Win rate is the most quoted statistic in trading and the least informative. It is the number every signal service advertises, every strategy claims, and every trader asks about first — and on its own it tells you nothing about whether an approach makes money.
Expectancy: the number that matters
Expectancy is the average result of a trade, expressed in R. It combines win rate and reward-to-risk into a single figure that answers the only question that counts: does this approach make money over enough trades?
Expectancy =
(Win% × Average win in R)
− (Loss% × Average loss in R)
A positive number means the approach is profitable over a large enough sample. Negative means it is not, regardless of how good the analysis is or how it feels.
Three systems, compared
| System | Win rate | Avg win | Expectancy |
|---|---|---|---|
| A | 70% | +0.5R | +0.05R |
| B | 50% | +1.5R | +0.25R |
| C | 35% | +3R | +0.40R |
All assume an average loss of −1R.
System A wins seven times out of ten and is barely profitable. System C loses nearly two trades out of three and produces eight times the expectancy.
If someone advertises a 70% win rate without mentioning the reward-to-risk, they have told you nothing. It is compatible with excellent performance and with losing money.
Because it sells. A high win rate is intuitively appealing and requires no explanation, while expectancy requires a paragraph. Any provider quoting a win rate without the accompanying reward-to-risk and average loss is either not measuring properly or has decided the incomplete number is more persuasive. Both are worth knowing about before you act on their signals.
What system C actually feels like
Worth spending a moment on, because the arithmetic and the experience are very different.
A 35% win rate means roughly two losses for every win. It means strings of four, five and six consecutive losses appear regularly. It means most days end negative.
The expectancy says this is the best of the three systems. The experience says you are doing badly.
This gap is where most traders abandon a working approach. They have a genuinely positive-expectancy system and stop using it after a normal losing sequence, because the losing sequence feels like evidence.
Knowing the expectancy in advance, and knowing what streaks are normal at that win rate, is the defence. It converts a run of losses from evidence into weather.
Sample size
The reason most trader self-assessment is unreliable.
Twenty trades tell you almost nothing. At a 45% win rate, twenty trades could easily produce twelve wins or five wins purely by chance.
| Trades | What you can conclude |
|---|---|
| 10 | Nothing |
| 30 | Very little |
| 100 | A rough indication |
| 300+ | Something meaningful |
This cuts both ways. A trader with twenty winners does not have a proven system either — and is at more risk, because a favourable start encourages increasing size.
The practical implication: risk small enough that reaching a few hundred trades is possible. At 1% risk, that is achievable. At 5%, most accounts do not last long enough to find out whether the approach works.
Where expectancy comes from
Four levers, in order of how much they typically matter.
Average loss size. Keeping losses at −1R rather than letting them run to −1.5R or −2R. This is the largest and most controllable lever, and it is pure discipline — stops placed properly and not moved.
Average win size. Targets set where price realistically reaches, and not closing winners early out of anxiety.
Win rate. Better setup selection and better entry timing. Improvable, but slowly, and with a ceiling.
Trade frequency. Not part of expectancy per trade, but it determines total return. Fewer high-quality trades often beat more marginal ones.
Traders overwhelmingly focus on the third. The first is where the available improvement usually is.
How expectancy degrades
A positive-expectancy approach can become negative without the strategy changing at all.
Widening stops. Average loss moves from −1R to −1.6R. At a 40% win rate with +2R wins, expectancy falls from +0.2R to −0.16R. The system is now unprofitable and nothing about the analysis changed.
Closing winners early. Average win falls from +2R to +1.2R. Similar effect.
Taking trades outside the rules. Marginal setups dilute the average.
Revenge trading. Larger positions on worse setups after a loss.
Baseline: 40% win, +2R win, −1R loss
= +0.20R per trade
Stops widened to −1.6R average
= (0.4 × 2) − (0.6 × 1.6)
= 0.8 − 0.96 = −0.16R
Same analysis. Now losing money.
This is the mechanism by which traders with genuinely good analysis lose. Not through bad reads, but through execution drift that shows up only in the aggregate.
Measuring your own
Requires a journal in R, as covered in the previous lesson.
From 100 journal entries:
Wins: 41 → win rate 41%
Average win: +2.1R
Average loss: −1.05R
(0.41 × 2.1) − (0.59 × 1.05)
= 0.861 − 0.620
= +0.24R per trade
Two things to check beyond the headline number.
Is the average loss close to −1R? If it is −1.3R, stops are being moved or slippage is significant. This is the most actionable finding a journal produces.
What is the expectancy on rule-following trades only? Splitting the journal by whether the rules were followed frequently reveals that the strategy is fine and the execution is not.
Applying it to signals
Directly relevant for readers following signal services, including ours.
A provider’s headline win rate is not enough information. What you need to know is the average R per trade, and the honest answer for any provider is that your result will differ from theirs — because you will not take every signal, you will not fill at the same price, and your position sizing is your own.
The practical approach is to journal the signals you actually take, in R, and calculate your own expectancy after fifty or a hundred. That number is real in a way that any advertised figure is not.
Our signals use one entry, one stop, one target and a full close specifically so that every trade produces a single clean R figure. It makes this calculation possible for you to do.
What to do next
The next lesson covers drawdown — what is normal, how to manage a losing period, and the arithmetic of recovery.
Frequently asked questions
What is expectancy in trading?
The average result per trade in R, combining win rate and reward-to-risk. Calculated as win rate times average win minus loss rate times average loss. A positive figure means the approach makes money over a large enough sample.
Is a 70% win rate good?
Not necessarily. A 70% win rate with an average win of +0.5R produces +0.05R per trade — barely profitable. A 35% win rate with +3R wins produces +0.40R. Win rate without reward-to-risk tells you nothing.
How many trades before I know if my strategy works?
At least a hundred for a rough indication, three hundred or more for something meaningful. Twenty trades could easily produce twelve wins or five wins by chance alone at the same underlying win rate.
What is the easiest way to improve expectancy?
Keeping average loss close to −1R. Most traders focus on improving win rate, which is slow and has a ceiling, while letting losses drift to −1.5R through moved stops — which can turn a profitable system negative on its own.
Can a profitable system become unprofitable without changing?
Yes. Widening stops so the average loss moves from −1R to −1.6R can flip a +0.20R system to −0.16R with no change in analysis. This is how traders with good reads still lose money.
How do I evaluate a signal service’s win rate?
Treat the headline number as incomplete without average R per trade. Better still, journal the signals you actually take in R and calculate your own expectancy after fifty to a hundred — your fills and selection differ from theirs.