Your copier maps ES to the wrong contract, lags four seconds, and keeps firing after you flatten the master. You find all three out with real funded money on the line, at the worst possible moment. That is the default outcome for traders who skip the pre-flight, and it is avoidable. A copier is infrastructure, and infrastructure gets tested before it carries load. What follows is a sequence for serious prop-firm traders: fire one contract, watch what actually happens on every follower, and only then scale.
Why you test first
A trade copier is not a strategy. It is plumbing that sits between one master account and N follower accounts, and it carries correlated risk by design. Every account fires the same signal at the same instant. That is the point. It also means a single mistake never fails quietly on one account. It multiplies across all of them at once.
Think about what that does to the arithmetic of error. A wrong contract, a fat-fingered multiplier, or a copier that keeps sending orders after you close the master is not a one-times problem. It is an N-times problem. Five accounts under one master turn one config mistake into five simultaneous breaches, and prop-firm drawdown rules do not wait for you to notice.
A copier multiplies one error across every linked account at the same moment. That is why a pre-flight test is mandatory, not optional.
So the case for testing fits in one line. Because a copier amplifies whatever you feed it, you verify the feed before you connect real size. If you are still choosing hardware, the mechanics of running this cleanly are covered in the VPS setup for futures copy trading, and the wider sync and slippage picture is in multi-account copy-trading sync and slippage risk.
The minimum-lot test
The baseline discipline is the minimum-lot test, and it is boring on purpose. Fire one minimum-size trade on the master, a single micro or the smallest tradable unit, then confirm five things in order.
- It lands on every follower.
- It lands at the expected scaled size on each account, matching the configured multiplier.
- The symbol matched correctly on each broker (right contract, right suffix).
- The close propagates to every follower when you close the master.
- Zero rejections in the copier log.
There is no partial pass. If any one of the five fails, you are not cleared to scale. Four of five is not a near-miss you can round up. It is a failed test, because the one account that misbehaved is exactly the one that will surprise you with real money on it. Fix the failure, then re-run the whole sequence from the top.
Four followers copying correctly and one that did not is a failed test, not a near-miss. There is no partial pass.
Why is this discipline worth the two minutes it costs? At one contract, the discovery cost of any bug is trivial, and every bug that matters shows up here first. The full step-by-step of wiring the master and followers before you reach this point lives in how to copy trade on MT5 and cTrader.
Symbol and multiplier checks
Two categories of silent error live here, and both are cheaper to catch than to survive. The first is symbol mapping. The same instrument is spelled differently across venues: ES on one broker, ESU5 on another, a suffixed variant on a third, MES for the micros. A mismatch can silently route to the wrong contract or reject outright, so you verify the mapping per broker, suffixes included, against the actual fill and not just the config screen.
The second is sizing math. Confirm each per-account multiplier against the real fill size, and set two hard limits while you are there.
- A maximum-position cap per follower, so a runaway master or a fat finger cannot push any single account past its risk limit.
- An explicit point-value mapping whenever you copy futures to CFD. Futures and CFD contracts do not carry the same notional per point, so a raw 1:1 copy from a futures master to a CFD follower mis-sizes risk unless you map the point value on purpose.
Here is the arithmetic that makes this concrete. Say you run five funded accounts on one master at multipliers of 1x, 2x, 2x, 3x, and 3x, trading ES. The CME E-mini S&P 500 carries a point value of $50 per full point (one tick is 0.25 points, or $12.50), correct at time of writing, though you should confirm current specs with the exchange before sizing since they can change. Fire one contract on the master and take a 4-point adverse move as the test.
| Account | Multiplier | Contracts | Risk on a 4-pt move |
|---|---|---|---|
| A | 1x | 1 | 4 x $50 x 1 = $200 |
| B | 2x | 2 | 4 x $50 x 2 = $400 |
| C | 2x | 2 | 4 x $50 x 2 = $400 |
| D | 3x | 3 | 4 x $50 x 3 = $600 |
| E | 3x | 3 | 4 x $50 x 3 = $600 |
| Total | 11 | $2,200 |
Now break it two ways. First, a point-value error: one follower routes ES to a broker CFD such as a "US500" contract worth (illustratively, an example figure, not a real quoted spec) $1 per point instead of $50. Account D then shows 4 x $1 x 3 = $12 on the same move. You believe you are copied at 3x. You are effectively 50x under-hedged and not really copied at all. Second, a multiplier fat finger: Account D gets typed as 30x instead of 3x. That is 30 contracts, 4 x $50 x 30 = $6,000 on one account, and the fleet total blows from $2,200 to $200 + $400 + $400 + $6,000 + $600 = $7,600. Account D almost certainly breaches its trailing drawdown on the first losing trade.
