Expectancy, not win rate, determines whether a trading strategy makes money long term. A high win rate feels reassuring, but it says nothing about the size of your wins versus your losses, and a strategy can win most of its trades and still lose money. The formulas and worked examples below show how to calculate real expectancy from your own trade log.
TL;DR:
- Calculate expectancy as win rate times average win minus loss rate times average loss; express results in R to compare trades across account sizes.
- A 40% win rate, $300 wins, and $100 losses yields $60 per trade; a 70% rate, $80 wins, and $220 losses loses $10.
- Assess expectancy across 50 to 200 trades, and split results by market regime because performance can shift between trending, ranging, and volatile conditions.
- Subtract commissions, spreads, and slippage from net results, and compare backtests with verified live performance before trusting a strategy's stated edge.
- Positive expectancy does not prevent losing streaks; oversized positions can still ruin an account, so keep risk consistent and well below full Kelly sizing.
Table of Contents
- Win rate, loss rate, and the expectancy formulas you need
- Two worked examples: 40% win rate vs 70% win rate
- Calculating expectancy from your trade journal step by step
- Why win rate alone can hide a losing strategy
- What positive expectancy means and how to raise it
- Common misconceptions about win rate versus expectancy
- How expectancy connects to Sharpe ratio and drawdown
- What expectancy means for automated trading systems
- Expectancy in discretionary trading versus automated systems
- Reading expectancy on leveraged or amplified accounts
- Why we check the math before we check the win rate
- FAQ
- Sources
Win rate, loss rate, and the expectancy formulas you need
Win rate measures how often a strategy wins. Expectancy measures how much it makes, which is the number that actually matters for an account balance.
- Win rate = winning trades divided by total trades.
- Loss rate = 1 minus win rate.
- Expectancy (dollars) = (Win% × average win) minus (Loss% × average loss).
- Expectancy (R) = (Win% × average win in R) minus (Loss% × average loss in R), where R is your initial risk per trade.
The R-multiple version strips out position-size noise, since a $500 win on a $5,000 account and a $500 win on a $50,000 account mean very different things in dollar terms but can mean the same thing in R terms. Expressing every trade as a multiple of initial risk lets you compare strategies, timeframes, and account sizes on equal footing.
Breakeven win rate follows directly from reward-to-risk. A strategy with a 1:3 reward-to-risk ratio only needs to win a relatively low percentage of trades to break even, which is the core reason win rate and expectancy diverge so sharply at the extremes, as expected value theory from betting mathematics demonstrates.
Two worked examples: 40% win rate vs 70% win rate
Numbers settle the argument faster than intuition does.
- Example A, 40% win rate. Average win is $300, average loss is $100. Expectancy = (0.40 × $300) minus (0.60 × $100) = $120 minus $60 = $60 per trade. Over 1,000 trades, that strategy nets roughly $60,000 before costs.
- Example B, 70% win rate. Average win is $80, average loss is $220. Expectancy = (0.70 × $80) minus (0.30 × $220) = $56 minus $66 = negative $10 per trade. Over 1,000 trades, that strategy loses roughly $10,000 despite winning most of the time.
A 40% win rate strategy can produce six figures in positive expectancy while a 70% win rate strategy bleeds money, which is exactly the trap that expected value principles warn against: hit rate alone is meaningless without knowing the payout attached to each outcome. Converting both examples to R-multiples (Example A averages roughly 0.6R per trade, Example B averages roughly negative 0.045R) makes the comparison cleaner across different account sizes and risk settings, since dollar figures alone can mask how much capital was actually at stake.
Calculating expectancy from your trade journal step by step
Your broker export or trading journal already has everything you need. The process just requires pulling the right fields and running the right aggregations.
Start by extracting these fields for every closed trade: entry and exit timestamps, realized profit or loss, position size, and any commission or slippage adjustment that affects net P/L.
- Filter the export to closed trades only, discarding open positions that have not realized a result.
- Label each trade as a win or loss based on net P/L after costs.
- Calculate average win (mean P/L of winning trades) and average loss (mean P/L of losing trades, as a positive number).
- Divide winning trades by total trades to get win rate, then subtract from 1 for loss rate.
- Plug those four numbers into the dollar expectancy formula.
- Divide each trade's P/L by its initial risk amount to convert to R, then repeat the expectancy calculation in R terms.
