Trade Copier: How They Work and What Silently Breaks Them

A trade copier mirrors trades from one account to others automatically. One account opens a position, and within a fraction of a second the same position appears on every account connected to it, resized for each one. That is the whole idea, and it is simple enough that the interesting part is entirely in what goes wrong.

This page covers how a copier works underneath, local versus cloud setups, the four failures that silently cost money, how to size positions across accounts of different sizes, and where copiers are genuinely useful rather than just clever.

• A copier multiplies an outcome, not an edge. Losses copy exactly as well as wins.
• Local copiers are faster and fragile; cloud copiers are resilient and slower.
• Symbol naming is the most common cause of a copier that "does nothing".
• Balance-proportional sizing is the only sane default across different account sizes.
• On prop firm accounts, copying is restricted almost everywhere. Read the clause first.

How a trade copier works

Three parts. A master where trades originate, a transport that carries the instruction, and a slave that reproduces it.

On MetaTrader the usual build is an expert advisor on each terminal. The master EA watches for open, modify and close events and publishes them. The slave EA subscribes and executes the equivalent action on its own account. Transport is a shared file, a local socket, or an HTTP endpoint when the accounts sit on different machines.

The slave has to do considerably more than "open the same trade":

• Resolve the master's symbol to whatever this broker calls it, suffixes included.
• Convert the master's lot size into one appropriate to this account, then round to the broker's lot step and clamp to its minimum and maximum.
• Translate stop and target prices, which will not match because the two brokers quote differently.
• Respect this broker's minimum stop distance, which may be wider than the master's.
• Reconcile after a disconnect so a reconnecting terminal does not duplicate open positions.

That last one separates a copier you can leave running from a script that works until it does not. Without reconciliation a copier will eventually double a position, and a doubled position is how an account blows a risk limit in a single move.

The four kinds of trade copier

"Trade copier" covers four fairly different things, and picking the wrong category is a more expensive mistake than picking the wrong product inside a category.

Terminal-side copier

An expert advisor on each MetaTrader terminal, talking over a shared file or a local socket. It is the cheapest and the fastest, and it is what most retail copiers are. The catch is that every terminal has to be running somewhere, which in practice means a VPS you now have to keep alive.

Server-side copier

The same EA model, but the transport is an endpoint rather than a file, so the accounts do not need to share a machine. You trade some latency for the ability to put accounts at different brokers, in different countries, without one reboot taking everything down.

Broker-side allocation

MAM and PAMM accounts, where the broker performs the allocation and no copying software exists at all. Nothing to install and nothing to break, but every account has to be at that broker, and opening one normally involves a money-management agreement. That is a different regulatory relationship than running software on your own accounts, and it is worth understanding before you sign one.

API or bridge copier

For platforms that expose an API instead of an EA model. The logic is identical, the failure modes shift: you inherit the platform's rate limits and its idea of what an order is, and you are responsible for retry behaviour that an EA would have handled locally.

Local versus cloud

The honest rule: if your strategy needs single-digit millisecond copying to work, the edge is too thin to survive a live spread anyway. For almost everything else, cloud is the better trade because uptime beats speed.

Four failures that cost money

1. Symbol suffixes

The master trades EURUSD, the slave broker calls it EURUSD.raw or EURUSDm. Without suffix handling the slave silently does nothing while the terminal looks perfectly healthy: EA attached, smiling face, no errors. You discover it when you compare statements a week later. This is the number one cause of "the copier isn't working".

2. Lot rounding

Proportional sizing produces fractional lots that must be rounded to the broker's step. On a small slave account that rounding is a large share of the intended position. Rounding up quietly increases risk beyond the master's; rounding down to zero drops the trade entirely. Neither is visible from the outside.

3. Partial closes and modifications

Copying opens and closes is the easy half. A master that scales out, moves a stop to break even, or trails a stop produces a stream of modify events. A copier that ignores them leaves the slave in a materially different trade, and by the time it matters the positions have already diverged.

4. Latency during the moments that count

The slave always fills after the master. On a swing system this is irrelevant. On a system targeting a few pips it can be most of the edge, and the gap is widest exactly when it hurts: at the open and around news, when the queue is longest. Measure master-to-slave delay under load, not at rest.

Sizing across different account sizes

Whichever you pick, every slave needs its own drawdown guard. Inheriting the master's risk logic is not enough when the slave carries limits the master does not have.

What to look for in trade copier software

Feature lists are close to useless here because every product claims the same six things. These are the questions that actually separate them, and most of them are answered by testing rather than by reading a page.

