Prop firm challenges: how many of you failed because of the loss limits? - page 5

[ShowMeTheTrades]

Christian's post in #1 is the answer to the question, and it is worth separating two things that got mixed together there.

He failed because the EA's worst day was bigger than the firm's limit. That is not a strategy failure and not a rule failure. It is a unit mismatch: the strategy was sized in average-day units while the constraint is measured in worst-day units.

The arithmetic worth doing before paying for a challenge: take the backtest, sort days by equity change, and read the fifth worst day out of 100. If that day breaches the daily limit, the challenge is not testing the strategy, it is testing whether the bad day arrives before the profit target. No entry logic changes that ordering.

The number Christian asked for - how much all open positions together can lose if price jumps - is computable, and it is not the sum of the stops:

worst-case gap = sum over open positions of (adverse distance beyond the stop x lot value), plus the floating loss already carried

The second term is the one that gets forgotten. A grid already sitting in floating loss has spent part of the daily budget before the gap happens. That is precisely his case: the gap did not have to do the work alone.

On the "smaller size makes the challenge take forever" conflict - that is also arithmetic, not feel. If the drawdown has to come down by a factor of k, the position size comes down by about k, and so does the return. The ratio does not move. So the question is not how much to shrink, it is: what is the smallest k that keeps the worst day under the limit, and does the resulting return still clear the target inside the window? If it does not, no sizing choice fixes it. The strategy is simply incompatible with that challenge, and a different challenge is a cheaper answer than a different lot size.

One more thing that bites quietly and is worth checking before blaming the firm: many EAs measure daily loss on closed trades (balance), while the firm measures it on equity. A grid can be 4% down on equity and show zero realised loss, so the EA reports itself inside the limit right up to the moment it is not.