Fooled by Randomness
A streak of success can arise from randomness when enough attempts run in parallel. Before calling the survivor skilled, reconstruct the losing field and ask what chance alone would produce.
Give 32 monkeys a yearly 50/50 stock pick, and after five years one monkey will have chosen winners every time. Its flawless record looks extraordinary—yet the experiment needs no investing ability at all.
E1The crowd hidden behind the streak
Randomness becomes convincing when selection hides its failures. Each round removes the monkeys that guessed wrongly, while attention follows the shrinking set that remains. After five rounds, the lone survivor has a coherent-looking history, but that history was selected from 32 parallel attempts. The relevant question is therefore not, “How unlikely is this streak for one monkey?” It is, “How likely was some monkey to produce it given all the chances available?” This is the baseline correction captured by Fooled by Randomness #2.
E1Where it shows up
The five-year stock-picking champion
The surviving monkey’s record demonstrates how repeated filtering can turn ordinary coin-flip luck into apparent expertise. Looking only at the winner erases the 31 other starting attempts that made a perfect streak possible.
E1Chance is not the same as no skill
A large field can manufacture an impressive winner, but that does not prove every winner is merely lucky. The model tells you that a streak alone is insufficient evidence; separating skill from chance requires a baseline that includes the number of attempts and the failures you cannot see.
E1Rebuild the missing field
Before trusting a standout record tomorrow, write down three numbers: how many contenders started, how many opportunities each had to succeed, and how many failures disappeared from view. Then compare the winner’s streak with what that full field could plausibly generate by chance.
E1Episodes that teach this
-
Why We Only Hear About The Winners? Survivorship Bias Explained | FutureIQ
8,091 views
Starting with 32 monkeys picking stocks, random 50/50 outcomes leave one monkey that picked winners five years in a row.