Insights · Performance Analysis

Why a high win rate can be misleading

Published: 15 September 2026 · by the QuantGovernor team

Win rate — the percentage of trades closed in profit — says nothing about profitability on its own, because it ignores how much is won and lost per trade. A strategy winning 90% of the time loses money if the rare losses are large enough; a strategy winning 40% of the time can be excellent if its wins are large. The number that actually decides is expectancy: the combination of win rate and average win/loss.

The arithmetic, in one paragraph

Expectancy per trade = (win rate × average win) − (loss rate × average loss). Two examples with round numbers: a system winning 90% of trades with average win 10 and average loss 100 expects 0.9×10 − 0.1×100 = −1 per trade — a slow loser with a beautiful win rate. A system winning 40% with average win 30 and average loss 10 expects 0.4×30 − 0.6×10 = +6 per trade — a solid earner that looks mediocre if you only read its win rate. (Illustrative arithmetic, not performance figures.)

How pretty win rates get manufactured

The uncomfortable part: a high win rate is easy to build on purpose. Take profits tiny and quick, and let losing positions run — or average into them — and most trades will close green while risk accumulates in the few that don't. This is the signature pattern of martingale and aggressive grid systems: months of small wins, then one event that returns them all. On a track record it shows up as a superb win rate next to a scary average-loss figure and a deep equity line — exactly the combination we flagged in How to read a Myfxbook page.

What to read instead

• Win rate + average win/loss, always as a pair. Either alone is decoration.
• Profit factor — gross profit over gross loss — summarizes the same trade-off in one number, provided the trade count is meaningful.
• The worst trades — the size of the largest losses tells you what the strategy does when it is wrong, which matters more than what it does when it is right.
• Equity vs balance — where accumulating floating losses hide while the win rate stays pretty.

Where our systems stand

We publish win rate on the strategy pages, but always in context — next to profit factor, drawdown and the full Myfxbook page where the loss distribution is visible. XAUUSD Adaptive Breakout is the cleanest illustration of the philosophy: a defined stop and target on every trade means the average loss is capped by construction, so its win rate means what it appears to mean. Where a component works differently — the declared averaging inside Copy Gold — we say so, and the metrics should be read accordingly.

Related: Maximum drawdown is not enough · Backtest vs live trading. Past performance is not indicative of future results.