Backtesting

Abstract

When evaluating a trading strategy, it is routine to discount the Sharpe ratio from a historical backtest. The reason is simple according to the authors: there is inevitable data mining by both the researcher and by other researchers in the past. In this article, the authors provide a statistical framework that systematically accounts for these multiple tests. They propose a method to determine the appropriate haircut for any given reported Sharpe ratio. They also provide a profit hurdle that any strategy needs to achieve in order to be deemed “significant.”

TOPICS: Statistical methods, portfolio management/multi-asset allocation