Socius NewsA 1% win rate can be profitable. Van Tharp's example is 99 losing trades at one dollar each and one winner at 500 dollars: 99 losses, one win, a net gain of 401 dollars. A 90% win rate can lose money by the same arithmetic in reverse. The win rate tells you how often you were right. It says nothing about what being right paid and what being wrong cost.
The number that does is expectancy: the average result per trade, measured in units of the risk you took. Two studies of real brokerage accounts explain why traders still watch the wrong number. Barber and Odean followed 66,465 households from 1991 to 1996 and found that the most active fifth earned 11.4% a year against a market return of 17.9%. Odean found that the same population sold winners far more readily than losers, and that the winners they sold went on to beat the losers they kept by 3.4% over the following year. Both behaviours raise your win rate. Both lower your expectancy.
What does expectancy actually measure?
Start with R. Van Tharp defines 1R as the amount you risk on a trade at entry: the distance from your entry to your stop, multiplied by your position size. A trade that loses exactly what you planned to risk is -1R. A trade that makes twice that is +2R. Every closed trade becomes an R-multiple, and expectancy is simply the mean of those R-multiples across many trades.
Written out, expectancy equals the win rate times the average winning R-multiple, minus the loss rate times the average losing R-multiple. A system that wins 40% of the time at +2.5R and loses 60% of the time at -1R has an expectancy of 1.0 minus 0.6, which is +0.4R. A system that wins 80% of the time at +0.5R and loses 20% of the time at -2.5R has an expectancy of 0.4 minus 0.5, which is -0.1R. The second system feels better four days out of five. It loses.
+401 USD
Net result of Van Tharp's 1% win-rate example
99 losses of 1 USD, one win of 500 USD
11.4%
Annual return, most active 20% of households
Market 17.9%, 1991 to 1996 (Barber and Odean 2000)
0.148 vs 0.098
Share of paper gains realized vs share of paper losses realized
10,000 accounts, 1987 to 1993 (Odean 1998)
3.4%
By how much winners sold beat losers kept over the next year
Odean 1998
Expectancy is a per-trade number. It says how much, per dollar risked, you can expect on average over many trades. It does not say how much you will make this month, because that depends on how many trades you take and how much you risk on each, which is position sizing. Tharp's point is that position sizing and psychology, not the entries, separated bankruptcy from a 13 million dollar outcome when different people were handed the same trades. Expectancy is the input that position sizing multiplies.
Why does a high win rate feel better than it pays?
Because losses hurt more than gains please. Kahneman and Tversky's 1979 paper in Econometrica describes a value function that is steeper for losses than for gains. A closed loss of 1R feels worse than a closed gain of 1R feels good. A trader who wants to feel good more often will, without deciding to, take gains early and let losses run. That is a strategy for a high win rate and a small average win.
Odean measured it. Across 10,000 discount-brokerage accounts between 1987 and 1993, investors realized 14.8% of their available paper gains and 9.8% of their available paper losses. They were about one and a half times more likely to sell a winner than a loser. The winners they sold then outperformed the losers they held by 3.4% over the following twelve months. Holding losers did not just lower the win rate's meaning. It lowered the returns.
Win rate versus expectancy, three demo strategies
Illustrative demo data. Expectancy in R per trade
Source: Socius Trades demo journal
The three strategies above are illustrative demo data. Strategy A wins 82% of the time and loses money, because its average loss is more than five times its average win. Strategy C wins one trade in three and has the best expectancy of the three, because its winners average more than three times its losers. The trader running A has the most pleasant month. The trader running C has the most money at the end of it.
Barber and Odean's 66,465 households show the same mechanism at the scale of a whole population. The average household turned over 75% of its portfolio each year and earned 16.4% net against the market's 17.9%. The most active fifth earned 11.4%. Activity is not the same as edge, and a stream of small realized gains is not the same as expectancy.
What do the prop firms measure instead?
Look at what a funded-account evaluation asks for. FTMO's published trading objectives for its two-step challenge, as read on 26 August 2026, are a 10% profit target in the first phase, 5% in the second, a maximum daily loss of 5% of initial capital, a maximum total loss of 10% and at least four trading days. There is no win-rate requirement anywhere in the list. FTMO is a trademark of its owner; Socius Trades is not affiliated with FTMO.
Read those rules as a statement about expectancy and drawdown. Reaching 10% before losing 10% is a question of how many R you can gain before a run of losses costs you 10% of the account. With 1% risk per trade, the 10% limit is 10R of losses. With 2% per trade it is 5R. The win rate determines how long a losing streak is likely to run. The expectancy determines whether the account climbs between streaks. The risk per trade determines whether the account is still there when the streak ends. Three numbers, none of them the one most traders quote.
What your journal would show
The following figures are illustrative demo data, not a real account.
A demo account with 300 closed trades shows a win rate of 68%. That number sits on the summary screen and it looks like a good year. The R-multiple view underneath it shows an average winner of +0.6R and an average loser of -1.4R. Expectancy: 0.68 times 0.6, minus 0.32 times 1.4, equals -0.04R. Three hundred trades, two-thirds of them winners, and the account is slightly smaller than when it started before costs.
The breakdown by instrument shows where the -1.4R comes from. On one index CFD the average loser is -2.3R against a planned -1R: the losses ran well past the level the trader had written down. On the two currency pairs, the average loser is -1.0R and the expectancy is +0.15R. The same trader, with the same entries, is profitable in two instruments and unprofitable in one, and the win rate is almost identical across all three.
You can see the same cut on your own history. Socius Trades imports your executed trades from cTrader or MetaTrader 4 and 5, converts each one to an R-multiple against the risk you defined, and shows expectancy by instrument, session, hour and day. Ask Socius AI, in plain language, which instrument has your worst average loser, and the answer comes from your fills. We look. We never touch.
“Your win rate is the number you tell people. Your expectancy is the number your account already knows.”
Trading involves risk of loss. Socius Trades is an analytics tool, not investment advice.