Win Rate vs Risk/Reward: 2,292 Tracked Trades

Ask a trading room which number matters and you'll hear about win rate. It's the stat in every screenshot, every Telegram pitch, every "87% accuracy" banner ad.

We've been publishing every signal our desk fires — winners, losers, and the ones that quietly expire — since February. That gives us 2,292 completed trades to check the claim against. And our own book says win rate is close to useless on its own.

Here's the number that makes the case. Our strategies win between 19.2% and 49.0% of the time. Every one of them is profitable.

Five strategies, a thirty-point spread, all in the black

Each of these runs on the same instruments, with the same stop discipline, differing mainly in what triggers the entry and how far the target sits.

(A sixth strategy, weekend gap fill, has two completed trades. That's not a sample, so we're leaving it out.)

Read the first and last rows together. Asian range breakouts lose four times out of five and make money. Mean reversion is nearly a coin flip and makes money. If win rate were the thing that pays you, those two couldn't both be true.

What actually separates them is the average winner column, and it moves in exactly the opposite direction to win rate. The 19% strategy pays 5.16R when it lands. The 49% strategy pays 1.30R. Neither is better — they're the same trade-off priced differently. Push your target further out and you win less often for more; pull it in and you win more often for less.

The column that decides everything is the last one. Margin is how far your real win rate sits above the rate you'd need just to break even. Every strategy on our book clears its bar by three to six percentage points. That's the entire profit — a handful of percentage points of daylight. Nudge the average winner down by half an R and most of these rows go red.

Where the money actually comes from

Sort all 2,292 trades by outcome and the picture gets uncomfortable.

167 trades — 7.3% of everything we've fired — produced +1,196R. The other 2,125 trades, added together, lost 754R.

That's the whole business in two lines. Not the entries, not the win rate, not the strategy names. Seven percent of trades pay for the other ninety-three, with change left over. Remove that top bucket and our record goes from +442R to deeply negative.

It reframes what a losing trade even is. A −1R stop-out isn't a mistake to be engineered away; it's the entry fee for still being in the seat when a 3R-plus runner shows up. There's no version of this where you keep the runners and skip the tolls — the same setup produces both, and you can't tell which is which at entry. We've looked.

This is also the arithmetic behind the most expensive habit in retail trading: closing early to feel safe. At our numbers, capping every winner at 1R would have converted a +442R record into a losing one using the exact same trades, the exact same entries, the exact same win rate. Nothing about the strategy changes. Only the exit does, and the exit is the business.

What a 19% win rate actually feels like

We should be honest about the cost of the approach we just recommended. A strategy that wins one trade in five spends most of its life losing. Six, eight, ten losses in a row aren't anomalies at that hit rate — they're Tuesday.

That's survivable at 1% risk per trade and fatal at 10%. The math has nothing to do with the strategy and everything to do with the size you took:

This is why "let your winners run" is only half the instruction. The other half is sizing small enough that the streak between runners doesn't end you first.

Your stop isn't the variable — your exit is

There's a reason we keep pointing at the winner side of the table. Look at how consistent the loss side is:

1,432 of our 1,501 losing trades — 95.4% — closed at a full stop, averaging −0.96R.

That's a stop doing precisely what it was built to do, over and over, with almost no variance. Which means the loss side of the equation is essentially a fixed cost. It's solved. Whatever separates a good month from a bad one is happening entirely on the other side of the ledger.

Most trading education has this backwards. Stop placement gets the attention because it feels like the dangerous decision — it's the one with the visible downside. But if your stop is mechanical and honored, it stops being a variable at all, and every remaining question is about targets, trailing, and patience.

Losers die fast. Winners take their time.

Since exits are where the money lives, here's how long our trades take to resolve:

A losing trade is usually finished inside two hours. A winner takes three and a half times as long.

The gap widens on higher timeframes, which is where our profit is concentrated:

On H1 — the timeframe producing 68% of our total R from 23% of our trades — winners run seven times longer than losers. Three hours in, a failing H1 trade has typically already failed. A working one is barely getting started.

The practical read: time in a trade is information. A break that's going to work tends to start working; one that sits around the entry going nowhere is usually telling you which column it's headed for. That's not a license to bail on anything slow — it's a reason to stop staring at the ones that are still alive and running.

What we'd want you to be skeptical about

A number is only worth citing if you know what's inside it, so here's what's inside these.

This is our customer-facing book only. Every trade counted here was actually delivered to subscribers. We also track signals that never clear our quality filters, and they're excluded — including them would flatter the sample, since as a group they lose money.

Six months is one regime. Our window is 24 February to 25 August 2026. That covers a single stretch of market conditions, and we've watched our own numbers turn inside it — our Asian range strategy ran +44R in June and −20R in August with no change to the rules. Nobody should treat a six-month sample as a law of nature, us included.

Recent performance is worse than the headline. Our trailing-90-day breakout result is −5.9R. The +442R figure is cumulative, and cumulative figures hide bad quarters. This one has a bad quarter in it.

Two strategies have thin samples. Pivot extreme fade has 25 trades and Asian range breakout has 130. Their ratios are directionally interesting and statistically fragile. Treat the 1,442-trade breakout row and the 592-trade mean reversion row as the load-bearing evidence, and the rest as suggestive.

The short version

Win rate is the number that gets advertised because it's the number that sounds like skill. Expectancy is the number that pays.

If you take one thing from 2,292 trades: stop optimizing your entries and go look at where you exit. Our loss side is a fixed cost, solved by a mechanical stop. Our profit lives in 7% of trades that needed room to run. And the difference between a good year and a blown account is a margin of three to six percentage points that lives entirely in the target.

We publish every signal, closed and scored, on our market pages — including the strategies having a bad month. If you want the breakout-specific version of this analysis, with the session and timeframe splits, that's in our breakout failure rate study.