The Metric That Actually Determines If You Make Money

Ask a retail trader how their system is performing and you will usually get one number: the win rate. Eighty percent. Ninety percent. It sounds like proof, and it is the figure that appears in every strategy advertisement, because it is the figure that makes people feel competent.

On its own, it tells you almost nothing.

You can be right ninety percent of the time and still lose money. Here is the arithmetic that never makes it into the advertisement.

The 90% loser

Take ten trades. Nine of them win, each producing $100. One loses, and that single loss costs $1,000.

The win rate is 90%. The account is down $100.

Nine wins at $100 is $900. One loss at $1,000 is negative $1,000. Net result: a losing system that was correct nine times out of ten. Run it for a year and the win rate stays attractive while the equity curve bleeds.

The issue is not how often you win. It is how much you make when you are right relative to how much you lose when you are wrong. Win rate ignores that entirely. It is one variable describing a system that requires at least four.

The number that matters

Professionals optimize for expectancy, the average amount a system produces per trade across a large sample. It combines frequency of winning with the size of wins and losses:

Expectancy = (Win% × Average Win) − (Loss% × Average Loss)

Apply it to the 90% winner:

(0.90 × $100) − (0.10 × $1,000)
= $90 − $100
= −$10 per trade

Negative ten dollars every time you place a trade. The high win rate concealed a losing system.

Now invert the profile. Consider a system that is wrong more often than it is right:

Win rate: 40%
Average win: $300
Average loss: $100

(0.40 × $300) − (0.60 × $100)
= $120 − $60
= +$60 per trade

A 40% win rate producing sixty dollars per trade. The system loses six times out of ten and remains profitable, because the winners are large and the losers are small.

That is the entire mechanism, and it explains why win rate in isolation carries so little information.

Why the payoff ratio changes everything

There is a clean way to see how little the win rate means on its own. For any given payoff ratio there is a breakeven win rate, the point at which a system neither makes nor loses money over time:

Breakeven Win% = 1 / (1 + Reward-to-Risk Ratio)

If your average winner is twice your average loser, a 2:1 ratio, the breakeven win rate is 33%. Anything above that is profitable. You can be wrong two out of every three trades and still finish ahead. At 3:1 the breakeven falls to 25%.

The larger your winners relative to your losers, the less often you need to be right.

This is the mechanical reason behind cutting losers short and letting winners run. It is not a motivational slogan. Small losses and large gains pull the breakeven win rate down to a level that is straightforward to clear.

What to track instead

Win rate is one input of four, and the least informative of the four. To know whether a system makes money you need all of them:

Win rate: how often you win
Average win: what you make when you are right
Average loss: what you lose when you are wrong
Expectancy: the figure that ties the first three together

Track expectancy per trade, multiply by the number of trades you take, and you have a defensible estimate of what the system earns. Track win rate alone and you have a feeling.

The difficulty is that a high win rate is pleasant. Being right feels good, and a system that is right most of the time is easier to hold. But the market does not pay for accuracy. It pays for the size of the edge, and the size of the edge lives in expectancy rather than in how often you get to feel correct.

That is the difference between trading and guessing. Guessing optimizes for the feeling. Trading optimizes for the number that survives when the feeling does not.

One caveat. Expectancy means something only if the inputs are real, and most of the win rates and average returns people quote come from backtests that quietly inflate every one of them. That is the subject of the next piece.