Trading expectancy and win rate
Calculate historical average profit or loss per trade, distinguish payoff ratio from planned reward-to-risk, and account for costs and sample limits.
Expectancy summarises average profit or loss per trade in a defined sample. It helps explain why win rate alone does not describe performance. A positive historical result is not a prediction that the next trade, or a future sample, will be profitable.
Start with completed trades
Section titled “Start with completed trades”Choose one account currency, a defined period and a consistent set of completed trades. Keep simulated backtest results separate from broker results. For the historical sample, the direct calculation is:
Average net result per trade = total net profit or loss ÷ number of completed trades.
TradingView’s expected payoff definition also uses completed trades. Open positions are not part of that realised result.
Calculate from wins and losses
Section titled “Calculate from wins and losses”For a sample with winning and losing trades and no breakeven trades:
Expectancy = (win fraction × average win) − (loss fraction × average loss magnitude).
Use positive amounts for both average win and average loss magnitude. The win and loss fractions sum to 1 in this two-outcome example. If you have breakeven trades, include them in the total count and use the actual winning and losing fractions; do not assume loss fraction equals 1 minus win fraction.
Worked example
Section titled “Worked example”Suppose 20 completed trades contain eight wins averaging 200 currency units and 12 losses averaging 100 currency units, before costs:
- Win fraction: 8 ÷ 20 = 0.40.
- Loss fraction: 12 ÷ 20 = 0.60.
- Average result: (0.40 × 200) − (0.60 × 100) = 20 currency units per trade.
Relative to the average loss of 100, that is 0.2 times the average loss. It is not a 20% return on the account. This is an arithmetic illustration, not a performance estimate for a strategy.
Payoff ratio and breakeven win rate
Section titled “Payoff ratio and breakeven win rate”The observed payoff ratio is average realised win ÷ average realised loss magnitude. A planned take-profit/stop-loss ratio may differ from that observed ratio. TradingView explains this distinction in its average profit / average loss reference.
With no breakeven trades and before costs, the win fraction needed to break even is 1 ÷ (1 + payoff ratio).
| Average win ÷ average loss | Breakeven win rate before costs |
|---|---|
| 0.5 | 66.7% |
| 1 | 50% |
| 2 | 33.3% |
| 3 | 25% |
| 5 | 16.7% |
| 10 | 9.1% |
Include costs and the sample’s limits
Section titled “Include costs and the sample’s limits”Use net outcomes after relevant commissions, financing and other charges. Spread and slippage affect realised fills; do not subtract them a second time if already reflected in your results. If the example above excluded an average cost of 25 units per trade, its average net result would become −5, not +20.
Record the sample size, dates, strategy version, market and sizing method. Changing these changes what the average represents. A small sample, a few unusually large winners, or a change in market conditions can make the historical average a poor guide to what follows. Average profit alone also does not describe drawdown or the order of wins and losses.
For TradingView testing assumptions, review its strategy testing guidance. For displaying supported trading history in PineConnector, see Trading analytics. To check whether your automation sends the intended request and produces the expected demo order, use Test your setup; that operational check does not establish profitability.