Risk per trade is the fraction of the account that a single trade can lose if its stop is hit. Common figures are half of one per cent, one per cent and two per cent. The number looks small and its consequences are not, because losses arrive in runs and the account has to survive the run.
Consecutive losses
Every strategy has losing streaks, and their length is set by the win rate, not by skill. A strategy that wins 40 per cent of its trades will, over a few hundred trades, routinely see six or seven losses in a row and occasionally ten. A strategy that wins 55 per cent will see five or six regularly and eight or nine at some point.
Ten consecutive losses at one per cent risk leaves the account down about 9.6 per cent, because each loss is one per cent of a slightly smaller balance. At two per cent it is down 18.3 per cent. At five per cent, 40 per cent. At ten per cent, 65 per cent. The strategy did the same thing in each case. The risk setting decided whether the run was a bad month or the end.
Recovery is not symmetrical
A loss of 10 per cent needs a gain of 11 per cent to recover. A loss of 25 per cent needs 33 per cent. A loss of 50 per cent needs 100 per cent. The deeper the drawdown, the more disproportionate the climb back, which is why keeping drawdowns shallow matters more than any amount of upside.
Open risk across positions
Risk per trade is only half the picture when more than one position is open. Three positions each risking one per cent is three per cent at risk if all three stop out, and if they are in correlated markets, three US stock indices for instance, they are likely to stop out together. Total open risk, the sum of the risk on every open position, is the figure that describes the account's real exposure at any moment, and a sensible ceiling on it is as important as the per-trade figure.
A trader risks two per cent per trade and runs a strategy with a 45 per cent win rate. Over 200 trades the longest losing run is nine. Starting from £10,000, nine losses take the account to about £8,340. That is uncomfortable and survivable. The same trader at five per cent per trade after the same nine losses is at £6,300, and needs a 59 per cent gain to get back to the start. A strategy that is perfectly good has been made dangerous by the risk setting alone.
An EA's risk setting is the same decision, made once and applied to every trade. A multi-strategy EA adds a second layer: because several strategies can be in the market at the same time, the EA should size each so that total open risk stays within a ceiling, and its documentation should say what that ceiling is. The risk settings such an EA offers, usually named as low, medium and high or similar, scale that whole structure up and down.
The right setting is the one whose worst drawdown, taken from the EA's live record, you would actually sit through. If the record at a given setting shows a 20 per cent drawdown and you know you would switch the EA off at 12, that setting is too high for you regardless of its returns, because you would leave at the bottom.