High-Frequency Trading: A Practical Guide to Algorithmic Strategies and Trading Systems

Chapter 5. Evaluating Performance of High-Frequency Strategies

The field of strategy performance measurement is quite diverse. Many different metrics have been developed over time to illuminate a strategy's performance. This chapter summarizes the most popular approaches for performance measurement and discusses strategy capacity and the length of time required to evaluate a strategy.

BASIC RETURN CHARACTERISTICS

Trading strategies may come in all shapes and sizes, but they share one characteristic that makes comparison across different strategies feasible—return.

The return itself, however, can be measured across a wide array of frequencies: hourly, daily, monthly, quarterly, and annually, among others. Care should be exercised to ensure that all returns used for inter-strategy comparisons are generated at the same frequency.

Returns of individual strategies can be compared using a variety of performance measures. Average annual return is one such metric. An average return value is a simplistic summary of the location of the mean of the return distribution. Higher average returns may be potentially more desirable than lower returns; however, the average return itself says nothing about dispersion of the distribution of returns around its mean, a measure that can be critical for risk-averse investors.

Volatility of returns measures the dispersion of returns around the average return; it is most often computed as the standard deviation of returns. Volatility, or standard deviation, is ...