Cognitive Biases · CB-22
After a long run of one outcome, people feel the opposite outcome is 'due' — even in a process with zero memory, where each event is completely independent of the last.
The mistaken belief that if a random, independent event has occurred more frequently than expected in the recent past, it becomes less likely in the future (or vice versa) — when in fact, for genuinely independent events, each occurrence has exactly the same probability regardless of what came before.
Named for a famous 1913 incident at a Monte Carlo casino, where a roulette wheel landed on black 26 times in a row, and gamblers lost enormous sums betting on red, believing each successive black spin made red more 'due.'
The Mechanism
26 blacks in a row — the wheel has no memory of any of them
The probability of red on spin 27 was identical to the probability of red on spin 1 — a roulette wheel has no memory of its own history, yet gamblers at Monte Carlo lost fortunes betting increasingly heavily on red as the black streak extended, certain each spin made red more overdue.
01 · IT REQUIRES GENUINE INDEPENDENCE TO BE A FALLACY AT ALL
This is the critical, often-missed condition
The gambler's fallacy specifically applies to processes where successive outcomes are truly statistically independent (a fair coin, an unweighted roulette wheel) — in genuinely non-independent processes (drawing cards without replacement, or systems with real momentum or feedback), past outcomes legitimately do change future probabilities, and it is not a fallacy to account for that.
02 · IT REFLECTS A MISTAKEN INTUITION ABOUT WHAT 'RANDOMNESS' SHOULD LOOK LIKE
People expect small samples to look representative of the true long-run distribution
The underlying intuition (sometimes called belief in the 'law of small numbers') is that a truly random process should quickly self-correct to look balanced even in a small sample — real randomness, however, permits long streaks perfectly consistently, and expecting rapid self-correction is itself the error.
03 · IT DRIVES REAL, DOCUMENTED FINANCIAL LOSSES BEYOND THE CASINO
The same reasoning shows up in trading, lending, and forecasting
Research has documented the same fallacious reasoning among loan officers (judging an application more favorably after several rejections in a row) and stock traders (expecting a reversal after a run of gains or losses) — the casino example is the clearest illustration, but the underlying error recurs anywhere people face a sequence of ostensibly independent outcomes.
Where It Fails / Inversion
Where it fails / inversion
The closely related 'hot hand' question shows the opposite error is also possible: assuming a streak indicates real momentum in a process that's actually closer to independent draws can be just as mistaken as assuming a streak makes reversal 'due' — correctly diagnosing whether a given process has real memory/momentum or true independence is the actual skill, not defaulting to either assumption.
How To Use It
Worked example · resisting the urge to 'average down' after a losing streak
An investor who has seen a stock decline several days in a row may feel a rebound is 'due' and increase their position size — but unless there's a genuine, independent reason to believe the asset is now undervalued, the sequence of recent price moves alone provides no valid statistical basis for that belief, and treating a losing streak as inherently self-correcting is the gambler's fallacy applied to markets.
How to use it
Before treating a streak of any kind as making the opposite outcome 'due,' explicitly verify whether the underlying process is actually independent from one event to the next. If it is, the streak provides zero predictive information about what comes next, however long or improbable-feeling the streak has become.
See Also