Sample size matters more than most traders assume. A handful of trades can show a misleadingly high or low expectancy purely from chance. NIST's statistical sampling guidance on estimating means and variance supports using at least 50 to 200 trades before trusting an expectancy figure, since smaller samples carry wide estimation error. Our internal breakdown of win rate versus risk walks through this same calculation with R-normalization in more detail.
Pro Tip: Split your sample by market regime (trending, ranging, high volatility) before averaging, since a strategy's expectancy often looks very different in each condition.
Why win rate alone can hide a losing strategy
Win rate is a single number pulled from a distribution of outcomes, and distributions can be deceptive.
- Variance describes how much individual trade outcomes swing around the average, and high variance means your realized results can diverge from expectancy for long stretches.
- Skew describes asymmetry: a strategy with many small wins and a few rare, large losses has positive skew in frequency but negative skew in dollar outcomes, which is exactly the profile of Example B above.
- Fat tails (kurtosis) mean extreme losses happen more often than a normal distribution would predict, a pattern common in strategies that sell options or hold through news events.
- Metrics worth tracking alongside win rate include expectancy, average win and loss size, maximum drawdown, and risk of ruin, since no single number tells the full story.
Backtested results and live results frequently diverge once slippage, fills, and real-time execution enter the picture, which is why verifying backtest performance against live trading matters before trusting any stated expectancy, a distinction covered in our backtest versus live trading comparison.
What positive expectancy means and how to raise it
A positive expectancy means that, given consistent position sizing and enough trades to smooth out variance, a strategy has a mathematical edge over time. It does not guarantee any single trade or even any single month will be profitable, since variance can produce losing streaks even with a strong long-run edge.
Risk of ruin ties directly to expectancy and position sizing: a strategy with weak positive expectancy but oversized positions can still blow up an account during a normal losing streak. The Kelly criterion offers a conceptual ceiling for position size based on edge and odds, though most traders size well below full Kelly to reduce variance. Our risk of ruin breakdown covers how sizing changes survival probability for a given expectancy.
Practical levers to improve expectancy include:
- Tighten trade selection criteria to raise average win size without proportionally raising average loss.
- Cut stop-loss distance or improve exit timing to shrink average loss size.
- Reduce commission and slippage drag, which erodes expectancy on every single trade.
- Size positions consistently rather than increasing risk after losses or wins.
Any performance figures used to illustrate expectancy, whether backtested or hypothetical, should carry full disclosure of assumptions and methodology, a standard outlined in NFA rules on hypothetical performance, since hypothetical results can mislead without that context.
Common misconceptions about win rate versus expectancy
The most persistent misconception is that a high win rate equals a good strategy. Traders gravitate toward strategies that win often because losing streaks feel psychologically costly, even when the underlying math produces negative expectancy, as Example B illustrated earlier.
A second misconception treats win rate and expectancy as competing metrics rather than complementary ones. They are not opposed: win rate is one input into the expectancy formula, and the goal is never to maximize win rate in isolation but to maximize the product of win rate and payout ratio together.
A third misconception assumes expectancy calculated from a small sample is stable. A strategy that shows strong expectancy over 20 trades can regress sharply once sample size grows, which is why the sampling guidance mentioned earlier recommends 50 to 200 trades minimum before drawing conclusions.
A fourth misconception ignores costs. Commissions, spreads, and slippage apply to every trade regardless of outcome, and a strategy with marginal positive expectancy on paper can turn negative once real execution costs are subtracted. Traders who only check win rate from a platform's summary screen often miss this entirely, since that screen rarely nets out trading costs the way a full expectancy calculation does.
A fifth misconception treats expectancy as static. Market regimes shift, and a strategy's average win, average loss, and win rate all drift as volatility and trend conditions change, which is why ongoing recalculation matters more than a single historical snapshot.
How expectancy connects to Sharpe ratio and drawdown
Expectancy answers whether a strategy makes money on average. Sharpe ratio and drawdown answer how smoothly and how riskily that money gets made, and the three metrics work best read together rather than separately.
Sharpe ratio divides average return by the standard deviation of returns, which means two strategies with identical expectancy can post very different Sharpe ratios if one has far more volatile outcomes. A strategy with modest but highly consistent positive expectancy often posts a better Sharpe ratio than one with higher expectancy achieved through a few outsized wins and many small losses.
