Running one strategy across many prop accounts: the copy-trading and correlation rule that voids them all

The single most common scaling plan in prop trading is also the one that gets the most accounts banned at once: take a strategy that passed one evaluation and copy its trades, tick-for-tick, across five or ten more accounts. The maths looks irresistible, one edge, ten times the funded capital. The problem is that every major firm has a rule, and increasingly an automated detector, aimed squarely at exactly this behaviour. Breach it and you do not lose one account; you lose the entire cluster, plus any payout already in flight, in a single compliance action.

Three different questions people conflate

The phrase "can I run my strategy on multiple accounts" hides three separate questions with three different answers. Sorting them out is the whole game.

1. Can I run the same logic on several accounts? Generally yes, owning the edge and re-applying it is not prohibited. Firms cannot and do not ban ideas.
2. Can I copy-trade between my own accounts at the same firm? Almost universally no. Master-to-slave copiers across accounts you hold at one firm are an explicit violation at FTMO, FundingPips, MyForexFunds and most others.
3. Can I run the same strategy across competing firms simultaneously? Usually contractually allowed, but if the fills are timestamp-identical, some firms treat cross-firm correlation as a manipulation flag at payout review. Allowed is not the same as safe.

Why copy trading inside one firm is the fast path to a mass ban

When you mirror one master account onto several accounts at the same firm, every account fills the same symbol, same direction, same size-ratio, within the same few milliseconds. To the firm's risk engine this is indistinguishable from a single oversized position split to dodge per-account exposure limits, which is precisely the abuse the rule exists to stop. The detector does not need to read your strategy; the fill-timestamp correlation alone is the evidence. Because all the accounts belong to one verified identity, the enforcement action closes them together. Traders who scaled to ten accounts this way routinely report losing all ten in one email.

A subtler trap: even without a formal copier, manually placing the same trade across accounts in quick succession produces near-identical timestamps and trips the same heuristic. "I clicked fast" is not a defence the review process recognises.

What each supported firm says

Copy-trading and correlation policy across the supported firm set, as captured in the most recent terms audit. "Within firm" = copying between accounts you hold at that firm; "across firms" = the same strategy run at a different firm at the same time.

The pattern is uniform: copying within a single firm is banned everywhere; running across firms is contractually fine everywhere; and almost every firm reserves a correlation review at payout. The risk is therefore concentrated not in whether you may scale, but in whether your fills look mechanically copied.

How to scale without tripping the detectors

Scaling is legitimate when each account makes its own decisions rather than mirroring a master. The goal is statistical independence of fills, same edge, different fingerprints. Concretely:

1. Never run a master-to-slave copier across accounts at the same firm. This is the single highest-risk action and is banned everywhere in the supported set.
2. If running the same strategy across competing firms, deliberately desynchronise: stagger entries by a randomised delay, vary position sizing per account, and let each instance evaluate its own signal rather than relaying one signal to all.
3. Vary the parameter set slightly per account (entry filter, stop distance) so fills diverge naturally instead of clustering on identical timestamps.
4. Keep per-account risk independent, do not size account B as a fixed multiple of account A, which reproduces the correlation the detector looks for.
5. Pre-flight each account's configuration through the firm-rule-aware backtester separately, so each one passes its own firm's rules rather than inheriting a single shared config.

Practical workflow on Glitch Executor

Glitch Executor arms each connected account as its own agent with its own signal evaluation, sizing, and execution clock, there is no master-to-slave relay, so fills are independent by construction rather than mirrored. When you arm the same strategy on more than one account, the position-sizing calculator sets risk per account against that account's own firm rules instead of as a multiple of a master, and the per-arming controls let you stagger and vary execution so the accounts do not produce a timestamp-identical fill cluster. That keeps a multi-account scale-up on the right side of the correlation rule that the copier-EA approach walks straight into.