Grid Trading Strategy: Mechanics, Costs and Account Limits

A grid trading strategy places pending orders at fixed price intervals above and below the market, so that movement in either direction fills something. Most explanations stop at the pattern and call it profit without prediction. What they leave out is that whether a grid can run at all is decided before the first order is placed: by the account mode, by a minimum order distance the broker enforces, and in one large jurisdiction by regulation.

This page covers the mechanics briefly, then spends its time on the four constraints that the pattern itself hides: the account mode a two-sided grid requires, the spacing floor, what an open grid costs to hold overnight, and how fast the open loss grows when price trends instead of oscillating.

What a Grid Trading Strategy Actually Does

A grid starts from a reference price and lays orders at a fixed spacing: every 25 pips, every 50 pips, whatever the plan specifies. Each order carries the same size, and each has a take-profit one grid step away. When price crosses a level, that order fills; when it travels one step further in the profitable direction, the leg closes for a fixed gain.

The appeal is that no single order needs a forecast. Price wandering up and down through the grid keeps filling and closing legs, and each round trip banks one step of profit. The structure replaces timing decisions with geometry.

The cost of that replacement is the subject of the rest of this page. A grid is not a neutral machine: it is a standing commitment to keep buying or selling at every level price reaches, and the commitment is largest exactly when the market moves furthest against it.

The Two Grid Types: With the Trend and Against It

Grids come in two orientations, and they fail in opposite ways.

A with-trend grid places buy stops above the market and sell stops below it. Movement in one direction fills legs that point the same way, so a sustained trend stacks winners. Its bad day is a range: price triggers a level, reverses, triggers the opposite side, and the grid buys high and sells low repeatedly. Each whipsaw pays the spread and books a step-sized loss.

An against-trend grid does the reverse: buy limits below the market, sell limits above it. In a range it harvests every oscillation, which is why demonstration charts almost always show one. Its bad day is a trend: each new level fills another position pointing the wrong way, and nothing closes. The losing legs accumulate until the trend ends or the margin does.

Many published grid systems run both orientations simultaneously, holding long and short legs in the same symbol at the same time. That version has a precondition most descriptions never mention, and it is the next section.

Why a Grid Needs a Hedging Account to Exist

A bidirectional grid assumes the account can hold a buy position and a sell position in the same symbol simultaneously. Whether it can is a property of the account itself, set when the account is opened.

MetaQuotes documents two position-accounting modes for MT5. In netting mode, only one position can exist per symbol: a sell fill while a long position is open reduces or reverses that position. In hedging mode, every fill opens its own position, so multiple positions in one symbol can coexist. The full mechanics, including how margin is charged on opposing positions, are on our page about netting and hedging account modes.

The grid consequence is blunt. On a netting account, the moment the first sell-side grid order fills against an open long leg, it does not open the short leg the grid design expects: it closes part of the long. The grid dismantles itself with every opposite fill. One-directional grids survive netting; the classic two-sided grid does not run on it at all.

Checking the mode takes one glance at the terminal before any order is placed, and it decides whether half the grid designs published anywhere are even available to the account.

The Spacing Floor: Stops Level and Rejected Orders

Grid design guides discuss spacing as a free parameter: tighter grids for quiet pairs, wider for volatile ones. On a live account the broker has a say first.

MT5 carries a per-symbol property called the stops level: the minimum distance in points from the current price at which stop and pending orders may sit. An order placed inside that distance is rejected. The general rule, why it also affects closing, and where to read the number for a specific account are on our stops level page.

For a grid the property acts as a density cap. A strategy calling for 10-point spacing on a symbol whose stops level is 30 points cannot be placed as designed: every level within the floor is refused as it comes due. The grid that actually runs is the one the broker permits, which may be three times coarser than the one that was backtested. Wider spacing means fewer, larger steps, a different profit rhythm, and a different worst case.

The number varies by broker, by symbol and sometimes by market conditions, so a grid ported between two brokers is not the same grid until the spacing has been rechecked.

What a Grid Costs to Hold: Swap on Every Open Leg

A grid in profit closes its legs quickly. A grid under pressure does the opposite: it warehouses open positions, sometimes for weeks. Warehousing is not free.

Each open position is charged or credited swap for every night it is held, at per-symbol long and short rates the platform publishes, with one day of the week carrying a triple charge to cover the weekend rollover. MetaQuotes documents these as per-symbol properties; the values themselves are set by each broker and change.

The grid multiplies whatever the rate is. Eight open legs accrue eight swap entries per night. If the warehoused side of the grid is the side that pays rather than receives, the floating loss deepens a little every rollover even while price stands still. A backtest that ignores financing, or a demonstration that runs for a week, never surfaces this line item; a statement from a month of trending market makes it hard to miss.

