Win Rate Is a Trap. Here's What to Measure Instead.

Win rate is the most quoted metric in trading communities and one of the least useful in isolation. It's intuitive — more wins than losses sounds good. But it's incomplete in a way that gets traders into serious trouble.

The math most traders skip

Imagine two strategies:

Strategy A: 70% win rate. Average win $50. Average loss $200. Expectancy per trade: (0.70 × $50) − (0.30 × $200) = $35 − $60 = −$25

Strategy B: 35% win rate. Average win $300. Average loss $100. Expectancy per trade: (0.35 × $300) − (0.65 × $100) = $105 − $65 = +$40

Strategy A wins 70% of the time and loses money. Strategy B wins 35% of the time and makes money. Win rate alone tells you nothing.

What expectancy actually measures

Expectancy is the average amount you expect to make per trade, across a large sample. It's the only number that tells you whether a strategy has a genuine edge.

Formula: (Win Rate × Average Win) − (Loss Rate × Average Loss)

A positive expectancy means that given enough trades, you will make money. A negative expectancy means the opposite — no amount of position sizing or risk management can fix it.

Why this matters for backtesting

When you run a backtest in Tradelybox, you get:

• Net P&L over the test period
• Win rate
• Total number of trades
• Individual trade list with entry price, exit price, and P&L per trade

From the trade list, you can calculate average win and average loss yourself. Net P&L divided by total trades gives you average expectancy per trade across the sample.

If a strategy has a high win rate but the net P&L is negative or flat, look at the trade list. You'll almost always find a small number of large losers dragging down the average.

The sample size problem

Expectancy is only meaningful over a large number of trades. A strategy that fires 8 times in a year cannot be statistically validated. A run of 3 consecutive losers on an 8-trade sample looks catastrophic but is statistically meaningless.

• Expand the date range
• Drop to a lower timeframe
• Loosen your entry criteria and test whether the edge holds with more signals

R-multiples: a cleaner way to think about it

Some traders think in terms of R — the ratio of reward to risk on each trade. A trade where you risk $100 to make $200 is a 2R trade.

• Average R per trade = (Win Rate × Reward R) − (Loss Rate × 1)
• Positive average R means positive expectancy

This removes the dollar amount from the equation and makes it easier to compare strategies across different position sizes and account sizes.

What to do with a strategy that has negative expectancy

Don't trade it. Not in demo, not with small size. Negative expectancy compounds against you.

Instead, isolate the variable causing the problem. Is the stop too tight — getting stopped out before the move happens? Is the take-profit too aggressive — not getting hit before price reverses? Is the entry occurring in the wrong session or market structure context?

This is where the strategy builder pays off. Change one variable, re-run the backtest, compare results. Iterate until you either find a positive-expectancy configuration or confirm the concept doesn't have an edge.