Prop firm rules your EA breaks without meaning to

Prop firms rarely fail an account because the expert advisor was unprofitable. They fail it because the EA did something the rulebook forbids, often once, often by a small margin, and often without the trader knowing until the payout review. This guide lists the rules automated strategies break most, shows what each one looks like in telemetry, and gives a way to audit your own EA before the firm audits it for you.

Rules differ between firms and change over time. Treat the list as a map of where to look, then read the current terms of the firm you actually trade with.

The rules, and how an EA breaks each one

Common prop firm rules and typical EA violations

Rule	How an EA breaks it	What to watch
Maximum daily loss	Floating loss spikes through the limit on an open position; several EAs share one account.	Equity vs day-start reference, per minute.
Maximum overall loss	Slow bleed over weeks; trailing versions move the floor up after profits and the EA's fixed stop does not follow.	Equity vs the firm's current floor, not vs initial balance only.
News trading restrictions	EA opens or closes a trade inside the restricted window around a high-impact release. A stop-loss that triggers in the window can count too. At FTMO the window is 2 minutes before and after the release and applies to funded accounts, not to the Challenge.	Order timestamps vs calendar events for the traded currencies.
Maximum lot / exposure per symbol	Risk-percent sizing with a tight stop produces a very large lot; grid levels add up past the cap.	Sum of open lots per symbol, not per order.
Martingale, grid or "all-in" prohibition	Lot size increases after a loss; averaging into a loser. Reviewers detect it from the pattern, whatever the EA is called.	Lot size sequence after losing trades.
Consistency rules	One day or one trade makes most of the profit, because the EA raised size after a good streak.	Share of total profit per day and per trade.
Minimum holding time / no tick scalping	Scalpers closing within seconds; latency-style entries.	Distribution of trade durations.
Copy trading and identical trades across accounts	The same commercial EA with default settings on many traders' accounts produces identical orders; firms group them and apply a shared exposure cap.	Whether your settings differ from defaults; magic numbers and timing.
Weekend / overnight holding	Swing EA leaves positions open through the Friday close on a programme that forbids it.	Open positions at the firm's cut-off time.
Inactivity	The EA stopped and nobody noticed; FTMO sends a warning after 30 days without a trade and closes the account after 30 more.	Time since last report and since last trade.

The lot you asked for is not always the lot you got

One class of violation deserves its own section because it is invisible in the EA's own logs. The EA computes a lot, a risk module multiplies it, the broker rounds it to the volume step, and a partial fill or a margin limit changes it again. The number in your backtest, the number in your log and the number on the account statement can be three different numbers.

An audit needs all of them side by side: base lot, requested lot, final executed lot and the multiplier that was applied, with its source. If the executed lot differs from the expected one by more than a tolerance (5% is a reasonable default), that trade is flagged. Most "I never trade more than 1% risk" disputes end when someone finally lines these four numbers up.

What the firm sees, and what you should have

The firm has a complete record of every order on its server, plus tooling that groups accounts by behaviour. You usually have MT5's history tab and memory. That asymmetry is the real problem: when a payout is denied, the conversation is the firm's data against your recollection.

An independent, time-stamped record changes that. It does not need to be able to trade; in fact it is more credible if it cannot. It needs to record, for every decision the EA makes: time, symbol, timeframe, the decision, whether trading was allowed, the reason if it was blocked, the spread at that moment, and the lot chain described above.

A self-audit you can run this week

1. List the rules of your current programme in one page, with the numbers.
2. Export the last 30 days of trades from MT5 history.
3. Daily loss: for each day, find the worst equity point against the day-start reference. How close did you get?
4. Exposure: find the maximum simultaneous lots per symbol. Compare with the cap.
5. News: cross the order times with a high-impact calendar for the currencies you trade. Any order inside the restricted window?
6. Sizing pattern: sort by time and look at lot size after each losing trade. Does it ever go up?
7. Consistency: what share of the month's profit came from the best day?
8. Lot audit: compare the lot your EA logged with the lot in the history for ten random trades.
9. Silence: what was the longest gap without any EA activity while the market was open, and do you know why?

If any answer is "I can't tell from the data I have", that is the finding. The fix is not a better strategy; it is better telemetry.

Where monitoring fits

Quantisentry is built around this problem. Each EA reports through a fixed contract that includes the decision, the block reason, spread and the full lot chain; the platform runs the 5% lot-audit tolerance and the freshness window automatically; and everything is served read-only, so the record cannot be accused of influencing the trades it describes. It does not replace your firm's rules or guarantee a payout. It gives you the same quality of evidence the firm already has.

This article is educational and is not investment advice. Trading leveraged products carries a high risk of loss.