Running an EA on a Prop-Firm Account: Why Max Drawdown Is the Wrong Number

Buyers increasingly ask whether an EA is suitable for a prop-firm evaluation. The honest answer for most EAs on the marketplace, ours included, is that they were not designed around a daily loss cap, and the number the product page shows you, maximum drawdown, is not the number a prop account fails on. Translating one into the other is work the buyer has to do, and nobody else will do it for them.

Maximum drawdown in a report is peak-to-valley over the whole test, often years. It can be built from twenty small losing days spread across three months. A prop account has a daily cap, commonly 4 to 5 percent, and a total cap, commonly 8 to 10 percent, both measured on equity, so open floating loss counts, and the daily figure usually resets at a fixed server time. An EA with 25 percent maximum equity drawdown is disqualified before you start, because the total cap is below its normal behaviour. An EA with 10 percent can still fail, because 10 percent over years says nothing about how much of it can arrive in one day.

Open the backtest report or export the trade list and find the worst calendar day by closed loss plus the floating loss at its low point. That is the figure to compare against the daily cap, not maximum drawdown. Then scale. If the worst day at the tested risk was 6 percent and the cap is 5, halving the lot gives 3 percent, and applying a safety margin of two, since the future contains days worse than the past, brings you to a quarter of the tested lot. A quarter lot also means a quarter of the return, which is the point most buyers stop at, because the profit target of the evaluation is now far away. That arithmetic is the real product-fit test, and the product page cannot do it for you.

Because the daily cap includes open positions, an EA that holds trades overnight can be inside the limit at every close and breach it at the reset with a position that later recovers. A backtest counts that trade as a win; the prop account counted it as a failure hours earlier. Know exactly when your firm resets, look at what the EA typically holds at that time, and treat the floating loss at that moment as a hard number, not a temporary one.

Any EA that adds to a losing position is designed to run a large floating loss and wait for it to recover. That is exactly the shape a daily equity cap is built to catch. It does not matter how good the long-run backtest looks; the strategy's normal operation is a breach. This is not a settings problem and no lot size fixes it, because the ratio between the first entry and the full basket is the strategy.

Many firms ban holding through scheduled news, some require lot sizes to be consistent across trades, some restrict weekend holds, and some judge consistency by how much of the profit came from the best day. An EA that scales lots after a win, trades the open of a release, or makes its month on two trades can fail an account while being profitable. Read the rules of the specific programme against what the EA actually does, trade by trade, not against its description.

We publish period, balance and settings for our backtests, and our forward accounts are demo. We do not publish a prop-firm setting because there is no single one; the right lot depends on the cap, the reset time and the margin you want, and any set file we labelled for prop use would be guessing at your numbers. Take the worst-day figure from the report, scale it yourself, and run it on a demo with the same rules before you pay an evaluation fee. If that scaled version cannot reach the target, the EA is not a fit, and it is better to learn that from the arithmetic than from the fee.

Our three best-selling EAs: GOLD NEURON (AI) https://www.mql5.com/en/market/product/187329 · ATLAS PORTFOLIO https://www.mql5.com/en/market/product/182751