Win Rate, Expectancy, and R-Multiples: The Numbers a Backtest Actually Needs
A high win rate feels like proof a strategy works. It isn't — not on its own. Here's why expectancy, not win rate, is the number that actually tells you whether a backtested setup is worth trading.
Once you've logged enough occurrences to calculate a real win rate, it's tempting to treat that number as the verdict: high win rate means good strategy, low win rate means bad one. That instinct is wrong often enough to matter. A strategy that wins 70% of the time can still lose money. A strategy that wins 35% of the time can be genuinely profitable. Win rate alone doesn't tell you which is which.
What does tell you is expectancy — a single number that combines how often you win with how much you win and lose when you do. This guide covers what it is, how to calculate it, and how to read it alongside the other numbers a solid backtest produces.
Why win rate alone is misleading
Win rate only answers one question: out of every occurrence, how many ended in a win? It says nothing about size. A setup that wins small and loses big can have an excellent win rate and still bleed money over time, because the losses that do happen outweigh the frequent small wins.
Setup A looks better on a headline win rate. It's the one that's actually losing money. Setup B, which would feel discouraging to trade day to day — you lose more often than you win — is the one with a genuine edge. This is the entire reason expectancy exists: it corrects for exactly this kind of illusion.
What an R-multiple actually is
Before expectancy makes sense, wins and losses need to be measured in a unit that's comparable across different trades — which is what an R-multiple does. R simply means "your initial risk on that trade." If you risk $100 on a trade and it wins $200, that's a +2R win. If it loses the full $100 you risked, that's −1R.
Measuring in R instead of raw dollars matters because it makes results comparable regardless of position size. A $50 win on a $500 account and a $50 win on a $5,000 account mean completely different things in dollar terms — but if both were 1R, they represent the same result relative to what was actually risked. Once every trade in your backtest is expressed in R, averaging and comparing them becomes meaningful.
Calculating expectancy
Expectancy is the average result you'd expect per trade, in R, over a large enough sample. The formula:
Expectancy = (Win rate × Average win) − (Loss rate × Average loss)
Using Setup B from above: a 35% win rate with an average win of 3R, and a 65% loss rate with an average loss of 1R:
(0.35 × 3R) − (0.65 × 1R) = 1.05R − 0.65R = +0.40R expectancy
That means, on average, this setup returns 0.40R for every trade taken — across enough occurrences for variance to even out. Positive expectancy is the minimum bar for a strategy worth risking real money on. Negative expectancy means the strategy loses money on average no matter how good any individual trade felt.
Reading average win, average loss, max win, and max loss together
Expectancy is the headline number, but the four figures that feed into it are each worth checking individually — they catch different problems.
Average win and average loss
These set your risk-reward profile. A strategy with a small average win and a much larger average loss needs an unusually high win rate just to break even — worth knowing before you assume a modest win rate is a problem.
Max win and max loss
These flag outliers. If your max loss is dramatically larger than your average loss, it's worth checking whether that came from a stop that didn't hold, a gap, or a rule you broke mid-trade — a single outlier can distort an otherwise solid average, especially in a smaller sample.
A note on sample size: expectancy calculated from a handful of trades is close to meaningless — a couple of lucky or unlucky outcomes can flip the sign entirely. Treat expectancy the same way as win rate from a small sample: informative, but not conclusive, until the occurrence count is large enough to trust.
Putting it together
None of these numbers are useful in isolation. Win rate without expectancy hides losing strategies behind a comfortable-looking percentage. Expectancy without knowing your average and max win/loss hides exactly why a number is positive or negative. Read together, they tell you not just whether a setup has an edge, but what shape that edge actually has — which matters for the next step: deciding how to size a position around it.
This is exactly what STRATLOG's PnL stats are built to calculate automatically from your logged rows — expectancy, average win, average loss, max win, and max loss, recalculated the moment you add a new occurrence.