Why is passing a prop firm challenge different from being profitable?
Passing an evaluation is different because it imposes path constraints that a normal account does not. Profitability is about where you end up; an evaluation is about every point along the way.
The constraints are specific: a maximum loss in any single day, a maximum drawdown that may trail your high-water mark, a minimum number of trading days, and often a cap on how much of your total profit may come from a single day.
A strategy that comfortably makes money over a year can still breach a daily loss limit during its worst week and end the attempt in an afternoon. The distribution of your losses matters as much as their sum — and a backtest reporting annual return says nothing about it.
This is why traders with genuinely profitable systems fail evaluations repeatedly. They are being measured on something their strategy was never tuned for.
What rules should you test against?
Test against four rule types, and test each one against your actual trade history rather than against an average.
Daily loss limit. Group historical trades by trading day and find the worst day. Then count how often days approached that threshold, because one bad day near the limit is survivable and a pattern of them is not.
Maximum drawdown, static or trailing. Trailing versions are considerably harsher: the limit rises with your equity high-water mark and never falls back, so a strategy that runs up and gives back is punished in a way a static limit would not punish it.
Minimum trading days. A selective strategy may pass on profit and fail on activity. If your rules produce four trades a month and the firm requires ten trading days, the strategy needs adjusting or the evaluation needs to be a different one.
Consistency rules. Some firms void an attempt where a single day contributes more than a set share of total profit. This specifically penalises strategies with occasional large winners — which includes most trend-following approaches — and it is the rule people most often discover after passing everything else.
How do you adjust a strategy to pass an evaluation?
The adjustment is almost always position size, not strategy logic, and that is fortunate because changing the logic to pass a rule usually breaks the edge.
The same strategy at half the risk per trade has roughly half the daily loss exposure and roughly half the drawdown depth, at the cost of taking longer to reach the profit target. For most evaluations that trade is strongly favourable, because the time limit is usually generous and the loss limits are not.
The second adjustment is trade clustering. If your strategy can take five correlated positions simultaneously — say, five USD pairs on the same signal — your effective daily risk is not five times one trade's risk but something closer to one large trade. Capping concurrent correlated exposure addresses the daily-loss rule directly.
The third is scheduling. If a disproportionate share of your worst days involve trading through major scheduled news, sitting those out costs a little expectancy and removes a large share of the tail risk that ends evaluations.
How do you test prop firm rules properly?
Test properly by replaying your actual trade sequence day by day against the specific rule set, rather than estimating from summary statistics.
Averages are useless here. What matters is whether any single day breached the limit, and a monthly average return tells you nothing about your worst Tuesday. The test has to walk the equity curve in order, applying each rule at each point, and stop at the first breach the way the firm would.
Run it repeatedly from different starting points, too. An evaluation beginning in January and one beginning in March are different tests of the same strategy, and a strategy that passes from most starting points is a different proposition from one that passes from a lucky few.
Finally, be honest about which historical trades you would actually have taken. Backfilling an evaluation with trades you identified in hindsight produces a pass that means nothing.