How to read your prop firm's automation rules

Almost every version of "can I automate this" reduces to three separate questions. They are independent: a firm can say yes to one and no to another, and people get into trouble by getting one answer and assuming it covers all three.

Can you copy between your own accounts?

This is about one order becoming several. Most firms have a position on copying within their own platform across accounts you own, and a separate position on copying across firms. The distinction matters because the risk the firm is managing is different in each case. Within one firm, the concern is usually correlated exposure across accounts they are all underwriting. Across firms, the concern is usually about who is actually trading, and whether one person is effectively multiplying a single evaluation.

Ask it in that shape: my own accounts, at this firm, at this stage. An answer about copying somebody else's signals is an answer to a different question.

Can orders be placed by software rather than by you?

This is where the most consequential distinction lives, and it is not yes or no. Many firms permit what is usually called semi-automation, meaning software may place orders while you are present and monitoring, and prohibit fully unattended operation, meaning software trading an account nobody is watching. Some state a presence requirement explicitly. Some address bots and algorithms as a category without addressing webhooks at all.

If your firm draws that line, it changes what you should build. Routing an alert to a broker while you sit in front of the chart is a different activity from leaving a strategy running overnight, even when the software is identical. Find out which side of that line your firm puts you on before you build for either.

What is the consistency rule, and at which stage?

Consistency rules cap how much of your total profit may come from your best day, and they are the rule automation interacts with most awkwardly. A strategy that has one outstanding session and then behaves normally can pass every risk rule and still fail a consistency check, because the check is about the shape of your returns rather than their size.

These rules also commonly apply at only one stage. A firm with no consistency requirement during the evaluation may apply one before a payout, which is exactly the point at which discovering it is most expensive.

The automation rule is rarely the one that ends an account. The risk rules you already know are, and automation changes how you hit them. Four interactions are worth thinking through before you switch anything on.

Copying makes your accounts correlated on purpose

This is worth saying plainly because it is the part that is easy to miss. The point of a copier is that every account does the same thing. That is also what it means to have no diversification whatsoever across those accounts: a bad session is a bad session everywhere at once, and a rule breach caused by the strategy is a breach in every account simultaneously rather than in one.

That is a perfectly reasonable thing to want. It is just worth deciding it rather than discovering it, and worth checking whether your firm has anything to say about correlated accounts before you assume the copier question is only about the copier.

Assume you have read your firm's rules and they permit what you want to do. There is still a gap between permitted and survivable, and it is mostly operational.

1. Get the answer in writing, from the firm, about your programme and your stage. A dated support reply in your inbox is worth more than any third-party summary on the internet, including anything on this site.
2. Start on one account. A copier makes a configuration mistake happen everywhere at once, so make the mistakes while the blast radius is one evaluation.
3. Know your stop before your entry. If the protective legs go on with the entry rather than after it, there is no window where the position exists unprotected.
4. Keep a way to intervene by hand. Cancel a working order, move a stop, close one account or all of them, from wherever you actually are when it matters.
5. Keep a record you did not write yourself. When a firm asks what happened, a per-order timeline with broker-side timestamps answers the question. A screenshot does not.
6. Re-read the rules on a schedule. They change, and the version you agreed to is not necessarily the version being enforced at payout.

The last one deserves its own sentence, because it is the one people skip. A rule you read once is a rule you read once.

What to actually put in the support ticket

Vague questions get vague answers, and a vague answer is worth nothing at a payout review. Ask about your programme, your stage and your specific setup, one question per line, and ask for the answer in the ticket rather than on a call. These four cover almost everything that goes wrong later.

1. May I use a trade copier to mirror orders across the accounts I hold with you, at my current stage, and does the answer change when I am funded?
2. May orders be submitted to my account by software I have connected, rather than by me clicking, and do I have to be present and monitoring while it runs?
3. Is there a consistency requirement on my account, what is the threshold, and at which stage is it applied?
4. Does your trailing drawdown follow my closed balance or my unrealised equity, and is there anything I must do before the session closes?

Keep the reply. Note the date on it. If the answer contradicts something published, the published page is what gets enforced by whoever reviews your account next, so it is worth asking them to reconcile the two while you have their attention rather than discovering the gap yourself.

Two mistakes that are not about the rules at all

The first is testing on the account that matters. Every configuration mistake you are going to make, you will make in the first week, and a copier makes each one happen on every account at once. A single evaluation account is a cheap place to make them.

The second is not knowing how to stop. People spend a great deal of care on the entry logic and almost none on the exit from the whole arrangement, and then discover during a bad session that flattening several accounts means logging into several platforms in sequence. Rehearse that before you need it, and time yourself doing it.

• If your concern is one signal reaching several accounts, the trade copier page shows how the fan-out and the single way out actually work.
• If your concern is what happens when the software stops, read what happens to your open position when your automation goes down.