Survivorship bias

Term explained · 2 min read · Last reviewed on Nov 9, 2026

Survivorship bias arises when a study only looks at the cases that are still around today.

What does survivorship bias mean?

In stock market terms this means: anyone who tests a rule only with today’s members of an index leaves out all companies that fell out of the index earlier, were taken over or went bust. It is precisely these companies that often had weak prices. The test therefore looks better than would have been possible in the past. The same applies to funds, strategies or success stories of which only the successful ones are known.

Example with model figures

A model index has 100 members ten years ago. 20 of them are no longer included today, many after heavy price losses. A test with today’s 100 members instead contains 20 companies that were only added later – mostly after strong rises. A rule that back then would have had to check all 100 members of that time therefore looks friendlier in the test.

How to read it

The bias can only be avoided if you know the members of that time for every day and also have the prices of companies that disappeared. Where that is missing, honesty helps: label the result as too favourable and check it with probes, for example without the biggest winners. In concentrated portfolios with few stocks the bias is especially strong. You meet it in everyday life too: anyone who only hears about successful traders underestimates how many gave up with the same ideas. The question “who is missing from this selection?” guards against many false conclusions, also outside the stock market.

Common mistakes

A common mistake is to take success stories as proof of a method without asking how many failed with the same method. Second, tests with today’s index members are described as “proven over 20 years”. Third, people overlook that data itself has to survive: companies without a complete price history silently drop out of every analysis.

How to use it on TickerEagle

Our methodology page states openly which stocks we test with and what limits that has. The backtesting lesson shows the bias with a simple example.

Related terms

↑ ↓ select · Enter open · Esc close