How can you reduce survivorship bias in your technical analysis performance evaluations?

1 Use a comprehensive database

The first step to reduce survivorship bias is to use a comprehensive database that includes both active and inactive stocks, funds, or strategies. This will allow you to capture the full range of outcomes and performance measures, and avoid missing important information or trends. You can use online sources, such as CRSP, Compustat, or FactSet, that provide historical data on delisted or merged securities, or create your own database by tracking and updating the changes in your universe of interest.

2 Adjust for splits and dividends

The second step to reduce survivorship bias is to adjust for splits and dividends that affect the prices and returns of the securities you are analyzing. Splits and dividends can distort the historical performance of a stock, fund, or strategy, and create false signals or patterns in your technical analysis. For example, a stock that splits 2-for-1 will appear to have a 50% drop in price, which may trigger a sell signal, even though the value of the stock has not changed. To avoid this, you need to adjust the prices and returns for splits and dividends, using either adjusted closing prices or total return indexes.

3 Use appropriate benchmarks

The third step to reduce survivorship bias is to use appropriate benchmarks that match the characteristics and objectives of your technical analysis methods. Benchmarks are useful for comparing and evaluating the performance of your technical analysis methods, but they can also introduce survivorship bias if they are not representative or consistent. For example, if you use a market index as a benchmark for your technical analysis strategies, you may be comparing apples to oranges, since the index may have different risk, return, or diversification profiles than your strategies. To avoid this, you need to use benchmarks that have similar features and goals as your technical analysis methods, such as factor-based indexes, peer groups, or custom portfolios.

4 Test for robustness and sensitivity

The fourth step to reduce survivorship bias is to test for robustness and sensitivity of your technical analysis methods. Robustness and sensitivity tests are useful for assessing how your technical analysis methods perform under different scenarios, assumptions, or parameters, and how sensitive they are to changes in the data or the market conditions. For example, you can test how your technical analysis methods perform with different time periods, frequencies, indicators, thresholds, or filters, and how they react to extreme events, outliers, or noise. This will help you identify and eliminate any spurious or unreliable results that may be due to survivorship bias or data mining.

5 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?