Survivorship Bias
IntermediateSurvivorship bias is the error of focusing only on people, companies or investments that have survived or succeeded, while overlooking those that failed or disappeared, which leads to overly optimistic conclusions.
When failures are removed from a data set, the remaining group looks better than the full group really was.
In investing, survivorship bias is a common problem in fund performance data. Poorly performing funds are often closed or merged into other funds, and their records may disappear from databases. If you look only at funds that exist today, average past returns will look higher than the returns investors actually achieved, because the losers have been removed.
For example, suppose 100 funds started ten years ago and 30 have since closed after poor results. Studying only the 70 survivors might suggest the average fund beat the market, when including the closed funds would show that most did not.
Survivorship bias also affects other areas: studies of successful entrepreneurs that ignore failed businesses, analyses of stock indices that reflect only today's members or advice based only on people who succeeded with a particular strategy.
To avoid it, researchers try to use complete data sets that include all members of the original group, both those that survived and those that did not.
For worked examples and full lessons, visit notesbysara.org.