Both errors surface at one contract: Account D fills 30, or shows nonsense per-tick P&L, before any size is on. Worst-case discovery cost is $200 across the fleet instead of $7,600 or a blown account.
The latency and fill test
Latency is the number vendors love to quote and traders love to benchmark, so measure it yourself rather than trusting the marketing page. Time the leader-to-follower delta during a liquid session, then confirm the measured figure matches the vendor's claim. If a vendor claims X and you measure 5X, the claim is not describing your setup, full stop.
For an order-of-magnitude sense of what "server-side and fast" looks like, a copier such as Thor operates at roughly 17 ms leader-to-follower. Treat that as a reference point, not a number to expect from every setup. Real-world delta varies with venue, session, and load, so your own measurement is the only one that counts.
Slippage and copier lag hide on quiet markets and appear at news and session open, which is exactly when a few hundred ms becomes a meaningful price gap between master and followers.
The critical part is when you measure. A calm tape flatters every copier. Force the test during a fast move, a news print or the session open, because that is where slippage and lag actually bite, and where a delta of a couple hundred milliseconds turns into a real price difference across your accounts. If the copier holds its claimed delta while the tape is moving, you have learned something worth knowing.
The flatten and failover test
Opening a trade is the easy path, and it is the only path most vendor demos show you. The states that kill accounts are on the way out and after a break in the connection, so you force those on purpose.
Start with the flatten. Close the master and confirm every follower flattens. A copier that closes four of five accounts is a liability wearing the costume of a tool, and an asymmetric close (master flat, one follower still open) is precisely the exposure you did not sign up for. Then attack the connection deliberately. Kill the master link and confirm the copier fails safe: it must not keep pushing stale orders, and it must not leave followers in an inconsistent state. Finally, bring the connection back and confirm a clean reconnect, state re-syncing without duplicate or phantom orders.
Here is the read a generic guide will miss. Most blown copier accounts are not blown by latency. Latency is easy to quote, so everyone benchmarks it, but the real account-killers cluster in symbol mapping, multiplier and point-value math, and flatten propagation. A copier that is 17 ms fast and maps ES to the wrong contract will drain you faster than a 400 ms copier that is configured correctly. That framing is operator experience, not a measured statistic, but it is why the traders who survive test the boring stuff first. Does the close reach all five? Does one broker's rejection get flagged, or swallowed? Chase correctness first, then chase speed.
Edge cases that break copiers
Happy-path testing certifies nothing. Force each of these states and watch how the copier reacts, because every one of them has broken a copier somewhere.
- Partial fills on the master. Does the follower copy the filled quantity or the intended quantity?
- A one-broker rejection. An order is accepted at four venues and rejected at the fifth. Does the copier flag it, or swallow it silently?
- A disconnect mid-trade. Open on master, connection drops before the follower confirms.
- A manual master flatten. You close the master outside the copier UI rather than with a copier close command. Does the close still propagate?
- Weekend and session gaps. Position state carried across a market close and re-open.
- A follower already at its cap. A new copy arrives at an account already at its contract limit. Is it rejected cleanly, or does it break the config?
The news gate belongs in this list too, and it is not something you leave on default. A proper setup disables copying around restricted economic-event windows, and you test that it actually stands down the correct configs. In Thor specifically, when a connection is restricted the disable filter matches either the follower ID or the leader ID on that connection, so any config touching the restricted account is stood down and every position on it is flattened. The takeaway is broker-agnostic. Confirm your news gate disables the right configs and flattens, on both the leader side and the follower side, before you rely on it. And remember that firm-specific news windows, contract limits, and allowed instruments vary and change, so verify the current rule with your firm rather than trusting a number you read once.
A go-live checklist
Turn all of the above into a decision rule. Do not carry real funded size until every box is checked.
- Min-lot test passed: lands on every follower, correct scaled size, symbol matched per broker, close propagated, zero rejections in the log.
- Symbol maps verified per broker, including suffixes.
- Multipliers set, and a hard max-position cap set per follower.
- Point-value mapping verified for any futures-to-CFD leg.
- Latency measured under load (a liquid session and a fast move) and matching the vendor claim.
- Flatten confirmed: master flat means every follower flat.
- Failover confirmed: kill the master connection, it fails safe, then reconnects clean with no dupes or phantoms.
- News gate configured and confirmed to disable the correct configs.
One failed checkbox means not cleared to scale. Fix it, re-run the min-lot test, then proceed. When everything passes, go live small and scale up one step at a time, never test-lot straight to full size.
Now the honest tradeoff, because no pre-flight test is bulletproof. You cannot summon a real CME news spike or a real broker outage on demand, so testing narrows the failure surface without eliminating it. That is exactly why the correct posture is to go live small and scale, not test-then-full-size. A fast server-side copier like Thor (about 17 ms, flat pricing, a free trial you can run this whole sequence inside) narrows the surface further by removing your local PC and home internet as points of failure, but it does not remove the need for a per-setup min-lot test. Every new configuration earns its own test before it carries real money.
