Grid Trading EAs: How They Work and How They Blow Up

Grid EAs are the great seducers of the EA market. Months of small, steady gains; an equity curve that barely wobbles; glowing reviews from users who haven't hit the bad month yet. Understanding exactly how they work is the best inoculation available.

How a grid EA works

A grid system places orders at fixed price intervals - say, every 20 pips - above and/or below the current price. When price oscillates, the grid harvests each swing. When price trends hard in one direction, the EA accumulates a stack of losing positions, each one deeper underwater than the last, waiting for a reversal to close the basket at breakeven or a small profit.

In a range, this looks like magic. The catch is structural: the strategy's losses are open, not closed. The equity curve shows realised profits while unrealised drawdown balloons out of sight. You're not watching performance; you're watching the visible half of it.

How they blow up

• The trend that doesn't reverse. Every market eventually makes a move bigger than the grid budgeted for. The basket grows, margin erodes, and either the EA's "equity protection" closes everything at a devastating loss or the broker's margin call does it less politely.
• Gaps. Weekend and news gaps skip right over grid levels and stop levels alike. Index and gold grids are especially exposed.
• Compounding the stack. Many grids increase lot size as the basket deepens - sliding into martingale territory - which converts a bad drawdown into a terminal one.

The pattern shows up constantly in why most EAs fail live: years of small wins, then one week that removes them all, plus the account.

Why the backtests mislead

A grid backtested over a rangy period looks flawless - and vendors get to choose the period. Even long backtests mislead, because the metric that matters is peak open drawdown, which many reports underplay or omit. If you don't know your drawdown definitions cold, a grid seller is counting on it. And a grid tuned to one instrument's historical range is curve-fitting in its purest form.

Questions that expose a grid EA

1. What was the maximum open drawdown in the backtest - not just realised drawdown?
2. Does lot size increase as the basket grows?
3. What happens, in exact terms, when the trend never comes back? Is there a hard stop on the basket?
4. What margin does the deepest historical basket require on my account size?
5. Has it survived a genuinely trending year on this instrument, out-of-sample?

Build instead of buy: our biased take

Bias on the table: we make Nebula, a no-code builder that breeds strategies with a genetic algorithm and validates them with walk-forward, out-of-sample and blind forward testing. Nebula's approach is the opposite philosophy to grids: every position carries its own stop, and risk is defined per trade rather than deferred into an open basket - the discipline behind the 1% risk rule. We don't claim its strategies outperform anything; we claim you can see exactly what they risk, which is precisely what a grid hides.

The bottom line

Grid EAs aren't a scam by definition - they're a risk structure most buyers don't understand until it's demonstrated on their account. If you run one, size it as though the blow-up is coming, because historically, it does. And before buying any EA, grid or otherwise, work through our evaluation checklist.

Trading carries a high level of risk and automated systems do not eliminate it. Nothing here is financial advice; past performance does not predict future results.