How to Configure and Evaluate Quantitative Trading Backtests
Summary
This guide treats backtesting as replaying historical market data while applying a strategy’s trading rules. It walks through configuring a test, including its date range, bar interval, data quality, fees, starting capital, execution mode, order size, contract settings, and slippage. A trend and oscillation strategy illustrates how signals can be tested, while the guide describes reviewing trade logs, account curves, returns, drawdown, volatility, and the Sharpe ratio. It emphasizes that simulated data may be faster, while more realistic data can better represent trading conditions.
The article cautions that strong results can reflect look-ahead errors, unrealistic fills, overfitting, or omitted slippage, and recommends checking logic, testing out of sample, and using simulated live trading. It suggests spanning different market regimes and having a substantial number of trades, but historical performance cannot establish future results. Performance measures also have limits: volatility alone does not distinguish upward gains from stagnant or harmful movement, and the Sharpe ratio is only one view of risk and return.
Key ideas
- Backtests should represent historical data, fees, order sizing, execution timing, and slippage as realistically as possible.
- Review trade logs and account performance measures alongside the equity curve.
- A high Sharpe ratio or smooth curve can result from look-ahead, overfitting, or unrealistic execution assumptions.
- Out-of-sample data and simulated live trading can help investigate suspiciously strong results.
- A historical test across varied market regimes does not guarantee future strategy performance.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.