Mechanical Trading Systems: Pairing Trader Psychology with Technical Analysis

MEASURING TRADING SYSTEM PERFORMANCE

The Sharpe ratio continues to be the traditional and standard measure of performance of both managed funds and trading systems. This ratio is defined as the expected return minus the risk-free interest rate (e.g., treasury bills²⁰) divided by the standard deviation of returns. The “expected return” is defined as the average past return of the entire data sampling in question. Standard deviation is a statistical measure of volatility of the entire data history. It measures the degree of dispersion of the individual data points in the history from the mean (or average) of that history. High standard deviation (high volatility) occurs when many of the individual time intervals within the history deviate dramatically from the average past return for the period.

My objective is not to provide a comprehensive exposition of all the shortcomings of the Sharpe ratio, but rather to outline some of the most dangerous flaws in utilizing this ratio to the exclusion of other measures of system performance.²¹ The basic premise of the Sharpe ratio is that the wider dispersal of individual returns from the average past return, the riskier the investment. Although it is true that a wider dispersal of individual returns from the average past return does suggest higher volatility, because the Sharpe ratio makes no distinction between profits and losses in the composition of its measure of volatility, high volatility of returns does not necessarily equate to a riskier ...