When to Switch an EA Off: Telling Losing Apart From Breaking

Knowing when to stop an EA is harder than knowing when to start one, and it is the decision that actually determines your result. Stop too early and you sell the strategy at its worst moment. Stop too late and you fund a broken system with real money. The way out is to separate two things that feel identical while they are happening: losing, and breaking.

A strategy that loses while doing exactly what it said it would do is working. Losing streaks are part of the design; every edge has them, and the published drawdown is a promise that this size of loss was anticipated. Turning off an EA that is inside its normal range is not risk management. It is buying high and selling low, one system at a time.

These are the things worth acting on, and none of them is the equity number. The trade count departs from the description — twice the trades, or none for a period the EA should have traded. Holding times change character: a strategy that held for hours now holds for days. Positions open outside the stated session or symbol. The largest loss exceeds anything in the published record, especially if the stop appears to have been absent. Any of these means the EA is no longer the thing you evaluated, and the reason matters more than the money.

Before funding, write two numbers: the drawdown at which you reduce size, and the drawdown at which you stop entirely. Both should be below the level at which you would start making decisions out of fear. A line chosen in advance is a decision; a line chosen during the drawdown is a reaction, and reactions are reliably worse.

When results diverge from expectation, the EA is only one of the candidates. Did the broker change the spread or the symbol specification? Did the VPS restart and leave the terminal without autotrading enabled? Did a news filter fail because the calendar feed was unavailable? Is the account in a different leverage tier than when you sized the position? Most "the EA stopped working" reports resolve to one of these, and switching the EA off hides the real fault instead of fixing it.

Halving the lot is almost always better than switching off, because it keeps the record continuous. You keep learning whether the strategy recovers, you cut the damage if it does not, and you avoid the worst outcome — stopping at the bottom and watching the recovery from outside. Off is a decision you can only make once; smaller is a decision you can reverse.

If you cannot explain what the EA is doing, stop it. Not because it is losing, but because an unexplained position is an unbounded risk. You can always restart after you understand it. You cannot un-lose money to a mechanism you never diagnosed.

Our three best-selling EAs: GOLD NEURON (AI) https://www.mql5.com/en/market/product/187329 · ATLAS PORTFOLIO https://www.mql5.com/en/market/product/182751