How to Use Expected Value to Decide Which Trades Are Worth Taking
Your Win Rate Cannot Tell You Whether a Trade Is Worth Taking
A high win rate can still hide a weak strategy. A low win rate can still produce a positive result. The missing variable is the size of your average win and average loss.
Expected value, usually shortened to EV, turns those outcomes into one number: the average R result you observed for a defined setup over a sample of trades.
It does not predict the next trade. It tells you whether your recorded process has produced more than it has cost, after the assumptions in your sample are made explicit.
What expected value measures in trading
Expected value is the average outcome of a repeatable setup, expressed here in R-multiples. One R is the amount you risked on that trade. A result of +2R earned twice that initial risk. A result of -1R lost it.
That framing separates the quality of a setup from account size. It also makes a sequence of trades comparable when your dollar risk changes.
EV = (win rate × average winner in R) - (loss rate × average loser in R)
Use the average loss as a positive magnitude in the formula. The subtraction accounts for the fact that it is a loss.
Calculate expected value with R, not dollar P&L
Dollar P&L changes with position size. EV should describe the setup, so it needs a unit that stays comparable from trade to trade.
R-multiples normalize each result by initial risk. That is what lets a $50 loss and a $500 loss describe the same -1R outcome when each followed its own risk plan.
A worked example
Assume you reviewed 40 completed trades from one defined setup. Sixteen were winners and 24 were losers.
- Win rate: 16 ÷ 40 = 0.40
- Loss rate: 24 ÷ 40 = 0.60
- Average winner: +2.625R
- Average loser: 0.83R
EV = (0.40 × 2.625) - (0.60 × 0.83) = +0.552R per trade.
The calculation is only as useful as the sample. It describes these recorded trades and this setup definition. It is not a return forecast.
Put fees inside the sample
Gross R can make a marginal setup look stronger than the account result. If entry and exit fees are part of the trade, they belong in the R result you record.
Fee-adjusted R keeps the EV calculation closer to account reality. The exact effect varies with venue, fee tier, order type, position size, and exit price.
Do not retrofit one blanket fee assumption across every trade unless it matches your actual records. Consistency matters more than a polished spreadsheet.
Define the setup before you count the trades
An EV number becomes vague when it mixes different decisions. A breakout taken during a trend, a mean-reversion entry, and a discretionary scalp do not automatically belong in one sample.
Start with a label you can apply the same way every time. Then record the context that could change the result.
- Setup or entry pattern
- Market condition or timeframe
- Exit rule and whether it changed during the trade
- Net R after relevant fees
Treat sample size as uncertainty, not a finish line
A small run of trades can be useful for finding questions. It is weak evidence for declaring that a setup works. A few outlier winners or one unusual loss can move the average sharply.
Keep adding comparable observations. When you review the number, look beside it at trade count, the date range, and whether one result dominates the average.
A positive EV is evidence from a sample and a rule set. It is not a promise about the next trade.
Use expected value as a pre-trade guardrail
EV is most useful before entry when it stops you from treating a familiar chart as evidence. It should confirm that the trade resembles a setup you have actually measured.
- Name the setup and its exit rule before the trade.
- Check whether you have comparable, fee-aware results for that setup.
- If the data is thin or mixed, call the evidence unknown instead of inventing confidence.
- After the trade, log the actual result and any deviation from the plan.
Review EV for drift, not drama
A single EV number can hide a changing process. Compare a recent window with a longer history, then inspect the underlying trades before explaining the difference.
The useful questions are operational. Did the setup change? Did fees, exits, or position management change? Did you mix a different market condition into the tag?
That review is more useful than chasing a threshold because it tells you what belongs in the next iteration of your rules and records.
Common expected value mistakes
Using one number for unlike trades
Mixed setup definitions create an average that describes nothing you can repeat.
Using planned R instead of realized R
EV should use the result that reached the account, including relevant costs and changes made during the trade.
Treating EV as a trade signal
EV describes an average across a group. It cannot tell you what one next price move will do.
Expected value FAQs
What is a good expected value for trading?
There is no universal target. The useful question is whether a clearly defined setup has a positive fee-adjusted average across a sample large enough to challenge your assumptions.
Can a low win rate have positive expected value?
Yes. A lower win rate can be offset by larger average winners, just as a high win rate can be offset by losses that are larger than the wins.
How often should I recalculate expected value?
Use a cadence that gives you enough new comparable trades to review. Recalculate after meaningful additions to the sample and when the setup, market conditions, or execution process changes.
The bottom line
Expected value makes your trading record answer a harder question than win rate: after you include the average win, average loss, and fees, what did this setup actually produce?
Use R-multiples, define the setup, keep the sample honest, and review the trades beneath the average. That is how EV becomes a measurement tool instead of another number to rationalize.
Track the inputs behind your EV
RiskReward Pro supports risk-first trade planning, fee-aware R:R, trade lifecycle tracking, and performance review by tag. Its performance analytics include win rate, net R, trade count, and EV by tag.
Use those records to review your own setup data. The point is not to predict trades. It is to make the evidence behind each decision easier to inspect.