Win rate and profit factor describe trade-level results. They are popular because they feel intuitive, but they can mislead if you ignore position size and tail losses.
Win rate
win rate = winning trades / total trades
A 60% win rate means six in ten closed trades were profitable before costs (definition depends on how you count partials and open trades).
Style expectations
| Style (broad) | Typical win rate intuition |
|---|---|
| Mean reversion | Often higher win rate, smaller average win |
| Trend following | Often lower win rate, larger average win when right |
| Carry / premium | Can look high until a rare loss |
There is no "good" win rate in isolation.
Profit factor
profit factor = sum of winning trade profits / |sum of losing trade losses|
Example: winners total $50k, losers total $25k.
profit factor = 2.0
Values above 1 mean gross profits exceed gross losses. Below 1 means the strategy lost money on a trade-sum basis before some costs.
The core trap
You can have high win rate and lose money if losers are much larger than winners (negative skew). Trend systems often live there by design: many small stops, occasional large gains.
You can have low win rate and strong equity if winners pay for many scratches.
Always read average win vs average loss and the distribution, not only percentages.
Relationship to portfolio metrics
Trade stats do not map cleanly to Sharpe or drawdown:
- Few large trades can dominate the equity curve.
- Rebalancing systems may blur what counts as one "trade."
Use win rate and profit factor for execution and signal diagnostics, not as substitutes for portfolio-level risk.
Costs matter
Slippage and commissions hit high-turnover, high-win-rate strategies hard. Recompute profit factor net of realistic costs.
Limits
- Sensitive to outliers (one huge winner inflates profit factor).
- Look-ahead in backtests can fabricate wins.
- Sample size: 30 trades is anecdote; hundreds matter more.
Win rate and profit factor are worth tracking. Pair them with expectancy (average profit per trade), drawdown, and a clear theory of why your edge exists.
