Balancing Win Rate, Payoff Ratio, and Drawdown in Trading Systems
Summary
The article proposes evaluating automated strategies with two linked measures: win rate and the ratio of average winning to average losing trades. It defines expected edge as win probability times average win minus loss probability times average loss, and illustrates the calculation with example profiles for trend following, mean reversion, and scalping. The examples show why a high win rate alone does not establish profitability: small average wins can leave a system vulnerable to occasional losses, while trend systems may accept fewer wins in exchange for larger winners.
It also considers net return and drawdown, using the Calmar ratio as a return-to-drawdown measure, and advocates testing with realistic commissions and slippage. Combining systems or parameter styles may diversify signals and alter their correlations. These are the author’s heuristics and examples, not independently validated universal thresholds. Backtest results, especially for scalping, can be distorted by execution assumptions and bias, so live performance may differ; the article does not provide a systematic comparative study.
Key ideas
- Expected edge combines win probability with average win and loss sizes.
- Trend, mean-reversion, and scalping systems often have different win-rate and payoff profiles.
- Return should be considered alongside drawdown when evaluating a system.
- Commissions, slippage, execution assumptions, and backtest bias can materially affect results.
- Combining systems with different styles or parameters may diversify signals.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.