Category
Strategy
Difficulty
Intermediate
Used in
Strategy designEA evaluationRisk management

Grid trading

A strategy that places orders at fixed price intervals above and below a reference price, profiting from oscillation rather than from predicting direction.

also: Grid system, Price grid, Grid EA

Updated · Reviewed

In plain English

Instead of guessing which way price goes, the EA lays a ladder of orders at even distances and lets price walk into them. Every step price retraces closes a rung in profit — and every step it keeps going adds another open position.

Why it matters

Grid EAs are the largest single family sold to retail traders, and they are the family whose equity curve lies most convincingly. A grid can print months of small, even gains while carrying an unrealised loss that never appears in the closed-trade record until the day it does.

  • The win rate is structurally high — most rungs do close in profit — so win rate is the least informative number on a grid EA's page.
  • The real exposure is the number of rungs the account can absorb before margin runs out, which is a function of grid step, lot size and leverage, not of the strategy's cleverness.
  • A grid with no per-position stop loss converts a directional loss into an open-ended one; the position is only closed by a recovery, a basket target, or a stop out.

How the strategy works

  1. 1Pick a reference price and a fixed step, e.g. every 20 pips.
  2. 2Place orders at each step above and below the reference — buy limits below, sell limits above for a mean-reverting grid.
  3. 3Each order that fills becomes an open position with its own small take profit, usually one step away.
  4. 4Price oscillating inside the grid closes rungs repeatedly, each for a small profit.
  5. 5Price trending out of the grid leaves the far side accumulating open positions that move further into loss with every step.

Market conditions that suit it

  • Ranging markets with repeated retracement inside a stable band.
  • Instruments whose spread is small relative to the grid step, so each rung's target survives the round-trip cost.
  • Sessions with mean-reverting character — the quiet hours of a range-bound pair rather than a news release.

Main advantages

  • No directional forecast is required; the edge comes from oscillation, which is more common than trend.
  • Trade frequency is high and results look smooth while price stays inside the band.
  • The logic is simple to specify and to audit — step, lot, target, maximum rungs.

Main risks

  • A sustained trend fills one side of the grid repeatedly and every fill adds unrealised loss.
  • Margin, not the strategy, decides the outcome once several rungs are open: a stop out closes everything at the worst price.
  • Combined with martingale lot scaling, the loss on the last rungs dwarfs every profit taken before it.
  • Backtests over a range-heavy window make a grid look far better than a window containing a strong trend, so period selection alone can manufacture a good record.

Typical EA behaviour

  • Multiple positions open on one symbol at the same time, in the same direction.
  • Individual positions carry small take profits and often no stop loss at all.
  • A basket target closes every open rung together once the combined profit reaches a threshold.
  • Equity dips much deeper than the closed-trade balance while a grid is open — the two curves separate.

How to spot it in an EA

  • Inputs named Step, Distance, GridSize, MaxTrades or Levels.
  • A lot multiplier above 1.0, which means the grid is also a martingale.
  • Reported maximum equity drawdown far larger than the balance drawdown.
  • A very high win rate paired with an average loss several times the average win.

Example

A grid with a 20-pip step and five levels below the reference price. Price falls through three of them, so three positions are open at once and the account is carrying the distance between each entry and the current price.

Filled rungs are solid; unfilled rungs are the exposure still ahead of the account.
Grid step
20 pips
Levels below the reference
5
The account must survive all five being filled, not just the first.
Lot per level
0.01
Fixed here. A multiplier above 1.0 makes this a martingale grid.

Calculation 3 filled levels × 20 pips average distance × 0.01 lots ≈ open loss of 60 pip-lots before the fourth level fills

Result 3 open positions, all in the same direction, all losing

Illustrative figures — not the record of any listed EA.

How it is used

Judge a grid EA by what happens when the range breaks, not by the equity curve while it holds.

  • Compute the worst case first: maximum rungs × lot size × step distance is the loss the account must be able to carry.
  • Ask for equity drawdown, never balance drawdown — the balance curve of a grid hides open positions by construction.
  • Check the backtest window contains at least one strong trend in the traded instrument. A range-only window proves nothing about the failure mode.
  • Look for a hard maximum on open positions. Without one, position count is bounded only by the account balance.

A grid can be a sound strategy with a fixed lot, a bounded rung count and a stop that closes the whole basket. It is the unbounded version that produces the equity curves that end at zero.

Common mistakes

Reading the win rate as an edge

A grid closes most rungs in profit by design. A 90% win rate says the grid was inside its range, not that the strategy is profitable over a full cycle.

Confusing grid with martingale

A grid is about where orders sit; a martingale is about how the size grows. They are frequently combined, and the combination is what turns a bounded loss into an account-ending one — but a fixed-lot grid is a different risk profile.

Assuming price always comes back

Mean reversion is a tendency, not a law. Currencies trend for months; the recovery a grid depends on may arrive after the margin call.

Judging a grid on a range-heavy backtest

Choose the window and you choose the result. A grid tested only on a sideways period is being asked the one question it always answers well.

In depth

What fourteen grid-free records say about the number a grid hides

Across the 14 EAs published hereFigure
Records where equity drawdown exceeds balance14 of 14
Median gap between the two0.61 points
Widest gap (Cairn)11.10 points
Cairn, balance then equity drawdown−14.52% then −25.62%
Longest losing streak, median6
Longest losing streak, worst (Cairn)14

Read the Cairn row twice. It wins 64.04% of 4,725 closed trades and still lost fourteen straight, because a streak follows how many times you roll, not how often you win. A grid answers that streak by adding rungs, and a doubling ladder covering the median six needs 64 times the first lot; covering Cairn’s fourteen needs 16,384.

The gap column is the other half. These fourteen hold one position per symbol at a fixed lot, with a stop attached — the design furthest from a grid — and equity still dipped below balance in every record. Cairn’s balance drawdown of 14.52% describes an account that was really 25.62% down. On a strategy that holds losing rungs open by design, that distance is the product, not a rounding error.

MT5 prints both. The Strategy Tester’s Results tab lists Balance Drawdown Maximal and Equity Drawdown Maximal on separate lines, and a grid listing tends to quote the first. Read them as a pair: their ratio is the closest thing to a grid detector a backtest report contains, and our methodology publishes both for every listing.

Frequently asked questions

Is grid trading profitable?
It can be, in ranging conditions with a fixed lot size and a bounded number of levels. What decides the outcome is not the entry logic but whether the account can carry every level the grid is allowed to open when price trends against it.
What is the difference between grid trading and martingale?
A grid defines where orders are placed — at fixed price intervals. A martingale defines how position size changes — increasing after losses. Many EAs use both, and it is the size progression, not the grid spacing, that produces the catastrophic tail.
Why does a grid EA show a small drawdown but a large loss?
Because the small figure is usually balance drawdown, which counts only closed trades. Open grid legs sit in floating loss and never enter the balance curve until they close, so the equity drawdown is the number that describes the risk.