Trading expectancy, calculated as (Win Rate multiplied by Average Win) minus (Loss Rate multiplied by Average Loss), measures the average profit per trade in units of R. A positive expectancy is the minimum condition for a tradeable system: it is the single metric that separates a genuinely profitable strategy from one that merely appears profitable due to a high win rate.
What is trading expectancy?
Definition and why it matters
Expectancy answers one foundational question: if you run this strategy 1,000 times under identical conditions, what is your average gain per trade? This is not a prediction for any single trade; it is the law of large numbers applied to trading system evaluation.
According to the European Securities and Markets Authority (ESMA), between 69% and 89% of retail investor accounts lose money when trading CFDs and leveraged FX products. These traders do not necessarily lose because they are wrong about market direction most of the time. Many lose because their system has a negative expectancy: they cut winners short and let losers run, or they trade high win-rate systems with unfavorable reward-to-risk ratios.
Expectancy makes this dynamic visible. It is the core metric of any rigorous strategy evaluation.
Expectancy vs win rate: which metric matters more
Win rate measures how often you are right. Expectancy measures how much you make on average per trade, accounting for both frequency and magnitude. These are different things, and confusing them is one of the most common mistakes in retail trading.
Concrete example: a strategy with a 72% win rate and a 0.25:1 reward-to-risk ratio has negative expectancy.
Calculation: (0.72 x 1 R) - (0.28 x 4 R) = 0.72 - 1.12 = -0.40 R per trade.
This strategy loses an average of 0.40 R per trade despite winning nearly three-quarters of its trades. Without expectancy, a trader looking only at win rate would believe this system is excellent.
| Metric | What it measures | Key limitation |
|---|---|---|
| Win rate | Proportion of winning trades | Ignores the size of wins and losses |
| Reward-to-risk ratio | Ratio of average win to average loss | Ignores how often you win |
| Expectancy | Average gain per trade (frequency x magnitude) | Requires a minimum sample of 100 trades |
For a deeper look at how win rate and expectancy interact, see our guide on the risk-reward ratio and expectancy relationship.
The expectancy formula step by step
Formula: E = (Win Rate x Avg Win) - (Loss Rate x Avg Loss)
The complete formula is:
Expectancy = (W x G) - (L x P)
Where:
- W = win rate (proportion of winning trades, between 0 and 1)
- G = average win per trade in R
- L = loss rate = 1 - W
- P = average loss per trade in R (absolute value)
The result is expressed in R per trade. Using R as the unit normalizes expectancy across different account sizes: a system with +0.45 R expectancy performs identically whether you risk $50 or $500 per trade.
Example calculation with real numbers
Take a Forex swing strategy backtested over 150 trades:
- 63 winning trades, 87 losing trades
- Average winning trade: 2.4 R
- Average losing trade: 1.0 R
Calculation:
- W = 63/150 = 0.42
- G = 2.4 R
- L = 87/150 = 0.58
- P = 1.0 R
Expectancy = (0.42 x 2.4) - (0.58 x 1.0) = 1.008 - 0.580 = +0.428 R per trade
This strategy generates an average of 0.428 R per trade. Over 150 trades risking 1% per trade, the theoretical capital growth before transaction costs is approximately 64.2% (150 x 0.428 x 1%).
Statistical significance of expectancy
An expectancy calculated on fewer than 50 trades is statistically unreliable. Results may be driven by luck rather than edge. For expectancy to be meaningful, aim for a minimum of 100 trades covering different market conditions: trending, ranging, and high-volatility environments. The AMF (Autorite des marches financiers) recommends validating systematic strategies across representative periods of at least two to three years.
Expectancy per trade vs expectancy per dollar risked
Expectancy can be expressed in two equivalent forms:
- In monetary terms: "This strategy earns an average of $43 per trade." Useful for evaluating the actual account impact.
- In R-multiples: "This strategy has an expectancy of +0.43 R per trade." Universal and directly comparable across traders and account sizes.
R-multiples are the preferred format for strategy comparison and position sizing. They decouple the mathematical edge from the dollar amount risked, making system evaluation portable.
What is a good expectancy in trading?
Positive vs negative expectancy systems
Any positive expectancy is theoretically viable. In practice, however, the minimum threshold depends on your trading frequency and transaction costs. A spread of 1 pip on EUR/USD can wipe out a +0.05 R expectancy in Forex scalping.
| Expectancy per trade | Interpretation | Practical verdict |
|---|---|---|
| > +0.50 R | Strong system, prop firm ready | Excellent |
| +0.30 to +0.50 R | Solid system, validate on more trades | Good |
| +0.10 to +0.30 R | Fragile: transaction costs may eliminate the edge | Acceptable with caution |
| 0 to +0.10 R | Too thin to survive real-world costs | Insufficient |
| < 0 R | Losing system, do not trade live | Negative |
Benchmarks by strategy type (scalping, swing, trend)
Expectancy benchmarks vary by trading frequency. Comparing a scalper to a trend follower on a per-trade basis is misleading.
- Intraday scalping: low per-trade expectancy (0.10 to 0.25 R), compensated by high volume (10 to 30 trades per day). Frequency amplifies the cumulative impact of expectancy.
- Swing trading: higher per-trade expectancy (0.30 to 0.60 R) with 2 to 5 trades per week. Fewer trades means each one carries more statistical weight.
- Trend following: expectancy can exceed 1.0 R per trade with win rates sometimes below 35%. Frequent small losses are offset by rare, large winners.
The correct cross-strategy comparison uses weekly or monthly expectancy: (expectancy per trade) multiplied by (number of trades per period).
How to improve your expectancy without more trades
Three primary levers for improving the expectancy of an existing strategy:
For a full methodology to compare strategies using expectancy alongside other metrics, see our guide on using the expectancy formula to compare trading strategies.
