Grid and recovery expert advisors

Grid and recovery expert advisors add to a position, or increase its size, when the market moves against them, then close everything once a small bounce puts the group of trades in profit. The results look calm, with long runs of winners, while the losses wait in open trades until one move turns them into a single large loss.

What grid and recovery EAs do

The names overlap, and many EAs combine them:

• Grid. Opens a new trade at fixed distances, such as every $5 on gold, as price moves against the first one.
• Averaging down. Adds to a losing position so its average entry moves towards the market, then closes the whole group, called a basket, at a small net profit.
• Martingale. Raises the lot size at each new level or after each loss, often doubling it.
• Recovery. After a loss, opens a larger trade, sometimes in the opposite direction, sized to win the loss back plus a profit.

All of them respond to being wrong by adding exposure rather than taking the loss.

Why the curve looks smooth

Markets retrace far more often than they run without pause, so a system that needs only a small bounce closes most baskets in profit. The balance, which counts only closed trades, climbs in steady steps. Meanwhile an open basket's loss sits in equity, so a balance graph can look flawless while equity is deep underwater. The pattern is a high win rate, small wins and an occasional loss that undoes months of them.

How exposure grows: a worked example

Take two buy grids on gold, each adding a level every $5 below the first entry. Grid A adds 0.10 lots per level. Grid B starts at 0.01 lots and doubles each time. A $1 move is worth $100 per 100-ounce lot, and costs are left out.

The figures are basket totals as each level opens. Grid A's loss grows faster than the distance, because every open trade loses on every step. Grid B starts tiny and overtakes it: at level 8 it loses $255 for each further $1, against $80 for Grid A. In return it needs only about $4.84 of bounce to break even, against $17.50, which is why doubling grids win so often.

A $35 fall is within gold's range on a busy day. If both grids stop adding at 8 levels and gold falls another $15, Grid A is down $2,600 and Grid B $5,060: 26% and about 51% of a $10,000 account.

Account size, levels and stops

• Maximum levels cap the number of trades, not the loss. Both grids above kept losing after their last level.
• A basket stop closes the basket at a set loss or distance, turning an open-ended loss into a known one.
• A hard equity stop closes everything the EA holds when equity falls by a set amount. Gaps can still fill it at a worse price.
• Account size should come from the worst case: the lots at the maximum level, their margin at your leverage and the loss at the basket stop, with equity well above your broker's stop-out level.

The margin calculator and pip value calculator give the figures for your settings.

Why backtests understate the tail risk

A backtest is one path through history, and a grid's risk lies in the paths that didn't happen:

• The breaking move may be missing. Ranging years can show no deep basket.
• Optimisation favours survivors. The top settings are often the ones that just missed the worst move.
• Coarse modelling hides spikes. Open-price testing and fixed spreads miss intrabar runs and spread widening.
• Balance gets the attention. The equity drawdown and the largest lot matter more.

Test on real ticks with variable spreads, and start runs just before the strongest trends you can find. Our backtesting guide covers the settings.

The risk, stated plainly

A grid or recovery EA exchanges many small wins for a rare large loss. Over a long enough period, a move that reaches the basket stop is likely, so the question is whether the profit made beforehand outweighs it. Without a basket or equity stop, the broker's stop-out sets the loss, which usually means most of the account.

Adding money to rescue a deep basket raises the eventual loss, and grids on correlated pairs such as EURUSD and GBPUSD can go deep together. On prop firm accounts, equity-based limits count floating losses; see our prop firm trading page.

Checks before running one

1. Read every input. Note the starting lot, multiplier, step, maximum levels and basket target and stop.
2. Work out the worst case. Find the lots, margin and loss at the maximum level with the margin calculator.
3. Set both stops. Put each at a loss you accept.
4. Read the equity line. Backtest on real ticks and record the equity drawdown and largest lot. See how to backtest an EA.
5. Check the recovery. The drawdown calculator shows the gain needed to make back a loss at the basket stop.
6. Start on demo. Confirm your broker's maximum lot and stop-out level.

Our guide to grid and martingale EAs shows how to spot one from its trade history.