Critique a Backtest Report Line by Line

[Fairman]

Critique a Backtest Report Line by Line

A report shows: 1,000 trades, win rate 92 percent, profit factor 4.1, average win 5 pips, average loss 40 pips, optimized over three years of data with seven parameters. It lists "no commissions" and does not mention spreads.

Check the economics. Expectancy in pips: 0.92 times 5 is 4.6. 0.08 times 40 is 3.2. Net: plus 1.4 pips per trade before costs. A 1.5-pip spread turns it into minus 0.1 pips. The strategy has no edge after realistic costs, despite the headline numbers.

Check the profit factor. Gross profit per trade is 4.6 and gross loss is 3.2, a ratio of 1.44, not 4.1. The reported 4.1 does not match the other figures, which means the report is inconsistent, or computed under assumptions not disclosed, such as dropping the largest losses.

Check the structure. A 92 percent win rate with small wins and large losses is the profile of a strategy that sells insurance: steady gains and rare large losses. The largest single loss is not given. The maximum drawdown is not given.

Check the process. Seven parameters optimized over three years of data suggests overfitting. No out-of-sample test is mentioned. The sample of 1,000 trades is large, but the trades may be clustered in a few periods.

Check what is missing: slippage, swap, the data source, the test period, and the equity curve.

Verdict: do not trade it. Ask for the missing figures first.

Exercise: write five questions to send to the author of any backtest you are shown.