The gambler's fallacy is the mistaken belief that a run of one outcome in a truly random process makes the opposite outcome 'due' on the next try.
No individual coined it. The error was recognised in probability writing long before it had a fixed name — Laplace discusses it — and 'gambler's fallacy' and 'Monte Carlo fallacy' both settled into use through the 20th century.
// Where it sits
The streak feels like debt. It isn't — nothing is keeping the ledger.
What it really means
On 18 August 1913, a roulette wheel at the Monte Carlo Casino landed on black twenty-six times in succession. The room's response is the reason the date is remembered: as the run lengthened, players piled money onto red, on the reasoning that the correction was overdue and getting more overdue by the spin. It wasn't.
What makes the gambler's fallacy so hard to shake is that it's built on a true fact used backwards. The law of large numbers is real: over enough spins, red and black converge. But the convergence works by dilution, not correction. The twenty-six blacks are never cancelled out by twenty-six compensating reds; they're simply swamped by the next hundred thousand spins. Nothing is keeping a ledger, and the wheel has no memory of what it just did.
The reason this matters beyond casinos is that a great many jobs consist of judging independent cases in a row. Studies of sequential decision-makers find streak-avoidance: after several decisions in one direction, the next comparable case gets nudged the other way — not from evidence but from a sense that the run has gone on long enough. Anyone working through a queue of applications, claims or audits is running the same mental wheel, with the difference that at Monte Carlo it was only money.
Where it comes from
Also called the Monte Carlo fallacy, after the 1913 roulette run that made it famous.
Myths & misconceptions
After a long streak, the opposite outcome really is more likely — that's how averages work.
The law of large numbers works by dilution, not correction. A fair coin does not remember; the next flip is 50/50 whatever preceded it. Long-run balance comes from the streak being swamped by future data, never from an offsetting streak being owed.
It's a gambling problem that doesn't touch serious decisions.
It's documented well outside casinos. Sequential decision-makers reviewing independent cases show streak-avoidance — after several rulings one way, the next comparable case gets pushed the other way. Anyone judging a queue of independent items is exposed.
Compare & contrast
Gambler's fallacy vs the hot hand fallacy
Mirror-image readings of the same streak: one says it must break, the other says it will continue. The crucial asymmetry is that the gambler's fallacy is genuinely a fallacy for independent events like roulette, while the 'hot hand' label is now contested for human performance, where outcomes are not independent.
| // | Gambler's fallacy | Hot hand |
|---|---|---|
| Predicts | The streak breaks | The streak continues |
| Domain | Independent events | Human performance |
| Status | A genuine error | Contested since 2018 |
| Root | Law of small numbers | Law of small numbers |
How it connects
- Hot Hand Fallacythe opposite reading — and the one whose 'fallacy' label has since been challenged.
- Apopheniaboth are pattern-finding in sequences that contain no pattern.
Tell it apart
Questions people ask
- If I've flipped nine heads, isn't tails more likely?
- No — still 50/50. What's unlikely is getting ten heads from a standing start, about 1 in 1,024. But you aren't at a standing start. Nine flips have already happened, and they cost the tenth nothing.
- Are lottery 'due numbers' worthless?
- Statistically, yes. A number undrawn for two years has exactly the same chance as one drawn last week. There is one real consequence of which numbers you pick, and it's social rather than statistical: numbers few other players choose are shared by fewer people, so a jackpot is split fewer ways. That is about popularity, not about being 'due' — a long-undrawn number can still be a crowd favourite. The odds of winning don't move.
- Where does it bite outside gambling?
- Anywhere independent cases are judged in sequence. A recruiter who has passed four candidates in a row may unconsciously stiffen on the fifth because 'they can't all be good' — but the fifth candidate's quality has nothing to do with the previous four. The same pattern has been measured in real decisions: loan officers who are not paid by commission are about 8% less likely to approve a loan if they approved the previous one, and US asylum judges about 5.5% less likely to grant a third case after granting two in a row.