Your Win Rate is Lying to You — Know Your Real Edge

Two numbers, one contradiction

Open any trading journal and win rate is usually the first stat on the dashboard, in the biggest font, often with a green tick if it's above 50%. It's the number traders quote to each other. It's rarely the number that predicts whether an account is growing or shrinking.

That's not a minor caveat — it's the whole problem with using win rate as a scorecard. Win rate answers exactly one question: out of all closed trades, what fraction closed in profit. It says nothing about how much those winners made or how much the losers cost. A strategy can be right most of the time and still bleed capital, and a strategy can be wrong most of the time and still compound steadily. Win rate can't distinguish between those two strategies. Expectancy can.

The formula

Expectancy is the average amount you can expect to make or lose per trade, computed across every trade a strategy produces — winners and losers together. The formula:

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

Three inputs, all measurable from your own trade log:

• Win rate — winning trades divided by total trades
• Loss rate — losing trades divided by total trades (simply 1 minus win rate, assuming no scratch trades)
• Average win / average loss — the mean profit of winning trades and the mean loss of losing trades, in rupees or in R-multiples if you size positions by risk unit

The result is a single number denominated in the same currency as your P&L. Positive expectancy means the strategy makes money on average, per trade, over a large enough sample. Negative expectancy means it loses money on average, no matter how confident any single trade felt.

A worked comparison — two hypothetical traders

The following numbers are a hypothetical illustrative example built to show the mechanics, not a record of real trades or a claim about actual trader performance.

Trader A trades an intraday breakout setup with a 75% win rate. Wins are booked quickly and are small — average win of ₹800. But losses aren't cut with the same discipline; when the setup fails, Trader A holds and hopes, and the occasional loss runs to an average of ₹3,200 — four times the size of a typical win.

Plugging into the formula:

Expectancy = (0.75 × ₹800) − (0.25 × ₹3,200) Expectancy = ₹600 − ₹800 Expectancy = −₹200 per trade

Trader A is right three times out of four and still loses money on average. Run this across 100 trades and the account is down roughly ₹20,000 — despite winning 75 of them. The win-rate number looked like an edge. It wasn't one.

Trader B trades a trend-following setup with a 40% win rate — wrong more often than right. But the exits are asymmetric by design: winners are allowed to run to an average of ₹2,500, and losses are capped tight at an average of ₹700.

Expectancy = (0.40 × ₹2,500) − (0.60 × ₹700) Expectancy = ₹1,000 − ₹420 Expectancy = +₹580 per trade

Trader B is wrong 60% of the time and still comes out ahead by ₹580 on average per trade — roughly +₹58,000 over the same 100 trades where Trader A lost ₹20,000. Same sample size, opposite outcome, and the trader with the "worse" win rate is the one compounding capital.

Nothing about Trader A's setup is fixable by trading it more. Higher frequency on negative expectancy just gets to the loss faster. Nothing about Trader B's setup needs a higher win rate to be worth trading — the size of the winners is already doing the work.

Why win rate feels better than it is

Win rate is psychologically loud in a way expectancy isn't. Every winning trade is an immediate, discrete "I was right" signal — a small hit of validation that arrives dozens of times a month for an active intraday trader. Expectancy only shows up as a trend across a large sample, and reading it requires deliberately pulling numbers together rather than just watching the P&L tick green on a closed position.

This is exactly the asymmetry that makes win rate addictive and expectancy easy to ignore: the feedback for win rate is frequent, immediate, and emotionally satisfying regardless of trade size. The feedback for expectancy is delayed, aggregated, and emotionally flat until you sit down and calculate it. A trader can go an entire month feeling like they're playing well — mostly green trades, mostly "called it right" — while the account value tells a different story. Rule-based behavioral tracking exists precisely to catch this gap between how a trading session felt and what it actually did to the account.

None of this means win rate is meaningless. A very low win rate can be a sign that stops are too tight or entries are poorly timed, and a very high win rate paired with poor R:R is often a sign that winners are being cut short out of impatience while losers are left to run out of denial. Win rate is diagnostic. It's just not the number that tells you whether the strategy is worth trading.

What should actually drive position sizing

Expectancy, not win rate, is the number that should set how much capital goes into a strategy — because expectancy is the only one of the two that's dimensionally about money made or lost, not about frequency of being right.

Any framework for sizing positions off historical performance — fixed fractional risk, a Kelly-style allocation, or a simple "risk 1% of capital per trade" rule — ultimately needs to know whether the underlying edge is positive before deciding how aggressively to lean into it. Sizing up a negative-expectancy strategy, even one with a high win rate, doesn't fix the leak — it just increases the rate at which capital drains through it. Conversely, a positive-expectancy strategy with a modest win rate can absorb a larger risk allocation precisely because the math behind it is sound, even though it will produce longer streaks of losing trades along the way — something a trader anchored to win rate alone will find much harder to sit through.

This is also why loss rate matters as much as win rate in the formula: a 40% win rate implies a 60% loss rate, and any trader running Trader B's setup needs to be psychologically prepared to lose more often than they win, for as long as it takes the expectancy to play out over a large enough sample. Position sizing decisions made without that context tend to get abandoned exactly when the edge is about to show up.

Calculating your own expectancy

You don't need a full trading system to compute this — you need a trade history and three numbers pulled from it:

1. Average win — sum the P&L of every winning trade, divide by the number of winning trades.
2. Average loss — sum the P&L of every losing trade (as a positive number), divide by the number of losing trades.
3. Win rate — number of winning trades divided by total closed trades. Loss rate is 1 minus that.

Then run the formula: (win rate × average win) − (loss rate × average loss). If the result is positive, the strategy has a real edge over the sample measured. If it's negative, no amount of confidence in individual setups changes the underlying math — the strategy loses money on average, and the fix has to come from either raising average win size, lowering average loss size, or a materially different entry/exit rule, not from trading it more.

A few things worth checking before trusting the number: the sample needs to be large enough that a couple of outlier trades aren't swinging the average (a few dozen trades minimum, ideally more), and the period measured should include at least one adverse market condition — an edge computed only over a strong trending month can look better than it behaves once conditions change. Expectancy calculated from trade history is a description of what a strategy has done, not a guarantee of what it will do going forward.