Deflated Sharpe ratio (DSR)

A confidence measure for a Sharpe ratio adjusted for multiple testing and non-normal returns.

The deflated Sharpe ratio estimates the confidence that an observed Sharpe ratio exceeds a benchmark that accounts for the number and quality of the strategy trials considered. It also adjusts for the sample length and for skewness and kurtosis in the returns.

DSR helps distinguish evidence of skill from a result that may have emerged because many alternatives were tested and only the best one was reported. It does not repair a biased backtest or replace out-of-sample validation; its inputs must describe the actual research process, including unsuccessful trials.