Survivorship bias is the silent killer of long-horizon backtests: you test your strategy on the current universe of assets, forgetting that the historical universe also included a long tail of names that have since delisted, gone bankrupt, been acquired, or in crypto, been rugged and frozen.
Equity backtests that use only currently-listed S&P 500 tickers ignore the dozens of stocks that fell out of the index over the years. Crypto backtests that use only currently-trading pairs ignore the hundreds of tokens that have died since 2017.
The fix is a 'point-in-time' dataset: at each historical date, you can only see assets that existed and traded at that date, including names that are now gone. Good data vendors price this in; cheap CSV dumps generally do not.
How Noon Barbari uses Survivorship bias
Every concept here is implemented in the platform. Open the relevant docs or tool to see it in action.
Backtest data sources →Related terms
- Backtesting
Look-ahead bias
Using information in a backtest that would not have been available at the time.
- Backtesting
Backtest
Simulating a trading rule on historical data to estimate how it would have performed.
- Backtesting
Out-of-sample
Data the strategy was not allowed to see during parameter selection.
- Backtesting
Overfitting
Fitting a strategy so closely to past data that it captures noise, not signal.