Win rate is the percentage of trades you close in profit. Profitability is how much money your account actually grows. They correlate weakly. A trader winning 70 out of 100 trades can lose money, and a trader winning 35 out of 100 can make money — because profitability is driven by expectancy, not frequency.
According to Brad Barber and Terrance Odean’s research on retail day traders (University of California), roughly 70-80% of active day traders lose money over any 12-month window. The top 20% who survive aren’t the ones with the highest win rates — they’re the ones with the highest expectancy. This guide breaks down the math, the psychology that sabotages it, and how to track the metric that actually predicts your P&L.
What is expectancy, and why does it beat win rate?
Expectancy is the dollar amount you expect to earn (or lose) per trade over a large sample. The formula:
Expectancy = (Win% × Average Win) − (Loss% × Average Loss)
A positive number means your edge is real. A negative number means every trade bleeds capital, regardless of how often you “win.”
The 70% vs 35% example
Two traders each take 100 SPY trades over a quarter.
Trader A — 70% win rate:
- 70 wins at +$50 each = +$3,500
- 30 losses at −$200 each = −$6,000
- Net: −$2,500
- Expectancy: (0.70 × $50) − (0.30 × $200) = −$25 per trade
Trader B — 35% win rate:
- 35 wins at +$300 each = +$10,500
- 65 losses at −$100 each = −$6,500
- Net: +$4,000
- Expectancy: (0.35 × $300) − (0.65 × $100) = +$40 per trade
Same instrument, same trade count. Trader B wins half as often and makes $6,500 more. The only variable is risk-reward discipline.
Break-even win rate by risk-reward ratio
Risk-reward ratio (R:R) is how many dollars you stand to make relative to what you risk. At 1:3 R:R, a $100 risk targets a $300 gain.
The break-even win rate formula is:
Break-even = 1 / (1 + R)
| Risk-Reward (R:R) | Break-even Win Rate |
|---|---|
| 1:1 | 50.0% |
| 1:2 | 33.4% |
| 1:3 | 25.0% |
| 1:5 | 16.7% |
| 1:10 | 9.1% |
The Turtle Traders of the 1980s famously ran ~35% win rates with 2.5:1+ R:R and produced 80%+ annual returns for nearly a decade, per Michael Covel’s documented history of the experiment. Their edge wasn’t accuracy — it was letting winners run 3-5x their risk while cutting losers at −1R.
Why most traders cut winners and hold losers
The psychology is Nobel-Prize-winning, not opinion. Daniel Kahneman and Amos Tversky’s prospect theory (1979) established that losses feel 2 to 2.5x more painful than equivalent gains feel good.
That asymmetry drives a predictable pattern:
- Winner at +$80? Brain says: “Lock it in before it turns red.” Close at +$80 instead of letting the planned +$300 target hit.
- Loser at −$120? Brain says: “If I close, it becomes real. Let me give it room.” Hold until −$400.
Over 50 trades, this behavior crushes expectancy even when the underlying strategy was sound. A setup that backtested at +0.6R becomes −0.2R in live trading because the trader overrides the exit rules on both sides.
The fix is mechanical: pre-define entry, stop, and target before the trade. Log every override. After 30 days of journal review, the pattern is impossible to ignore.
Expectancy by trading style
Different styles produce different win-rate ceilings. Judging a trend follower by scalper metrics is how traders abandon profitable systems.
- Scalping (SPY 0DTE, /ES futures): 60-70% win rate, 1:1 to 1:1.5 R:R. Edge comes from frequency — 20-40 trades/day. Requires tight execution; one missed stop erases a week.
- Swing trading (3-10 day holds): 45-55% win rate, 1:2 to 1:3 R:R. Most profitable retail traders operate here. Realistic expectancy: +0.3R to +0.6R.
- Trend following (weeks to months): 30-40% win rate, 1:3 to 1:5+ R:R. 5-10 trades per year drive 80%+ of P&L. Requires psychological tolerance for 10+ losses in a row.
Van Tharp’s widely-cited benchmark: +0.5R per trade is a strong edge. Anything above +0.3R over 200+ trades is genuinely tradeable. Below +0.1R, you’re breaking even after commissions and slippage.
Position sizing: the Kelly Criterion
Once you have a real edge, the next question is how much to risk per trade. The Kelly Criterion answers it mathematically:
f* = (bp − q) / b
Where b is your reward-to-risk odds, p is your win probability, and q = 1 − p.
Worked example: you win 40% of the time with 2:1 R:R.
- b = 2, p = 0.40, q = 0.60
- f* = (2 × 0.40 − 0.60) / 2 = 0.10, or 10% of account per trade
Full Kelly is brutal — a single bad variance cluster can produce a 40-50% drawdown. Most professionals use Half-Kelly or Quarter-Kelly (2.5-5% per trade) to account for estimation error in p and b. FTMO and Topstep prop firms cap risk around 1-2% per trade precisely because sub-Kelly sizing preserves capital through bad streaks.
