You can have a 40% win rate and print money, or an 80% win rate and slowly bleed out. The number that settles the argument is expectancy — and expressed in R, it’s the closest thing trading has to a single truth.
From single R-multiples to average R
An R-multiple is one trade’s result measured against what you risked: a full stop-out is −1R, a target that pays double your risk is +2R. That’s a per-trade snapshot. Expectancy zooms out.
Expectancy in R is just the average of every R-multiple you’ve booked — winners, losers, and scratches all thrown into one pot and divided by the number of trades. If your last five trades were +2R, −1R, −1R, +3R, −1R, the sum is +2R across 5 trades, so your expectancy is +0.4R per trade.
That’s it. No account size, no lot count, no dollar figure. Just: on average, what does one unit of risk return? If you’re new to the underlying unit, the R-multiple primer walks through how each trade becomes a clean multiple in the first place.
The expectancy-in-R formula
The running-average method above works, but the classic formula is more useful for planning because it separates the two levers you actually control:
Expectancy = (Win% × Average win in R) − (Loss% × Average loss in R)
Say you win 45% of the time, your average winner is +2.2R, and your average loser is −1.0R:
- Win side: 0.45 × 2.2 = +0.99R
- Loss side: 0.55 × 1.0 = −0.55R
- Expectancy: 0.99 − 0.55 = +0.44R per trade
The formula makes the tradeoff visible. You can lift expectancy by winning more often or by making your winners bigger relative to your losers — the two don’t have to move together. The expectancy calculator runs this instantly, and the broader expectancy explainer covers the dollar version alongside the R version.
What +0.2R per trade means over 100 trades
A small positive expectancy sounds unimpressive until you let it repeat. +0.2R per trade means that, on average, every 100 trades returns roughly +20R.
If your R is $100, that’s about $2,000 of expected profit per hundred trades — before you touch position size or compounding. Crank R up to $250 and the same edge, same win rate, same everything, returns about $5,000 per hundred. The edge didn’t change; the size did.
Two honest caveats:
- “Expected” is not “guaranteed.” +20R is the center of a wide cloud. A cold streak of eight −1R losers is completely normal inside a +0.2R system. Expectancy describes the long-run average, not the next ten trades.
- Expectancy per trade × trades taken = total return. This is why frequency matters. A +0.5R system taken 40 times a year and a +0.2R system taken 300 times a year are not close — the second one wins by a mile.
Why R beats dollar expectancy for comparison
You can compute expectancy in dollars, and plenty of journals do. The problem is that dollar expectancy is polluted by every sizing decision you made, so it can’t compare trades cleanly.
| Question | Dollar expectancy | Expectancy in R |
|---|---|---|
| Compare a scalp vs a swing? | No — different sizes | Yes — same unit |
| Compare last month vs this month after you grew the account? | No — size changed | Yes — R is size-agnostic |
| Compare your forex edge vs your futures edge? | No — different instruments | Yes — both reduce to R |
| Carry the number from a challenge to a funded account? | No — sizing resets | Yes — R survives intact |
R strips out size and instrument so you’re measuring the decision, not the bankroll. That’s why prop traders who juggle multiple accounts and multiple firms track everything in R — it’s the one currency that’s comparable across all of them. A plain spreadsheet can do this math, but it only works if you actually log every trade correctly and never fat-finger a cell. Shibiki’s auto-journal captures each closed trade in R automatically and computes your live expectancy as trades land, so the number is always current instead of always three weeks stale.
Sample size: how many trades before you trust it
The uncomfortable part: expectancy computed from 12 trades is noise. One lucky +5R runner can flip a losing system’s average positive, and one bad losing streak can bury a good one. Small samples lie in both directions.
Rules of thumb worth internalizing:
- Under ~20–30 trades, treat expectancy as a rough sketch, not a verdict.
- Around 100 trades, the number starts to mean something — though a fat right tail (rare big winners) needs more.
- The wider your win/loss spread, the more trades you need before the average settles.
This is exactly why Shibiki wraps its live edge health in a Wilson confidence interval instead of just showing a point number. A raw +0.4R headline hides whether you’re looking at solid evidence or a small-sample mirage; the interval shows you the range your true edge is likely to sit in, and it tightens as the sample grows. Watch the band, not just the middle — an edge you can trust is one whose whole interval has climbed above zero.
Related: Trading expectancy · Expectancy calculator · R-multiple