Survivorship Bias

Survivorship bias

The tendency to focus only on the successful

Survivorship bias refers to the tendency to focus only on the successful outcomes or survivors in a particular study or analysis, while neglecting the failures or non-survivors. In investment and finance, survivorship bias can occur when analyzing historical data or performance of a specific group of assets, such as stocks or mutual funds. If the analysis only considers the assets that have survived or still exist in the present, it can lead to an overestimation of their performance or characteristics. This bias can skew the perception of the overall risk and return profile of the group being analyzed.

Survivorship bias in trading and backtesting

In trading, survivorship bias rarely looks like a mistake. It hides in the data you choose, the strategies you remember and the traders you follow. A few common forms:

• Delisted and dead assets. A stock backtest that only uses companies listed today silently drops every company that went bankrupt, was delisted or was taken over at a low price. The losers vanish from the sample, so historical returns look better than any real investor could have earned. Crypto has the same problem: many coins that once traded actively no longer exist, while the ones at the top of today's rankings are survivors by definition.
• Today's index constituents. Testing a strategy on the current members of an index as if they had been members for the whole period quietly assumes you knew in advance which companies would grow enough to be included. A fair test uses the constituents as they were on each historical date.
• Only-surviving strategies. If you try fifty variations of an idea and keep the one with the best equity curve, you have found a survivor, not necessarily an edge. The other forty-nine are the failures your result depends on.
• Only-surviving traders and funds. Social media shows the accounts that made money, while the ones that blew up stop posting. Fund databases that drop closed funds overstate average performance for the same reason.

How to reduce survivorship bias

• When testing stocks or crypto baskets, use survivorship-bias-free data that includes delisted instruments and point-in-time index membership.
• Define your rules before you look at results, and record every variation you test, not just the winner.
• Check a strategy on an out-of-sample period it was not tuned on, then forward test it before trusting it with real money.
• Judge any track record by asking how many similar attempts failed and where they went.
• Prefer instruments that existed for the whole test period. Major currency pairs help here because a pair like EUR/USD does not get delisted the way a company or a small coin can.

Replaying history candle by candle, without seeing what comes next, also helps: you make decisions the way you would live, and every losing trade stays in your record. You can practise this way in the ForexThrive simulator.