Measures of consistency
- Sharpe ratio: average return divided by the variability of returns. It rewards steady results and penalises large swings in either direction.
- Sortino ratio: like Sharpe, but it only counts downside variability, so a strategy is not penalised for unusually good days.
- Win rate: the share of trades that made money. On its own it says little, because a strategy can win often with small gains and lose rarely with large ones.
Measures of payoff
- Profit factor: total gross profit divided by total gross loss. Above one means winners outweighed losers before anything else.
- Expectancy: the average amount made per trade, taking both win rate and size of wins and losses into account.
Measures of pain
- Maximum drawdown: the largest peak-to-trough fall in the equity curve. It tells you the worst stretch you would have had to sit through.
- Ulcer index: a drawdown measure that accounts for both how deep and how long the declines were, so a long shallow decline scores worse than a brief dip.
- Recovery factor: total profit divided by maximum drawdown. It shows how much return was earned for the worst pain endured.
How to read them together
Start with the number of trades and the trade ledger, because every ratio is meaningless on too few trades. Then read pain before gain: if the maximum drawdown is more than you could actually tolerate, the rest does not matter. Then check consistency and payoff.
Be careful with any single metric used as a target. Optimising for Sharpe alone can select strategies that sell tail risk, collecting small steady gains until a rare large loss. Ranking by a blend, or by drawdown when survival matters, is more robust.
What none of them can tell you
Every metric describes the past window it was computed on. None of them says anything about the future, and all of them are inflated by unrealistic fills, ignored costs, or settings tuned on the same data. Check that the report separates gross results, charges and the result after charges, and always note the dates.
Common questions
What is a good Sharpe ratio for a backtest?
There is no universal number, and a high Sharpe from a short or over-tuned backtest means little. Compare strategies under identical assumptions and costs, and check that the result holds on data the settings were not tuned on.
What is the difference between Sharpe and Sortino?
Sharpe penalises all variability in returns, upward and downward. Sortino counts only downside variability, so large gains do not lower the score.
What does maximum drawdown tell me?
It is the largest fall from a peak to a following low in the equity curve, showing the worst stretch the strategy would have put you through. It is often the most useful number for judging whether you could stick with a strategy.
Try it on the desk
Build the strategy without code, run it over history, and read the result before any money is involved.