• What happens when the computed lot rounds to zero. Silently skipping the trade and opening the broker minimum are both defensible. Not telling you which one it did is not.
• Does it map symbols, or just match strings. Suffixes are the easy case. Metals and indices are where naming genuinely diverges between brokers.
• Does it reconcile, and how. Ask what happens if the slave terminal is offline for ten minutes while the master opens and closes a position. The honest answers are "it is skipped" or "it is opened late"; the dangerous answer is that nobody knows.
• Are modifications propagated. Stop moves, partial closes and trailing stops, not just opens and closes.
• Can a slave have its own risk cap that overrides the master rather than inheriting it.
• Are rejections visible. When a slave broker refuses an order for margin or stop distance, you need to find out then, not at the end of the month.
• What it does on a slave margin call. A copier that keeps sending into an account that cannot fill is generating noise at the worst possible moment.

Test all of it on demo accounts with deliberately mismatched conditions: different brokers, different suffixes, a slave a tenth the size of the master. A copier that survives that is a copier you can leave alone.

Copiers on futures platforms

Most of what is written about copiers assumes MetaTrader, which is unhelpful if your accounts are futures evaluations. Those normally sit on NinjaTrader or Tradovate, or on a Rithmic feed behind whichever front end the firm provides, and none of them use the EA model. Copying there is done by platform add-ons or through the API, so the four failure modes above still apply but they surface in different places.

The bigger difference is not technical. Futures evaluations typically use a trailing drawdown that follows the account's high water mark, rather than a fixed floor set at the starting balance. A copied position that runs into profit and then gives it back can breach the limit on a slave while the master, sitting on a different balance and therefore a different trailing level, is nowhere near it. Two accounts running identical trades can end the week with one funded and one closed.

If you are copying across futures evaluations, model the trailing level per account before you connect anything. The drawdown calculator is the quick version of that check.

What a copier costs

Three pricing shapes are common: a one-off licence per terminal, a monthly fee per connected slave, and a bundled subscription that includes hosting. Sticker price is rarely the real number.

The costs people forget are the VPS to keep the terminals alive, and the time spent restarting things after a broker updates its terminal. If a copier saves you twenty minutes a day and costs an hour a week to keep running, it is not saving anything.

Free copiers exist and some are fine for a single pair of accounts on one machine. What they usually lack is exactly what matters when you stop watching: reconciliation after a disconnect, visible rejections, and per-slave risk caps. That is an acceptable trade while you are testing and a poor one once personal capital is connected.

Setting one up without breaking anything

1. Read the copy trading clause for every account involved, and for prop accounts get the answer from support in writing.
2. Connect one slave, on demo, at a different broker than the master so suffix handling is actually exercised.
3. Run a full trade lifecycle through it: open, partial close, stop to break even, trail, close. Compare both statements line by line.
4. Kill the slave terminal mid-trade and bring it back. Confirm what reconciliation did, and that it did not duplicate the position.
5. Set the slave's own drawdown cap below the account's real limit, so the copier stops before the broker or the firm does.
6. Only then add the second slave, and check sizing again, because rounding behaves differently at different account sizes.

Where a copier is actually worth it

• One strategy, several of your own accounts at different brokers, so you are not repeating entries by hand.
• Separating strategies from execution. Run the logic once on a master and let accounts subscribe, rather than maintaining the same EA in five places.
• Managing accounts for others where each needs its own sizing and its own risk cap.

And where it is not: a copier does not diversify anything. Four accounts on one master is the same position four times, so the losing streak arrives on all four in the same week. The correlation is exactly one, and calling it diversification is the most expensive mistake in this area.

On prop firm accounts, read the rules first

Copying from a third party or a signal service is prohibited essentially everywhere. Copying between accounts you own varies: some firms permit it, others treat identical fills across accounts as a shared strategy. Firms also cap total capital per strategy across all your accounts, which is aimed precisely at running one master across several evaluations.

Simultaneous identical fills are trivially visible in the firm's own data, and the assessment usually happens at payout review, after the work is done. The detail is in trade copiers on prop firm accounts.

What we run

Ours exists because we needed a copier that understood prop firm rules rather than one that simply mirrored fills: per-slave lot multipliers, balance-proportional sizing, symbol suffix resolution, reconciliation after disconnect, and drawdown limits enforced on the slave rather than assumed from the master.

If what you actually want is the automation without operating a copier fleet, the JPTC Algo runs on your own account at your own broker and we host it, so there is nothing to install and no VPS.

Related: how latency affects copied accounts and setting one up across multiple accounts.