Backtesting Options Strategies and Measuring Expectancy
Most retail backtests prove that a strategy worked in a specific regime, then get deployed into a different one. Honest testing is mostly about controlling the ways you fool yourself.
Define the rules before you look
Entry, exit, sizing and invalidation written down in advance. Any parameter you adjust after seeing results is a fit, not a finding. Keep a separate out-of-sample period untouched until the rules are frozen.
Model costs honestly
Options backtests die on execution assumptions. Fill at the mid only if you always trade liquid chains; otherwise assume mid plus a tick of slippage per leg, plus commissions per contract. A four-leg condor pays that spread four times on entry and four on exit.
Segment by regime
Report results separately for positive and negative gamma regimes and for high vs low IV rank. A strategy with an overall 1.2 profit factor that is 1.9 in positive gamma and 0.6 in negative gamma is not one strategy — it is a regime filter waiting to be applied.
- Minimum 100 trades per regime bucket before drawing conclusions.
- Report expectancy per trade in dollars and in R multiples.
- Always report max drawdown; it determines whether you could actually trade it.
Forward testing
Paper-trade the frozen rule set for at least a month before committing capital, then start at a quarter size. Live execution reveals frictions no backtest captures.
Key takeaways
- Freeze rules before measuring, keep an out-of-sample window.
- Slippage per leg is the difference between a profitable and unprofitable condor.
- Segment every result by gamma regime and IV rank.
Practice this in the terminal
Test your knowledge
2 questions. Score 80% or higher to count this guide as mastered.
Keep going
Educational content only. Options involve substantial risk and are not suitable for every investor. Nothing here is financial advice.