An Expert Advisor does the same thing every time. Its owner does not. Between a strategy's record and an owner's result sits a series of decisions the owner makes, almost all of them unplanned, almost all of them made under stress, and almost all of them costly.
The patterns
Switching off during a drawdown, which locks in the loss and misses the recovery that the record says usually follows. Restarting after the recovery, which puts the account back in just as the next drawdown begins. Raising the risk setting after a good run, which means the next drawdown arrives at double the size. Lowering it after a bad run, which halves the recovery. Closing winning trades by hand before their target, because the profit is there and might go away. Closing losing trades before their stop, or worse, moving the stop. Adding a second EA to make up a loss. Checking the account several times a day, which produces all of the above.
Each of these is an intervention that changes the EA's result, and there is no evidence that owners' interventions improve it. The pages on drawdown and losing streaks explain why the moments that feel like emergencies are usually ordinary.
Why it happens
A loss is felt more strongly than an equal gain, so drawdowns feel like disasters and recoveries feel like relief rather than profit. An open loss is a decision not yet made, and the pressure to make it grows with every look at the screen. And an EA gives no reassurance; it does not explain itself the way a human manager would, so an owner in a drawdown has only the equity figure and their own nerves.
An owner runs an EA with a verified record showing a maximum drawdown of 17 per cent. In month four the account is down 12 per cent. They lower the risk setting by half. The recovery over the next two months is 8 per cent at half size, where it would have been 16. The account is now 5 per cent below where the untouched EA would have been, and the owner, seeing the slow recovery, switches it off. The EA was fine. The record was fine. The owner cost themselves 12 per cent.
The answer is a written plan, on its own page, with the rule for stopping decided in advance and a review cadence measured in weeks. Run the EA on demo long enough to see it lose and recover, so that the first live drawdown is not the first drawdown you have watched. Do not look at the account daily. And treat the urge to intervene as information about you, not about the EA.