The number everyone quotes
Ask a trader how it's going and you'll get a win rate. “I win seven out of ten.” It's the number in every Telegram group, every screenshot, every end-of-month brag. And it's the most comfortable metric in trading, because it measures the thing your ego cares about: being right.
The problem is that being right and making money are different games. Win rate counts how often you win. It says nothing about how much you win when you win, or how much you give back when you don't. Half the equation is missing — and it's the half that decides whether your account grows.
Expectancy: the whole story in one rupee figure
Expectancy is the average result of one trade, in ₹, once wins and losses are both counted:
- Expectancy = (Win% × Avg Win) − (Loss% × Avg Loss)
Run a real profile through it. A NIFTY weekly option buyer wins 70% of trades. Profits get booked fast, so the average win is ₹800. Losers get room to “come back”, so the average loss is ₹2,600. Expectancy = (0.70 × ₹800) − (0.30 × ₹2,600) = ₹560 − ₹780 = −₹220 per trade.
This trader wins seven out of ten, feels sharp most days, and pays ₹220 for every trade taken — before brokerage. At 60 trades a month, that's −₹13,200 a month while the win rate says everything is fine. Profitable-feeling and slowly broke.
Now invert it: a 40% win rate with an average win of ₹2,400 and an average loss of ₹1,000 gives (0.40 × ₹2,400) − (0.60 × ₹1,000) = +₹360 per trade. Losing more often than winning, and building an account. The scoreboard was lying; the arithmetic wasn't.
Why comparing win rates across styles is meaningless
Win rate isn't even a property of the trader — mostly it's a property of the style. An option seller collecting premium wins most weeks by design: NIFTY stays inside a range, theta does the work, and the win rate prints 85%. The cost is the tail — the one trending expiry that hands back two months of credits in an afternoon. An option buyer runs the mirror image: strings of small debits lost to time decay, paid for by the occasional outsized winner, at a 35% win rate.
Both profiles can be positive expectancy. Both can be negative. A seller at 85% with one uncontrolled tail loss can be far worse off than a buyer at 35% who cuts fast. So when someone quotes a win rate without an average win, an average loss, and a largest loss next to it, they've told you nothing — including when that someone is you, reviewing your own month.
Chasing win rate teaches your two worst habits
Worse than measuring the wrong number is optimising for it. A trader who wants the win rate up will discover — without ever deciding to — the two most expensive habits in intraday trading:
- Cutting winners early. The trade is up ₹900 against a plan that said ₹1,500. Booking now protects the W. Do that for a month and your average win shrinks to a number that can't carry your losses.
- Holding losers. Exiting at the stop prints an L on the day. Widening the stop, or averaging down, keeps the loss hypothetical — until it's three times the size it was supposed to be. The win rate survives; the average loss balloons.
Look back at the worked example: ₹800 average wins against ₹2,600 average losses. Nobody chooses those numbers. They're what a 70% win rate costs when it's being protected instead of earned.
The three numbers to check every week
- Expectancy, in ₹ per trade — over your last 25–30 round-trip trades, not the last five. A smaller sample is mood, not measurement.
- Average win vs average loss. This ratio sets the win rate you actually need. At 70% wins, you break even only while the average loss stays under roughly 2.3× the average win — the trader above was at 3.25×, which is the whole story of the −₹220.
- Largest loss vs average loss. If your biggest loss is 6× your average one, your expectancy is one bad expiry away from irrelevant. Tail control is what makes the other two numbers mean anything.
None of this predicts the next trade. It measures the last hundred — which is the only dataset you control.
You can't compute this from a P&L statement
Here's the practical catch: your broker's P&L report can't produce these numbers. It shows daily and net figures, not per-decision results. Expectancy needs every fill paired into round-trip trades first — otherwise four partial exits on one position corrupt your averages four ways. This is exactly what PnL Book does with a screenshot of your order book: fills become round trips, and expectancy, average win vs average loss, and largest loss are computed per week and per setup, ready for your weekly review.
Keep quoting your win rate if you like — everyone does. Just make sure you know your expectancy first. One number is for conversations; the other one is your account.