Fooled by randomness, luck vs survivorship bias

Randomness can manufacture impressive winners when enough attempts run in parallel; survivorship bias then hides the failures and makes the surviving streak look like skill. Before trusting success, reconstruct the selection process and compare the winner with what chance alone could produce.

Start with 32 monkeys picking stocks at random. After five rounds, chance alone can leave one apparent market-beating genius—the monkey whose losing peers have quietly vanished from view.

The filter creates the genius

A large field produces many outcomes. Random variation eliminates most participants while allowing a few to accumulate striking streaks. You then encounter the survivor without seeing the original field, mistake a filtered sample for a representative one, and invent a causal story after the result. The streak is real; the inference that skill caused it may not be.

Genuine advantage—“alpha” in investing—requires information and analysis capable of explaining performance, not merely a winning record or a WhatsApp stock tip. The key distinction is between an outcome that survived and a process that could reliably produce it again. This is where Fooled by Randomness #2 helps: ask what success would look like if the claimed skill were absent.

E1 R1

Where it shows up

The untouched bullet holes

Abraham Wald’s returning bombers made the missing aircraft part of the evidence. Damage visible on survivors could distract from the places where a hit prevented a plane from returning at all.

The stock-tip survivor

A tipster’s correct calls prove little unless you can see the full population of predictions, failures, and abandoned forecasters. A credible investing edge instead rests on unusually deep investigation—speaking with company leaders, operational managers, and lenders.

The hospital sample

Claims about Covid reinfections drawn from hospital reports risk describing only the cases severe enough to enter the hospital dataset, not reinfections as a whole.

Luck does not erase skill

Survivorship bias is not proof that every winner is lucky. Sometimes the survivor has a repeatable informational or analytical advantage. The model tells you to demand evidence about the process and the missing comparison group—not to dismiss achievement automatically.

Rebuild the vanished sample

Before acting on a success story tomorrow, write down three numbers: how many attempts began, how many failed or disappeared, and how often chance could produce the advertised streak. If those numbers are unavailable, treat the result as a lead to investigate, not a method to copy.

Episodes that teach this

Why We Only Hear About The Winners? Survivorship Bias Explained | FutureIQ start here 8,091 views