Expectancy is the amount a strategy can expect to make or lose, on average, per trade. It combines the win rate, the average winner and the average loser into one figure, and it is the figure that all the others, the win rate, the ratio, the streaks, are really describing.
The calculation
Expectancy equals the win rate multiplied by the average winner, minus the loss rate multiplied by the average loser. A strategy that wins 45 per cent of trades with an average winner of £180 and an average loser of £100 has an expectancy of 0.45 multiplied by £180, £81, minus 0.55 multiplied by £100, £55: £26 per trade. Over 120 trades a year that is about £3,100.
Expressed in R, with the average loser as one unit, the same strategy has an expectancy of 0.45 multiplied by 1.8, minus 0.55, which is 0.26R per trade. The R version is more useful for comparing strategies, since it does not depend on account size or risk setting.
Working it out from a history
Export the closed trades from the MetaTrader 5 History tab, or read them from a verified record. Count the winners and losers, average the profit of the winners and the loss of the losers, including commission and swap in each, and apply the formula. A spreadsheet does it in a minute. The result is the realised expectancy, which already includes costs, slippage and every trade that was cut short, and it is the only expectancy that matters.
Frequency and costs
Expectancy per trade multiplied by the number of trades is the annual result, which is why a small positive expectancy at high frequency and a large one at low frequency can produce the same return. Costs work the other way: they are a fixed deduction per trade, so they take a larger share of a small expectancy. A strategy with an expectancy of £8 per trade before a £6 cost keeps £2. One with £40 before the same £6 keeps £34.
How many trades
An expectancy calculated from 20 trades is a guess. From 100 it is an estimate with a wide margin. From several hundred it starts to mean something, and the margin still depends on how variable the results are. The page on sample size covers this properly. For now, be suspicious of any expectancy, including your own, computed from fewer than a hundred trades.
A live record shows 240 closed trades over 18 months: 98 winners averaging £142 and 142 losers averaging £71, after costs. Expectancy is 0.408 multiplied by £142, £57.96, minus 0.592 multiplied by £71, £42.02: £15.94 per trade, about £3,830 over the period, and about 0.22R per trade. That is a modest, real edge. The same record shown as "41 per cent win rate" would have looked like a failure, and shown as "average winner twice the average loser" would have looked like a triumph. The expectancy says what it is.
An EA's live record contains everything needed to compute its realised expectancy, and computing it is the most useful ten minutes you can spend before buying. Do it at the risk setting you intend to use, in R rather than money if the record's account size differs from yours. Then multiply by the trade frequency to see the expected annual figure, and set that against the drawdown the same record shows.
If a vendor publishes only a return figure, a win rate, or a screenshot, the expectancy cannot be computed, and a strategy whose expectancy cannot be computed has not been shown to have one.