Tracking expectancy per setup, not in aggregate
Aggregate expectancy hides everything useful. A trader might be +$3,000 over 200 trades and assume the edge is real — when in reality one setup (breakout longs on /ES during the first hour) produced +$8,000 and three other setups lost −$5,000 collectively.
The journaling framework that surfaces this:
- Tag every trade with a setup name — e.g.,
opening-range-breakout,vwap-reclaim,earnings-fade. - Log R multiple per trade — if you risked $100 and made $230, that’s +2.3R. If you risked $100 and lost $100, that’s −1R.
- Compute expectancy per setup after 20+ trades —
Avg Racross each tag. - Kill any setup below +0.1R expectancy after 30+ samples.
A trading journal like JournalPlus automates the R-multiple calculation and setup tagging, so the per-strategy breakdown is available without spreadsheet math.
The profit factor shortcut
If expectancy feels abstract, profit factor is the one-number sanity check:
Profit Factor = Gross Winners ÷ Gross Losers
- Below 1.0: losing strategy.
- 1.0 to 1.2: marginal, likely breakeven after costs.
- 1.2 to 1.5: tradeable edge (prop firm minimum territory).
- 1.5 to 2.0: strong edge.
- Above 2.0: either exceptional skill or small-sample noise — test for another 100 trades.
FTMO’s evaluation criteria require a profit factor above 1.2 and minimum 1:1 R:R to pass challenge phase. That floor is low specifically because anything below it can’t survive live costs.
Conclusion
Win rate is vanity. Expectancy is sanity. A 35% win rate at 1:3 R:R beats a 70% win rate at 1:0.25 R:R every time, and the math is uncontroversial.
The behavioral trap is that humans are wired to optimize the wrong variable — we prefer feeling right (higher win rate) over being profitable (higher expectancy). The only defense is systematic journaling: track R multiples per setup, compute profit factor monthly, and kill any setup below +0.1R expectancy over 30+ samples. The traders who survive long enough to compound aren’t the ones who win most often — they’re the ones who stopped trying to.
People Also Ask
What is the expectancy formula, and why does it matter more than win rate?
Expectancy is (Win% × Average Win) − (Loss% × Average Loss). It tells you the dollar amount you expect to earn per trade over a large sample. A trader with a 70% win rate averaging $50 wins and $200 losses has a negative expectancy of −$25 per trade. A trader with a 35% win rate averaging $300 wins and $100 losses has a positive expectancy of +$40 per trade. Expectancy accounts for both frequency and magnitude; win rate only measures frequency.
What win rate do I need to break even at different risk-reward ratios?
At 1:1 R:R you need 50% to break even. At 1:2 R:R you need 33.4%. At 1:3 R:R you need 25%. At 1:5 R:R you need 16.7%. The formula is: Break-even win rate = 1 / (1 + R), where R is your reward-to-risk multiple. This is why trend-followers can run 30-40% win rates and still generate 80%+ annual returns — their R:R is 3:1 or higher.
Why do most traders end up cutting winners early and letting losers run?
Kahneman and Tversky's prospect theory (1979) showed that losses feel 2 to 2.5 times more painful than equivalent gains. The brain treats a $100 loss as worse than a $100 gain feels good. This bias pushes traders to lock in small wins (to feel 'right') and hold losers (to avoid realizing a loss). The result: high win rate, terrible expectancy. Journaling each trade's R multiple — how much you risked vs. how much you made or lost — surfaces this pattern in weeks.
What win rates should different trading styles expect?
Scalpers typically run 60-70% win rates with 1:1 or 1:1.5 R:R, relying on frequency. Swing traders average 45-55% with 1:2 to 1:3 R:R. Trend followers (like the famed Turtle Traders) historically hit 30-40% win rates with 1:3 to 1:5+ R:R, accepting many small losses to catch 5-10 outsized wins per year. All three can be profitable; all three have different psychological demands.
How do I use the Kelly Criterion to size positions?
Kelly formula: f = (bp − q) / b, where b is your odds (reward ÷ risk), p is win probability, q is loss probability (1 − p). If you win 40% of the time with a 2:1 R:R, Kelly says bet (2 × 0.4 − 0.6) / 2 = 10% of your account per trade. Most pros use Half-Kelly or Quarter-Kelly (2.5-5%) because estimation errors in p and b compound quickly at full Kelly, producing 50%+ drawdowns.
What's a good expectancy benchmark?
Van Tharp's widely-cited benchmark is +0.5R per trade, meaning you earn half your risk amount on average. If you risk $100 per trade, you'd average +$50 per trade across 100+ trades. Elite systematic traders often hit +0.8R to +1.5R. Prop firms like FTMO and Topstep require a minimum profit factor (gross wins ÷ gross losses) above 1.2 and a minimum 1:1 R:R, which implies positive expectancy regardless of win rate.
Why does optimizing for win rate kill profitability?
Traders chasing a higher win rate move stop-losses wider and take profits earlier. Both changes compress R:R. A strategy that once won 45% at 1:3 R:R (+0.8R expectancy) becomes a strategy that wins 60% at 1:1 R:R (+0.2R expectancy) — higher win rate, 75% less money. The only metric that matters is expectancy per setup, not win rate.