Expectancy and position sizing
Kelly Criterion and expectancy
The Kelly Criterion uses expectancy to calculate the mathematically optimal percentage of capital to risk per trade for maximum long-term capital growth. The simplified formula is:
f = W/P - L/G
Where f is the optimal capital fraction. Using our example numbers (W=0.42, G=2.4, L=0.58, P=1.0):
f = (0.42/1.0) - (0.58/2.4) = 0.420 - 0.242 = 17.8% of capital per trade
Full Kelly is extremely aggressive. In practice, nearly all systematic traders use half-Kelly (8.9% in this example) to reduce variance and drawdowns. J.L. Kelly published this criterion in 1956 in the Bell System Technical Journal, and it remains a cornerstone of quantitative risk management.
For a complete application to trading, see our guide on position sizing with the Kelly Criterion.
Why expectancy alone does not guarantee profits
Positive expectancy is necessary but not sufficient for guaranteed profitability. Two important constraints:
-
Variance: even a strategy with +0.43 R expectancy can string together 15 consecutive losing trades by pure statistical chance. The probability of such a sequence is low but non-zero. Position sizing must account for this reality.
-
Temporal stability: an expectancy derived from backtesting may not reproduce in forward testing if market conditions shift. A trend-following system showing +0.50 R expectancy in a trending period may show +0.08 R in a prolonged range.
Backtest expectancy vs live performance
A backtest can show high positive expectancy and still perform negatively in live trading due to overfitting, real transaction costs (spread, commission, slippage), or regime change. Backtest expectancy is a starting point: validate it on out-of-sample data before committing real capital. Backtrex includes automatic overfitting detection to flag strategies that look good on paper but may not hold up out of sample.
Calculating expectancy from your backtest results
Extracting win rate and average R from backtest data
Calculating expectancy from a backtest requires four raw inputs: total trade count, winning trade count, R values of winning trades, and R values of losing trades.
Collect the raw trade log
Calculate win rate (W)
Calculate average win (G)
Calculate average loss (P)
Apply the expectancy formula
Assess statistical significance
Using expectancy to compare two strategies
Expectancy is one of the best tools for choosing between two competing strategies. Here is a concrete comparison:
| Criterion | Strategy A (Intraday scalping) | Strategy B (Swing on major pairs) |
|---|---|---|
| Win rate | 60% | 40% |
| Average win | 0.90 R | 2.80 R |
| Average loss | 1.0 R | 1.0 R |
| Expectancy per trade | +0.14 R | +0.52 R |
| Trades per week | 25 | 4 |
| Weekly expectancy | +3.50 R | +2.08 R |
Strategy A looks more consistent with its 60% win rate, but Strategy B offers nearly four times the expectancy per trade. On a weekly basis, Strategy A outperforms only because of volume. This comparison illustrates why expectancy must always be read alongside trading frequency.
For a broader evaluation including profit factor and Sharpe ratio, see our guide on backtest metrics: expectancy, profit factor, Sharpe ratio.
Important Risk Warning
Conclusion
Trading expectancy is the foundational metric that separates viable systems from losing ones. It is superior to win rate alone because it integrates both frequency and magnitude of gains and losses. To evaluate a strategy rigorously, always calculate its expectancy in R over a representative sample of at least 100 trades.
Backtrex automatically calculates expectancy and all key metrics for every backtest you run. Create a free account and get a guided tour to see your first real expectancy numbers in under 30 seconds. For context on what good expectancy looks like across different strategy types, read our guide on what a good trading expectancy looks like.
The formula is: Expectancy = (Win Rate x Average Win) - (Loss Rate x Average Loss). The result is expressed in R per trade (R being the amount risked per trade). A positive result means the system is theoretically profitable over a large number of trades; a negative result means it destroys capital over time.
Take your win rate (e.g., 0.42), multiply by your average win in R (e.g., 2.4 R), then subtract the product of your loss rate (0.58) and your average loss in R (1.0 R). Example: (0.42 x 2.4) - (0.58 x 1.0) = 1.008 - 0.58 = +0.428 R per trade. This means the strategy earns 0.428 R on average for every trade executed.
Any positive expectancy is theoretically viable. A value above +0.30 R per trade is considered solid for systematic strategies. Elite systems often reach +0.50 R or higher. Below +0.10 R, real transaction costs (spread, commission, slippage) may eliminate the statistical edge entirely in live trading.
A low win rate can produce positive expectancy if average wins are large enough. A 30% win rate with an average win of 3 R has an expectancy of (0.30 x 3) - (0.70 x 1) = +0.20 R, which beats a 60% win rate system with 0.80 R average wins: (0.60 x 0.80) - (0.40 x 1.0) = -0.08 R (negative). Expectancy depends on the product of frequency and magnitude, not on frequency alone.
Yes, in the short term. Statistical variance can produce extended losing streaks even with a positive expectancy system. Over 15 trades, randomness can dominate. Over 300 trades, expectancy asserts itself statistically. Position sizing is the tool that protects you from variance in the short run while expectancy does its work over the long run.
Three main levers: (1) optimize exits to let winners run longer, which increases G (average win); (2) filter out low-quality setups, which improves either G or W; (3) refine stop placement to reduce unnecessary losses, which decreases P (average loss). Any adjustment must be validated with a fresh backtest before being applied in live trading.
Expectancy measures the average gain per trade in R. Profit factor measures the ratio of total gross gains to total gross losses across the entire backtest. Both are complementary: a profit factor above 1.5 combined with an expectancy above +0.30 R generally indicates a robust strategy. Expectancy is more useful for position sizing decisions; profit factor is better for assessing overall system stability and comparing results across different time periods.