Why win rate is a vanity metric — and what to look at instead

A 90%+ win rate is the loudest number in indicator marketing — and one of the least useful. It tells you how often a strategy wins, and nothing about how much. You can win nine trades out of ten and still bleed money. Here's why win rate flatters, and the two numbers that actually tell you whether an edge exists.

Why win rate flatters

Win rate feels like safety. “92% of trades win” reads as “I'll almost never lose,” and that's exactly why it sells. It's also trivial to engineer: set a tiny take-profit and a huge stop, and most trades tag the target before they wander into the stop. The wins are frequent and small; the losses are rare and enormous. The scoreboard says 92%. The account says otherwise.

Any number that can be inflated without improving the strategy isn't measuring the strategy. It's measuring the marketing.

The math it hides

What win rate leaves out is the size of wins versus losses. Take that 92% win rate: 92 winners at +0.2R and 8 losers at −3R.

• Gross profit: 92 × 0.2R = +18.4R
• Gross loss: 8 × 3R = −24R
• Net: −5.6R — a losing strategy, at a 92% win rate.

Flip it: a strategy that wins just 40% of the time, +2R on winners and −1R on losers, nets +0.2R per trade and compounds. A low win rate can be an excellent strategy; a high one can be a trap. Win rate alone can't tell them apart.

What to look at instead

Two numbers do the job win rate can't:

• Profit factor — gross profit ÷ gross loss. Above 1.0 the strategy makes money; below 1.0 it loses, whatever the win rate. It bakes in both frequency and size, which is the whole point.
• Expectancy — the average result per trade (in R). It answers the only question that matters: if I take the next hundred setups, what do I expect to happen?

Neither is as seductive as “92% win rate,” and that's the tell. Honest metrics are usually the boring ones — because they can't be gamed by shrinking the target and widening the stop.

Reading a track record honestly

When you're handed a performance claim, work past the win rate and ask:

• What's the profit factor, and over how many trades? A great-looking number on 20 trades is noise; it needs a real sample.
• Is it after costs — spread and commission — or a gross figure that quietly ignores them?
• Are the losers shown? A record with only winning screenshots is telling you what it's hiding.
• Was it measured out-of-sample, on data the strategy was never tuned on?

None of this guarantees a profit — no honest number can. But it does separate a strategy that makes money from one that just wins often. When someone leads with the win rate and goes quiet on profit factor, that silence is the answer.