Direction matters as much as size. The same pair can be cheap to warehouse short and expensive to warehouse long, so the identical grid design has different economics depending on which side the trend leaves open.

The Exposure Arithmetic Nobody Prints

Grid presentations show the profit per oscillation. Almost none work through what the account holds after a one-way move, so the arithmetic follows, with every input stated as an assumption.

Assume an against-trend buy grid: levels every 25 pips below the start, 0.10 lots per level, on a pair where 0.10 lots is worth one dollar per pip. Price falls 200 pips with no meaningful bounce. Eight legs have filled, so the account is long 0.80 lots. The first leg is 175 pips under water, the second 150, and so on down to the newest at zero. The floating loss is the sum: 700 dollars.

Now let the fall continue to 300 pips. Twelve legs are open and the same sum runs 275 down to zero: 1,650 dollars. The move grew by half; the loss grew by 136 per cent. That is the shape of the curve — with spacing s and N filled legs, the floating loss is s times N times (N minus 1) over 2, which grows with the square of the distance, not in proportion to it.

Margin compounds the squeeze. Every filled leg adds its own requirement — assume, for the arithmetic only, 3.33 per cent of notional — so used margin climbs in a straight line while equity falls along that square-law curve. The lines cross at the account’s margin call and stop out levels, and the platform then closes legs at the worst prices the grid has seen. Running an against-trend grid pre-commits the account to funding that worst case.

Grid and Martingale Are Not the Same Bet

Grids get grouped with recovery systems, and the association is half right. A martingale strategy responds to a loss by doubling size, so its exposure grows exponentially with each step of a losing sequence. A grid keeps size constant per level, so exposure grows linearly with distance — it is the floating loss, not the position size, that follows the square law shown above.

The distinction matters for failure speed. A martingale can reach an unplaceable order size within a handful of steps. A grid fails more slowly and more quietly: no single leg looks reckless, and the account dies of accumulation rather than escalation. Hybrid systems that widen grid size at deeper levels import the martingale failure mode into the grid one, and inherit both.

Where a Grid Cannot Run

Three environments rule out the classic bidirectional grid regardless of settings.

US retail forex accounts. NFA Compliance Rule 2-43 requires a forex dealer member to offset positions in a customer account on a first-in, first-out basis rather than carry offsetting positions. Long and short legs in the same pair therefore cannot coexist; the account behaves like an enforced netting account. The account-mode consequences are worked through on the netting and hedging page.

Netting accounts anywhere. As covered above, opposite fills reduce the existing position, so only one-directional grids are available.

Symbols whose stops level exceeds the designed spacing. The grid can technically be placed at wider spacing, but it is then a different strategy from the one that was tested, and it should be treated as untested.

Who This Is Not For

A grid is unsuitable for an account that cannot fund its worst case: the arithmetic section above is the sizing tool, run before the first order, not after the drawdown. It is equally unsuitable for anyone who saw it marketed as passive income — the pattern automates order placement, not risk. Whoever cannot state what the grid loses at 300 pips adverse should not place one at 25.

Frequently Asked Questions

Is grid trading profitable?

Sometimes, in ranging markets, and the gains arrive as many small steps. The structure pays for those steps by holding a floating loss that grows with the square of any sustained move against it, plus swap on every warehoused leg. Whether the steps outweigh the eventual trend depends on spacing, size, financing and the market that actually arrives, and no setting removes that trade-off. No return can be promised.

What is a grid trading strategy?

A method that places pending orders at fixed price intervals above or below the market, each with a take-profit one interval away. Price movement fills the orders mechanically, so no individual order carries a forecast. The design replaces entry timing with a standing commitment to trade at every level price reaches.

Does grid trading work on an MT5 netting account?

Only in one direction. Netting mode allows a single position per symbol, so a sell fill while a long position is open reduces that position instead of opening a short one. A grid that needs long and short legs at the same time requires hedging mode, which is a property of the account, not a platform setting to toggle.

Why does a broker reject some grid orders?

Most commonly because they sit inside the stops level, the minimum distance from the current price at which pending orders may be placed on that symbol. Levels closer than the floor are refused, which caps how dense a grid can be. The value differs between brokers and symbols, so a design that places cleanly on one account can be rejected on another.

Can a US retail account run a bidirectional grid?

No. NFA Compliance Rule 2-43 requires offsetting positions in a customer account to be closed first-in, first-out rather than held simultaneously, so long and short legs in the same pair cannot coexist. One-directional grids remain possible within the usual margin limits.