The reason a fair coin produces unfair-looking runs

A statistics class is split in two. Half toss a real coin two hundred times and record the results; half invent a plausible sequence. The lecturer identifies the invented lists in seconds, without a calculator, by looking for the longest run. The forgers alternated too politely. The coin did not.

What everyone sees

A fair process is expected to look even, so a streak reads as a departure from fairness. Six heads in a row invites explanations: a weighted coin, a practised thrower, a hot hand. The explanation is generated because the pattern feels like it needs one, and the feeling arrives before any calculation of how often such runs occur.

What is actually happening

The clustering illusion is a mismatch between the statistics of randomness and the intuition of it. In two hundred independent fair tosses the expected longest run is around seven or eight, and runs of five or six are near-certain. Gilovich, Vallone and Tversky showed the same misreading in basketball shooting records: sequences consistent with independence were confidently described as streaky by players and fans alike.

Why it stays hidden

The illusion survives because the correct baseline is never computed and rarely available. Nobody knows offhand how long a run should be, so intuition supplies a baseline that is far too smooth, and every real sequence then looks anomalous against it. The narrative reward compounds this: a streak with a cause is a story, while a streak without one is not worth repeating.