When evaluating a trading strategy, it is routine to discount the Sharpe ratio from a historical backtest. The reason is simple: there is inevitable data mining by both the researcher and by other researchers in the past. Our paper provides a statistical framework that systematically accounts...