Intermediate · Learn

Reading a backtest: what the numbers actually mean

A win rate is just one number — here's how to actually evaluate one like someone who does this for a living.

Win rate alone isn't enough

A "94.8% win rate" sounds impressive on its own, but it's incomplete without context. The two questions that matter most: how many trades is that based on, and over what time period?

Sample size

A win rate from 300+ trades tells you much more than the same win rate from 15 trades. Small sample sizes can look great by chance — a handful of lucky outcomes can produce a misleadingly high percentage that won't hold up over time.

Timeframe

A 5-year backtest has lived through different market conditions — rallies, corrections, high and low volatility regimes. A win rate that holds up across all of that is more meaningful than one calculated only during a calm, trending market.

What "counts" as a win

Always check the definition. Does a win mean the option expired worthless? Or does it include trades closed early for a partial profit? Different definitions can produce different-looking win rates from the same underlying data.

The one thing every backtest shares

No matter how rigorous the methodology, a backtest describes the past. It is not a prediction. Market conditions change, and a strategy that performed well historically can behave differently going forward.

See our methodology page for exactly how we calculate our own numbers.