Pages

06 January 2017

Survivorship bias.

One of my learning goals for the year is to become more familiar with biases.

Survivorship Bias is a bias of systematically over-estimating our chances of success because when we scan the environment for information on our chances, we see a lot of successes. However, this is because failures often leave no trace. Failures write no books, make no albums, etc. Failures are not usually celebrated. What we see are the survivors, e.g. Google rather than Altavista.

Most business ventures fail. Most musicians never get that record deal. Most authors never get published. Etcetera. The reason is not usually a quality issue. Good business ideas fail. Good authors and musicians are denied. A lot of it comes down to luck or persistence. Most small businesses are under-capitalised for example and are not able to persist long enough to get into profit, which can take five years of trading. A record deal often depends on the right person seeing an act on the right night and taking a personal interest. Or on what kind of pop music happens to be in fashion (a lot of people who make it by tapping into fashion don't last because fashions change).

This not to say that everyone is doomed or that we'll never succeed. It is a warning that in assessing our chances of success we need to look at examples of failures as well as successes. What did people get wrong as well as what they got right. We mostly underestimate the amount of persistence in the face of adversity required for success.

No comments:

Post a Comment