Maximum drawdown measures the worst peak-to-trough decline an account experiences, and it matters because expectancy is a long-run average that says nothing about the path to that average. Two strategies can share identical expectancy over 1,000 trades while one stays within a shallow drawdown band and the other swings through deep, account-threatening declines along the way. Our discussion of low-drawdown strategy design covers why drawdown control matters as a companion metric to expectancy rather than a replacement for it.
Reading expectancy alongside Sharpe ratio and drawdown gives a fuller picture: expectancy confirms the edge exists, Sharpe ratio shows how efficiently that edge converts to smooth returns, and drawdown shows the worst-case emotional and financial cost of staying invested through the strategy's rough patches.
What expectancy means for automated trading systems
Automated systems remove emotional decision-making from trade execution, but they do not remove the need to verify expectancy. An algorithm that claims an 80% win rate still needs its average win and average loss disclosed before that figure means anything.
Automated execution through the COPYX system also removes manual entry and exit delays that can quietly erode real-world expectancy versus a backtested figure, since every millisecond of slippage between signal and fill is a cost that subtracts from average win size or adds to average loss size.
Our verification checklist for gold trading performance walks through exactly which fields to check (win rate, average win, average loss, drawdown, and trade count) before trusting any automated system's published numbers, whether that system is ours or anyone else's.
Expectancy in discretionary trading versus automated systems
Discretionary traders calculate expectancy from a journal they maintain by hand, which means the quality of that expectancy estimate depends entirely on how consistently they log every trade, including the painful ones they might rather forget. Inconsistent journaling is the single biggest reason discretionary traders overestimate their own expectancy.
Automated systems generate a complete, unbroken trade log by design, since every entry and exit gets timestamped and recorded without the selective memory that affects manual journaling. That completeness makes expectancy easier to calculate accurately, though it does not make the underlying edge any larger. A discretionary strategy and an automated strategy with identical win rate, average win, and average loss produce identical expectancy. The real difference lies in execution consistency: a discretionary trader can deviate from their own rules during a losing streak, which changes the realized average loss away from the backtested or journaled figure, while an automated system executes the same rules on every single trade regardless of the trader's emotional state that day.
Verified, independently tracked performance such as MyFXBook verification matters more for automated systems precisely because the claim of consistency is central to their appeal, and that claim only holds up when backtest and live results are shown side by side.
Reading expectancy on leveraged or amplified accounts
Leverage changes the size of outcomes without changing the underlying win rate or the ratio between average win and average loss, which means expectancy measured in R-multiples stays constant across leverage levels.
What changes with leverage is the dollar expectancy and the variance of outcomes around it. This is also where risk of ruin becomes more sensitive to position sizing, since a losing streak that would be survivable on a standard account can erode a larger share of equity on an amplified one.
The 12X and 24X amplified account page outlines how the multiplier applies to the underlying Sonic AI strategy for traders considering that step.
Why we check the math before we check the win rate
Most traders ask about win rate first because it is the easiest number to understand. We have found the better question is always expectancy, since it is the number that actually predicts whether an account grows.
Hypothetical and backtested results deserve real skepticism. Any performance figure, ours included, should come with the assumptions behind it stated plainly rather than buried, and traders should always compare backtested numbers against live, verified results before relying on them. Our backtest versus live trading guide and our gold trading verification checklist both exist for readers who want to run that comparison themselves rather than take a published win rate at face value.
— Paulo
This article is general information, not a substitute for advice from a qualified financial advisor. Consult a qualified financial professional about your own circumstances before acting on anything here.
FAQ
Is a 40% win rate considered good?
Example A showed a 40% win rate producing roughly $60 of positive expectancy per trade when the win-to-loss ratio was 3:1.
Is 60% win rate good trading?
A 60% win rate paired with losses that are much larger than wins can still produce negative expectancy overall.
Is 30% win rate good?
A strategy with a 1:3 reward-to-risk ratio only needs to win a relatively low percentage of trades to break even.
What is a good win rate percentage?
There is no universal good win rate percentage, since the number only matters alongside average win and average loss size. A strategy's expectancy, calculated from at least 50 to 200 trades for statistical reliability, is the more reliable measure of whether a win rate is actually profitable.