Prop Firm EA Rules

Why firms restrict automation

A prop challenge sells a test and, for those who pass, access to a funded account. The firm carries the risk of that account, so it has reasons to ask how a pass was achieved. A bot is not the issue. Some methods of passing do not show skill at all, they exploit the structure of the test. A rule that limits them is a rule about behaviour, and a bot that behaves that way is caught by it, whether it was written for that purpose or not.

The reasons that recur are the same in rule sets of different firms, so they are worth understanding in general terms.

• Latency arbitrage. A bot trades on a price that has not yet reached the firm's feed. The result looks like a high win rate but is a timing gap. It is not a strategy that works on an ordinary feed.
• Copying. Many accounts following the same trades turn one bet into many. If it loses, every account loses. If it wins, many accounts pass on a single idea.
• Overlapping accounts. Taking opposite sides on two accounts so that one is sure to pass. The next section shows the arithmetic.
• Martingale. Doubling after a loss gives many small wins and a rare large loss. A challenge only needs a pass in a short window, so it can be passed on the small wins before the large loss arrives. See martingale and the rules against bot types page.

The kinds of restriction found

The wording differs by firm, and several firms may group these differently. The table lists the kinds of rule that turn up, why a firm might write one, and the bot designs it touches.

Other terms often sit beside these, such as limits on news trading, weekend holds and how many trades per day. They are covered with the loss limits in prop firm drawdown rules and the consistency rule.

A worked example: why paired accounts are restricted

Take two made-up accounts of $10,000, each with a 10 percent target ($1,000), a 10 percent loss limit ($1,000) and a fee of $150. Someone opens a long position of a size that gains $1,000 on a given move on account A, and the same size short on account B. Whichever way price moves, the sums are the same: one account is $1,000 ahead and the other is $1,000 behind. A move that reaches the target on one account reaches the limit on the other. Fees were $300 for the two accounts against $150 for one.

Seen from the firm, one account passed and one failed, and the person is paid once, at a cost to the firm that came from no skill at all. That is why overlapping accounts, hedging between accounts and group trading are the kind of behaviour a rule is written for. A bot that runs on more than one account, even honestly, can look the same in the data, so the safe reading is to treat multi-account use as a restricted behaviour until the terms say otherwise.

The same logic applies to copying. If 3 accounts each risk 1 percent on the same signal, the single idea puts 3 percent of combined capital at stake, while each account sees only 1 percent. The accounts are not independent bets. A loss on the idea is a loss on all of them at once.

A pre-flight checklist for reading any firm's terms

Treat the terms as a specification. Before you run a bot, read them in full, with the date they were published, and turn each clause into a question about your bot's design. The table shows five made-up clauses and the question each one asks. None is from a real firm.

1. Read the current version. Rules change, so note the date and read the version that applies to your account, not an old copy.
2. Find the section on automation. Look for words like automated, bot, expert advisor, copy, signal, hedge, latency and high frequency.
3. Write down the bot's behaviour in plain facts. Speed, holding time, sizing rule, number of accounts, where it runs, who wrote it.
4. Match facts to clauses, one by one. Mark each as clearly outside, clearly inside, or unclear.
5. Ask the firm in writing about every unclear clause and keep the reply with the date. A forum answer is not the firm's answer.
6. Check the sizing rule against the loss limits. A bot that fits the automation terms can still end an attempt on a loss limit, as the challenge guide explains.
7. Re-read when anything changes. A new account type, a rule update or a change to your bot starts the process again.

What a checklist does not do

It does not make a bot compliant, and passing the checklist does not mean a bot will pass a challenge. The terms may be vague, the firm may read a clause differently from you, and a restriction can be added after you start. If a clause is unclear, the firm's written answer is worth more than any guess.