And a copier is not always the answer. If you need the absolute lowest raw latency and can put a machine in the same datacenter as the broker, a colocated local copier beats a server-side one on raw milliseconds. Server-side wins on reliability and hands-off uptime, not on the last few ticks of speed. If you trade a single account, a copier adds correlated infrastructure risk for zero diversification benefit, so trade the account directly. If your accounts run incompatible rules (different news restrictions, contract limits, or allowed instruments), forcing them under one master can trip one firm's rules while satisfying another's, so reconcile the rulesets first. And a copier never fixes a losing strategy. It faithfully replicates whatever the master does at N times the size, which means an unprofitable master just loses money faster. Compare architectures with clear eyes in the rundown of futures trade copiers, then test the one you pick before it carries a single funded dollar.
Frequently asked questions
What is the minimum-lot test for a trade copier?
The minimum-lot test means firing one minimum-size trade (a single micro or the smallest tradable unit) on the master account before you run any real size. You then confirm five things in order: the trade lands on every follower, at the correct scaled size per the multiplier, with the symbol matched correctly on each broker, the close propagates to every follower, and the log shows zero rejections. If any one of the five fails, you are not cleared to scale. There is no partial pass, so you fix the failure and re-run the whole sequence.
Why does a trade copier carry more risk than trading one account?
A copier carries correlated risk because every follower account fires the same signal at the same instant. A single config error, such as a wrong contract or a mis-typed multiplier, does not fail on one account. It multiplies across all of them simultaneously. Five accounts under one master turn one mistake into five simultaneous breaches. That amplification is the core reason a pre-flight test is mandatory rather than optional.
What copier latency should I expect between master and follower?
There is no single right number, so you measure the leader-to-follower delta yourself during a liquid session and confirm it matches the vendor's claim. As an order-of-magnitude benchmark for a fast server-based architecture, a copier such as Thor operates at roughly 17 ms, but real-world delta varies with venue, session, and load. Test during a fast move, not only a calm tape, because slippage and lag show up under volatility. That is exactly when a few hundred milliseconds becomes a meaningful price difference across accounts.
Why does symbol mapping matter when copying trades across brokers?
The same instrument is spelled differently across venues, for example ES, ESU5, a broker-suffixed variant, or MES for the micros. A mismatch can silently route to the wrong contract or reject the order outright. You verify the mapping per broker, suffixes included, against the actual fill and not just the config screen. Symbol mapping errors are one of the most common account-killers precisely because they can fail silently while everything looks connected.
How do I test a copier's flatten and failover behavior?
Close the master and confirm every follower flattens; a copier that closes four of five accounts is a liability, since an asymmetric close leaves one account exposed. Then deliberately kill the master connection and confirm the copier fails safe, meaning it stops sending stale orders and does not leave followers in an inconsistent state. Finally, restore the connection and confirm a clean reconnect with state re-syncing and no duplicate or phantom orders. These out-and-recovery states never appear in a happy-path vendor demo, which is why you have to force them.
Do I need to check point values when copying futures to CFD?
Yes. Futures and CFD contracts do not carry the same notional per point, so a raw 1:1 copy from a futures master to a CFD follower mis-sizes risk unless you map the point value explicitly. For example, an ES point is worth $50 on the CME E-mini S&P 500, while a broker CFD equivalent might be worth far less per point, which would leave that follower badly under-hedged even though it appears copied. Verify the point-value mapping for any futures-to-CFD leg during your test, and confirm current contract specs before sizing since exchanges can change them.
When is a trade copier not the right tool?
A copier is the wrong tool if you trade only one account, because it adds correlated infrastructure risk for zero diversification benefit. It is also a poor fit when your accounts run incompatible rules, such as different news restrictions or contract limits, since forcing them under one master can trip one firm's rules while satisfying another's. If you need the absolute lowest raw latency and can colocate a machine in the broker's datacenter, a local copier beats a server-side one on raw milliseconds. And a copier never fixes a losing strategy, because it replicates the master faithfully at N times the size.
What is a safe way to go live after testing?
Treat the go-live checklist as a decision rule: min-lot test passed, symbol maps verified per broker, multipliers and a hard max-position cap set per follower, point-value mapping verified for any futures-to-CFD leg, latency measured under load, flatten confirmed, failover confirmed, and the news gate configured and tested. One failed box means you are not cleared to scale, so you fix it and re-run the min-lot test. When everything passes, go live small and scale up one step at a time. No pre-flight test reproduces every live condition, so you narrow the failure surface and never jump from test-lot straight to full size.