During World War II, the US military analysed returning bombers to decide where to add armour. Most damage was concentrated on the wings and fuselage. They were about to reinforce those areas.
Then a statistician named Abraham Wald pointed out the flaw: they were only looking at the planes that came back.
The planes hit in the engine and cockpit never returned. Those were precisely the areas that needed reinforcing.
That’s Survivorship Bias — and it distorts how we understand almost everything.
What Is Survivorship Bias?
Survivorship Bias is the logical error of focusing only on the subjects that passed through a selection filter — the ones that “survived” — while ignoring those that didn’t, because the failures are invisible or harder to study.
We see the winners. We don’t see the losers. We draw conclusions from an incomplete sample, and those conclusions are wrong in a predictable direction: we systematically overestimate success rates and underestimate difficulty.
Where It Shows Up
Entrepreneurship. Business books are almost exclusively written by successful founders. The lessons of people who tried the same strategies and went bankrupt are not recorded. When a successful entrepreneur says “I took bold risks and ignored the naysayers,” that may be true — but so did thousands of people who failed. The advice sounds like a formula for success; it’s actually a description of one path through a lottery.
Investing. Investment funds that perform poorly are closed and their records disappear. Only successful funds remain visible in historical data. Studies comparing fund performance are therefore biased toward showing better-than-actual results — because the worst performers are no longer in the dataset.
Celebrities and school dropouts. “Bill Gates dropped out of Harvard and became the richest man in the world.” True. But for every Gates, there are thousands of dropouts who simply struggled. We remember the exception because it’s remarkable. The rule — finishing your degree correlates with better lifetime earnings — is less quotable.
Ancient wisdom. Old sayings and traditional practices that survived centuries tend to sound wise because the foolish ones were forgotten. “Survivorship Bias in aphorisms” is a real phenomenon: we attribute wisdom to age rather than selection.
Buildings and art. Ancient buildings that still stand look remarkably well-built. Of course they do — the poorly-built ones collapsed centuries ago. We look at surviving Roman aqueducts and conclude Romans were extraordinary engineers, without accounting for all the aqueducts that failed.
The Missing Question
The antidote to Survivorship Bias is always the same question: where is the data that isn’t here?
Before drawing a conclusion, ask: who or what didn’t make it into this sample? What would the failures look like if we could study them? How would including them change the pattern?
Wald’s insight about the bombers was exactly this: he asked about the planes that weren’t in the hangar. Once you ask that question, the armour decision reverses completely.
How to Correct for It
- Seek out failure cases deliberately. If you’re studying successful companies, find the equivalent failed ones and compare.
- Be sceptical of datasets that exclude failures. Mutual fund rankings, university rankings, and “top performer” lists are all subject to survivorship effects.
- Ask what the selection process was. Any time you’re looking at a curated set of results, understand what was excluded and why.
Test Yourself
Survivorship Bias appears as a scenario in Mind Traps — a free 40-level quiz covering psychology laws and cognitive biases. The bomber scenario is one of the hardest questions: most players know the bias by name but can’t immediately identify it in context.
Leave a Reply