Connecting a live account to a copier takes about ninety seconds. Undoing what a misconfigured one does to that account takes considerably longer.
This is the list worth running first. It is deliberately boring, in order, and every item is something that has cost somebody real money.
Before you connect anything
- Run the master for two weeks on its own. If you cannot describe the strategy's expectancy and worst drawdown from its own history, copying it only multiplies an unknown.
- Open a demo follower at the target broker. Same instruments, same account type, same leverage. This is the environment every remaining check runs in.
- Verify the instrument mapping by notional value. Matching lot sizes across brokers do not imply matching exposure — contract sizes differ, especially on metals and indices.
Sizing and risk
- Choose the sizing mode deliberately, and know its failure case. Risk-percent needs a stop loss to compute against; decide what should happen to a trade that arrives without one before one does.
- Set a maximum lot size as a backstop. Not as your primary control — as the thing that catches a mis-keyed stop distance or a bad rounding result.
- Check the broker's minimum lot step against your typical calculated size. On a small account, a 0.01 minimum against a 0.013 calculation is a permanent 23% overshoot on every single trade.
- Set an equity stop on the follower. A hard floor at which copying halts, independent of any per-trade risk setting. This is the one control that limits a bad sequence rather than a bad trade.
Behaviour under stress
- Test a partial close. Open on the master, close a third, then close the rest. Confirm the follower's position ends at exactly zero — orphaned remainders are common and easy to miss.
- Test a rejected stop. Set a stop tighter than the follower broker allows and watch what the copier does with the resulting unprotected position. If the answer is "nothing, silently", that is a finding.
- Test a disconnect. Kill the connection with a position open, then restore it. The question that matters: on reconnect, does it resume, re-open, or duplicate? Find out on demo.
Every one of these is reversible on a demo follower and permanent on a live one. That asymmetry is the entire argument for the list.
The first live week
Go live at a fraction of intended size — a quarter is reasonable — and hold it there for a full week of normal trading, including at least one scheduled news event.
Watch three things, and nothing else:
- Fill differences between master and follower. A consistent bias in one direction is a routing or latency problem, not variance.
- Any trade that did not copy. One unexplained skip is worth an hour of investigation before it becomes a pattern.
- Aggregate exposure across all accounts, summed by currency and direction rather than by ticket. This is almost always larger than expected the first time it is measured.
Scale up only when a week has passed with nothing on that list unexplained.
The tedious version of this process takes about two weeks. The exciting version takes ninety seconds and occasionally costs an account. Copy trading rewards the tedious version more reliably than almost anything else in this business.
