Gambler's Fallacy and Streak Myths in Casino Games

The gambler's fallacy is the belief that a random result is more or less likely because of what has just happened. After seven reds, people feel black must be "due". After three wins in a row, they feel the luck is "running hot". In any game where the rounds are independent, both feelings are wrong, because the game has no memory: each round starts from the same odds. This guide explains why, shows how common streaks actually are with exact numbers, and describes the mistake behind related myths.

What the fallacy is

A common dictionary definition (Colman's Oxford Dictionary of Psychology, quoted by GoodTherapy) is the belief that one event will affect the outcome of a future event when the two are in fact independent. Psychologists link it to a more general habit that Tversky and Kahneman described in their 1971 paper "Belief in the law of small numbers": people expect even a short sample of random events to look like the population it came from, so a run of reds looks "unbalanced" and seems to call for black.

The habit shows up in real behaviour. Clotfelter and Cook's study of a lottery "The 'gambler's fallacy' in lottery play" found that betting on a number dropped sharply right after it was drawn and recovered over several months, a pattern consistent with players believing the number had used up its luck. Nothing in the lottery changed; only the players' beliefs did.

Independent events

Two events are independent when the result of one gives no information about the other. A fair coin, a roulette wheel and a random number generator all behave this way. Cashout Arcade's games draw every round's result fresh from a secure random source (see the RNG guide), with no stored "recent results" that feed into the next draw.

You can test the idea with arithmetic. The chance of red on a single-zero wheel is 18/37 = 48.65%. The chance of ten reds in a row, from a fresh start, is (18/37)^10 = 0.0743%, about 1 in 1,347. Now suppose nine reds have already come up. The chance that the tenth is also red is not 1 in 1,347 any more; it is 48.65% again, because the nine spins that already happened are no longer uncertain. The long odds apply to the whole sequence before it starts. They do not apply to the next step once the earlier steps are fixed.

Streaks are normal

The surprise comes from comparing one specific outcome with how often unlikely-looking things happen overall. A run of 10 reds from a given spin is rare. A run of 10 of something at some point in a long session is not.

Runs of the same colour on a single-zero roulette wheel (red or black; the zero breaks a run). Probability of at least one run that long in the given number of spins, computed exactly:

Runs of the same face with a fair coin (heads or tails), the cleanest case:

So in 100 coin flips, a run of 7 of the same face is more likely than not. If a sequence of 100 flips had no run longer than 4, that would be the suspicious one; real people, asked to write down a "random" sequence, tend to produce exactly that.

A second illustration: the idea that a number is "overdue". A given roulette number has a 36/37 chance of missing each spin. The probability that a specific number has not appeared for 111 spins (three times the 37 you might "expect") is (36/37)^111 = 4.8%. That sounds rare, yet the chance that at least one of the 37 numbers has missed 150 spins in a row is 47%. Some number is always overdue. The one you are watching is no more likely to come next.

The same logic covers Cashout Arcade's games. In Crash, each round has a 51.5% chance of ending before 2x, and over 100 rounds there is a 60% chance of at least one run of 6 such rounds in a row. In Mines with 5 mines, you lose on your first tap one round in five, and a run of three such first-tap losses in a row has close to even odds (47%) of turning up within 100 rounds.

What the law of large numbers does, and does not, do

People hear "it evens out in the long run" and conclude that a streak must be balanced by an opposite streak. That is not what the law of large numbers says. It says the proportion of reds settles toward 18/37 as spins accumulate. It does so by dilution, not compensation.

Suppose the first ten spins are all red. Over the next 1,000 spins the expected number of reds is 486.5, same as ever. The ten early reds are not paid back; they just count for less and less in the total. After 1,010 spins you expect about 496.5 reds, which is 49.2%, only a little above 48.65%. The gap shrinks because the denominator grows, not because the wheel corrects.

Hot hand and cold streaks

The reverse mistake is believing a streak will continue: "I'm on a roll", "this machine is hot". In games where results are independent, a run of wins does not predict the next round any more than a run of losses does. The two beliefs look opposite but have the same root: treating independent events as a story with a direction. (Researchers distinguish the hot-hand belief from the gambler's fallacy; in games of pure chance both are mistakes.)

Hot and cold numbers on a keno or roulette board, a slot that "hasn't paid in hours", a Crash round that "must" go high after a series of early crashes, and a Mines board where you "know" a safe pattern are all the same belief in different clothes. See how slot machines work and keno odds explained for the mechanics that make each independent.

Why this matters

Believing a result is due is not harmless. It is one of the beliefs most often connected with playing longer than planned, raising bets after losses and trying to win losses back. The fallacy turns bad luck into a debt that a game owes you, and nothing in a fair game does.

• It leads to bigger bets at the worst moment. "It has to turn" is the thought behind most doubling systems; see why betting systems fail.
• It makes fair games look rigged. If you expect results to alternate evenly, normal runs look like manipulation. Are casino games rigged? shows how to check properly.
• It hides the real cost. The price of a game is the house edge on everything you stake, whatever the pattern of results. See the house edge guide.

If you notice yourself thinking "I have to win this back" or "it's about to turn", treat it as a signal to stop and take a break. The self-check page and the play-safely page have more.

Related myths, one line each