If you measure trades in dollars, a $500 win on a $50,000 account and a $500 win on a $5,000 account look identical — but they represent wildly different risk decisions. R-multiples fix this by normalizing every trade to a multiple of what was risked.
Defining 1R
1R = Amount Risked on the Trade (Entry − Stop, converted to $)If you risk $200 on a trade and it closes for a $600 profit, that's a +3R trade. If it stops out, that's a -1R trade by definition (assuming you didn't move the stop). A trade closed early for a small loss might be -0.4R.
Expectancy Formula
Expectancy (R) = (Win% × Avg Win R) − (Loss% × Avg Loss R)Expectancy tells you the average R you should expect per trade, over a large enough sample, regardless of any single outcome. A positive expectancy system makes money over time even with a win rate below 50%.
Worked Example
A system wins 40% of trades averaging +2.5R, and loses 60% of trades averaging -1R. Expectancy = (0.40 × 2.5) − (0.60 × 1) = 1.0 − 0.6 = +0.4R per trade. Over 100 trades risking 1% each, that's roughly +40R of theoretical gain, compounding.
| Win Rate | Avg Win | Avg Loss | Expectancy |
|---|---|---|---|
| 30% | 3.5R | 1R | +0.35R |
| 50% | 1.5R | 1R | +0.25R |
| 60% | 1R | 1R | +0.20R |
| 70% | 0.5R | 1R | +0.05R |
| 40% | 0.8R | 1R | -0.28R |
A 70% win rate system with 0.5R average wins and 1R average losses has an expectancy of only +0.05R — barely profitable after costs. A 30% win rate trend-following system with 3.5R average wins can have far higher expectancy. Judge a system by expectancy, not win rate.
Sample Size Matters
Expectancy is a statistical average — it only becomes meaningful over 30-100+ trades. Judging a system after 5-10 trades is judging noise, not edge. Keep a trade journal recording R for every trade to calculate real expectancy.
Using R to Plan Targets
- ▸A 1:2 risk-reward setup needs only a 34% win rate to break even (before costs).
- ▸A 1:1 setup needs above 50% win rate just to cover spread/commission.
- ▸Track expectancy separately per strategy — a 'mixed' journal hides which setups actually work.