as saying that any gambling device will correct itself after a losing run. Games of chance for money typically have a built-in house edge, results of independent plays do not make later ones more favourable, and losses over time are the normal outcome. Anyone worried about their own play should look for support from a local service.