Ask a beginner how their trading is going and they will quote a win rate. Ask a professional and they will quote an expectancy in R. The gap between those two answers explains most of the confusion in retail trading — including why systems that are "right" most of the time can steadily lose money, and why systems that are wrong most of the time can steadily make it.
The risk-reward ratio (written R:R) compares what a trade risks against what it could reasonably pay. Both numbers come from the trade's own structure: the distance from entry to stop-loss is the risk; the distance from entry to target is the reward.
Example: you buy an asset at $100, place a stop-loss at $95 and a target at $110. The risk is $5 per unit, the potential reward $10 per unit, so the ratio is 1:2 — every dollar risked stands to earn two. Nothing about the ratio predicts whether the trade wins. It only defines the terms of the bet.
Traders shorten this into the R-multiple: R is simply "the amount risked on the trade". A win at the target above is +2R; a loss at the stop is −1R; an early exit halfway to target is +1R. If you risked $100, +2R means +$200.
Measuring results in R instead of dollars does something quietly powerful: it makes every trade comparable. A crypto trade, a forex trade and a stock trade all reduce to the same unit, regardless of account size or volatility — provided you sized each position so 1R was the same fraction of your account. (That sizing step has its own guide: position sizing.)
Win rate and reward size are not independent dials. Tighten your target and more trades reach it — win rate up, R down. Stretch your target and fewer trades get there — win rate down, R up. Every strategy lives somewhere on this curve, and the only question is whether its combination clears break-even.
The break-even win rate for a given payoff is 1 ÷ (1 + R):
| Average payoff per win | Break-even win rate | What that means |
|---|---|---|
| 0.5R | 66.7% | Must win two of every three trades just to tread water |
| 1R | 50.0% | Coin-flip territory — costs push you below water |
| 1.5R | 40.0% | Can be wrong 6 times in 10 and survive |
| 2R | 33.3% | One win pays for two losses |
| 3R | 25.0% | One win pays for three losses |
Costs — spread, fees, slippage — shift every break-even upward, which is why serious backtesting always includes them.
Expectancy is the average result per trade in R, combining win rate and payoff in one figure:
expectancy = (win rate × average win) − (loss rate × average loss)
Work through a concrete case. A system wins 42% of the time; winners average +2R, losers −1R. Expectancy = (0.42 × 2) − (0.58 × 1) = +0.26R per trade. Losing nearly six trades in ten, it still earns about a quarter of a unit of risk on every signal, on average. Over 200 trades risking 1% each, that compounds meaningfully — while feeling, day to day, like constant losing.
This is not a hypothetical shape. SRA Quant's published methodology reports exactly this profile for its validated historical baseline: a 42.2% win rate with an average expectancy of +0.257R per signal, with winners roughly twice the size of losers. The live track record exists precisely so that claim can be checked against ongoing results rather than taken on faith. And the honest caveats apply in both directions: historical expectancy is measured, not promised, and it says nothing about any individual trade.
A high win rate feels like skill, so people engineer it — usually by taking profits quickly and letting stops drift wider "to give the trade room". The result is a strategy that wins 80% of the time at +0.3R and loses 20% at −2.5R. Expectancy: (0.8 × 0.3) − (0.2 × 2.5) = −0.26R. It produces months of pleasant small wins, then gives everything back in a handful of disasters.
The inverse discomfort is why positive-expectancy trend systems are hard to follow: they feel bad most of the time. Long losing streaks are mathematically guaranteed at modest win rates — which is exactly why the sizing arithmetic matters, and why judging any system on ten trades is meaningless. Expectancy only reveals itself across large samples.
Three habits follow from all of this. First, define the stop and target before entry, so R is knowable — a trade without a defined invalidation point has no measurable risk. Second, filter setups by a minimum ratio: many frameworks refuse anything below roughly 1.5:1, because at low payoffs even good analysis can't clear break-even. SRA Quant's scorer, for example, automatically disqualifies any setup below 1.5:1 regardless of how strong the rest of the confluence looks. Third, review your trading journal in R, not dollars or win percentage — it is the only view that tells you whether the process is working.
SRA Quant applies these principles automatically. Every setup carries a defined invalidation and target, is rejected below a minimum risk-reward, and every outcome is logged in R — publicly.
SRA Quant provides market analysis and educational content only. Nothing on this page or the platform constitutes financial advice, and past or backtested performance does not guarantee future outcomes. Trading involves substantial risk of loss.