Survivorship Bias Simulator: Why Winners Tell Misleading Stories

simulator intermediate ~8 min
Loading simulation...
Survivorship Bias: Survivors appear to earn ~5-8% more annually than reality

With 200 companies over 10 years at an 85% annual survival rate, roughly 40 companies survive. These survivors show dramatically higher average returns than the full population, because companies with poor returns are more likely to fail and exit the dataset. This is exactly how mutual fund performance statistics, business success stories, and historical analyses are distorted by survivorship bias.

Formula

P(survive all years) = survival_rate^years
Survivor return = mean(annualized returns | survived)
True return = mean(annualized returns | all companies)
Bias gap = survivor return - true return

What Is Survivorship Bias?

Survivorship bias is one of the most insidious forms of selection bias. It occurs whenever we draw conclusions from a dataset that only includes entities that 'survived' some process — companies that didn't go bankrupt, funds that weren't closed, people who succeeded — while ignoring the far larger number that failed and became invisible.

The Wald Aircraft Problem

The most famous illustration comes from World War II. The U.S. military examined returning bombers and found bullet holes concentrated in certain areas. The intuitive recommendation was to reinforce those areas. But statistician Abraham Wald recognized the critical flaw: the data only represented planes that survived. The areas without bullet holes were where planes were being hit and not coming back. Wald recommended armoring the undamaged areas — a brilliant inversion that saved lives and became the canonical example of survivorship bias thinking.

How the Simulation Works

This simulator creates a cohort of companies, each starting at a value of 100. Each year, every surviving company earns a random return drawn from a normal distribution, and faces a probability of failure. Companies that fail stop being tracked in 'survivor only' statistics but remain in the full population data. The fan chart shows all trajectories: survivors are highlighted in cyan, while failures fade into darkness — mirroring how failed companies disappear from real-world databases and media coverage.

The Bias Gap

The bottom bar chart reveals the key insight: the average return of survivors is substantially higher than the true average return of all companies. This gap — the survivorship bias — arises because companies with unlucky return sequences are more likely to fail. When you study only the winners, you systematically overestimate how good the typical outcome is. This same logic applies to studying successful entrepreneurs, profitable trading strategies, or star fund managers.

FAQ

What is survivorship bias?

Survivorship bias is a form of selection bias that occurs when we analyze only the entities that passed some selection process (survived, succeeded, remained visible) while ignoring those that did not. This creates a systematically distorted picture because the failures — which may represent the majority — are invisible. The term was popularized by Abraham Wald's WWII aircraft armor analysis.

What is the Abraham Wald aircraft story?

During WWII, statistician Abraham Wald was asked where to add armor to bombers. The military had data showing where returning planes were hit. While others suggested armoring those areas, Wald realized the data only showed survivors — planes hit in other areas never returned. He recommended armoring the areas with NO bullet holes, because those hits were fatal. This is the canonical example of survivorship bias.

How does survivorship bias affect investing?

Mutual fund performance statistics typically include only funds that still exist. Funds with poor returns are often closed or merged, removing their bad performance from the historical record. Studies show survivorship bias inflates reported fund returns by 1-2% per year. Similarly, stock market index returns appear higher because indices drop failing companies and add successful ones.

How can you correct for survivorship bias?

To correct survivorship bias: (1) include data from entities that failed or exited, not just current survivors; (2) use inception cohort analysis — track all entities from their starting point; (3) be skeptical of 'lessons from successful companies/people' without studying equally the failures; (4) look for base rates — how many attempts fail for each success story.

Sources

Embed

<iframe src="https://homo-deus.com/lab/cognitive-biases/survivorship-bias/embed" width="100%" height="400" frameborder="0"></iframe>
View source on GitHub