The prop firm vertical has a new section — Copy Trading Risk. It addresses one of the most common ways a prop account is lost, and one that rarely involves any bad intent: correlation bans.
Firms close accounts for copy trading, confiscate profits and sometimes add the account to the lists they exchange with each other. The typical case is a trader who bought an Expert Advisor on a marketplace and had no way of knowing the same EA runs on dozens or hundreds of other accounts at the same firm. The firm sees a cluster of accounts producing identical trades and handles the cluster as a whole. The second most common case is one EA running across several of a trader's own challenges at the same firm — which most firms already define as copying.
Neither situation can be established from the inside. The firm sees all of its accounts; the trader sees only his own. The new section closes that gap by computing the same metric a firm computes, and showing it before a decision is made rather than after.
The section is split into four tools. Other traders shows the share of matched trades between your account and other accounts at the same firm, with a risk band — the other party's identity is never disclosed in any form. Your own accounts detects indistinguishable accounts among yours, including EAs that were never declared anywhere, by finding the same magic number actually trading on two accounts. Strategy crowding shows how many accounts at your firm run the same EA and how mass-market that EA is across the platform. Comparison reproduces the firm's algorithm across four axes — entry time, entry price, SL/TP and exit — for two of your accounts, and returns a sample of matched trades timestamp against timestamp.
Comparison runs at three strictness levels, from the firm's literal algorithm (1 second, 1 point) to a profile wide enough to catch deliberate jitter (30 seconds, 50 points). Matching is one-to-one, SL and TP are compared by distance from entry rather than absolute price, and systematic quote differences between brokers are compensated for.
A separate emphasis is on the honesty of the numbers. A raw match share proves little when both accounts trade the same session on the same instrument, so the section computes a randomness correction: the second account's timeline is cyclically shifted by whole days, the comparison is re-run a hundred times, and the excess over chance is reported. If the observed value does not stand out from the random spread, the risk band is lowered and the result is stated plainly — the accounts simply trade in the same sessions. Strategy crowding is likewise paired with a confidence figure, so that factory-default magic numbers shared by unrelated EAs are not reported as copies.
Cross-user calculations run only on data from users who have enabled anonymised data sharing. The setting is symmetric — disabling it removes your data from other users' calculations and stops cross-user figures from being shown to you — and analysis of your own accounts works regardless. Only anonymised trading data participates: symbol, direction, volume, time and price, algorithm identifier and trade comment. Personal data, account numbers, broker names and financial results do not.
The section measures and displays; it does not block, prevent or intervene. It does not produce a probability of a ban, since firms' thresholds are not public and no calibrated model exists. Only accounts of FX Monitor users are compared, so an absence of matches means an absence within that population rather than a clean result overall — coverage will improve as more trading accounts are connected. Copy Trading Risk is available now to users with the prop firm vertical enabled, for both MT4 and MT5.
