GRID TRADING — THE RANGE MACHINE AND ITS FATAL FLAW.
EducationStrategy~18 min readUpdated 30 September 2026
The short answer
Grid trading places buy and sell orders at fixed price intervals around a central price. Each time price moves one grid step, an order fills. In a range, the grid harvests profit from every oscillation. In a trend, the grid accumulates losing positions that compound into a single catastrophic drawdown. The grid does not have a stop-loss. That is the point and the problem.
Why this lesson follows the breakout lesson
Lesson 39 taught you how to trade a range that resolves. Lesson 40 teaches you how to trade a range that holds. These are opposite trades for opposite market states. A breakout trader needs the range to break. A grid trader needs the range to hold. If you run a grid during a breakout, you lose. If you trade a breakout during a range, you get chopped.
This lesson is not a recommendation. It is an honest assessment of a strategy that is heavily marketed and rarely explained. By the end, you will know exactly when a grid works, exactly when it fails, and whether it belongs in your toolkit.
T
Written by the Trade To The Top team|Reviewed 30 September 2026
Grid methodology cross-checked against The Mathematics of Money Management (Vince), the martingale and anti-martingale literature in Fortune's Formula (Poundstone), the grid-trading documentation published by major grid-bot vendors, and the risk-of-ruin mathematics used in casino and trading system analysis. The distinction between fixed-size grids and martingale grids verified against standard position-sizing literature.
Grid trading is the most marketed and least understood strategy in retail trading. Vendors sell it as passive income. Forum posts call it account suicide. Both are wrong. A grid is a tool with a specific use case and a specific failure mode. This lesson gives you both, with the math.
Key takeaways
A grid is a set of orders at fixed intervals. Buy below, sell above, repeat. No prediction, no analysis, no discretion.
The grid profits from oscillation. Every time price moves one grid step and returns, you book a profit. The more oscillations, the more profit.
The grid has no stop-loss. That is the design. A stop would defeat the purpose. But it means the grid has unlimited theoretical risk.
Fixed-size grids are not martingale. A martingale doubles after a loss. A fixed-size grid uses the same lot on every level.
The danger is not the grid size. It is the trend. A grid in a range is safe. A grid in a trend accumulates losses that can exceed the account.
Drawdown is the only metric that matters. A grid with a 1% monthly return and a 40% drawdown is not a strategy. It is a time bomb.
The grid works best in ranging pairs. AUD/NZD, EUR/CHF, and other low-volatility crosses are the classic examples.
The grid fails in trending pairs. GBP/JPY, gold, and indices can move 500 pips in a day. A grid on those instruments is a countdown to ruin.
You must cap the grid. A grid with no maximum position count is not a strategy. It is a bet against a black swan.
Know the regime. Grids work in ranges. Ranges end. When the range ends, the grid must be turned off.
A grid is a set of orders placed at fixed price intervals around a central price. Below the centre, buy orders. Above the centre, sell orders. Every time price moves one grid step, an order fills. Every time it moves back, a profit is taken.
That is the whole system. No analysis. No prediction. No discretion. The grid does not care which way price goes. It only cares that price moves back and forth.
This is why grids are popular in ranging markets and dangerous in trending ones. A range is a machine that oscillates. A trend is a machine that runs. The grid is designed for the first and destroyed by the second.
THE GRID CONCEPT · ORDERS AT FIXED INTERVALS
Conceptual diagram · buy orders below the mid, sell orders above, profit taken at each grid step
Buy orders below the mid, sell orders above. The grid does not predict direction. It harvests oscillation.
Common mistake
Thinking a grid is a strategy. A grid is an execution method, not an edge. The edge must come from somewhere else — either a genuine range regime, or a filter that turns the grid off when the range ends. A grid with no filter is a countdown to a blow-up. The countdown may take months. It always ends the same way.
How a grid works — the mechanics
Here is the mechanical sequence of a grid trade:
The grid mechanics
01
Set the centre. Pick a mid price — either the current price or a level you expect price to oscillate around.
02
Set the spacing. Choose the grid interval — for example, 20 pips. Every order goes 20 pips apart.
03
Place buy orders below. Buy at mid − 20, mid − 40, mid − 60, and so on.
04
Place sell orders above. Sell at mid + 20, mid + 40, mid + 60, and so on.
05
Take profit one step away. When a buy fills at mid − 20, set a take-profit at mid. When a sell fills at mid + 20, set a take-profit at mid.
06
Re-place the order. When the take-profit hits, the buy order at that level is placed again. The grid runs continuously.
That is it. The grid is a loop. Price oscillates, orders fill, profits are taken, orders are replaced. In a range, this loop prints money. In a trend, it prints losses.
GRID IN A RANGE · THE LOOP THAT WORKS
Twenty-eight candles · price oscillates between grid levels, each oscillation books a profit
Price oscillates between grid levels. Every round trip books a profit. The range is the machine.
The math — why it works in ranges
The grid profit formula is simple. If your grid spacing is S pips, your lot size per level is L, and the number of completed round trips is N, your profit is:
Formula — grid profit
Profit = S × L × N
Where S is the grid spacing in pips, L is the lot size per level (converted to pips value), and N is the number of completed round trips.
Example:
Spacing = 20 pips
Lot size per level = 0.10 (about $1 per pip)
Round trips per month = 40
Profit = 20 × $1 × 40 = $800 per month.
THE RANGE MACHINE PRINTS MONEY.
Now the counter. What happens when price moves 200 pips in one direction without returning? Every grid level below the mid fills. You now hold 10 losing positions at 0.10 lots each, all moving against you. The floating loss at 200 pips is:
The counter — floating loss in a trend
Floating loss = (Sum of open position losses)
10 open buys at 0.10 lots, spaced 20 pips apart.
Price has moved 200 pips below the mid.
The first buy (at mid − 20) is 180 pips underwater.
The last buy (at mid − 200) is at breakeven.
Average loss per position = 90 pips.
90 pips × $1 per pip × 10 positions = $900 floating loss.
And price is still moving.$900 FLOATING LOSS. AND COUNTING.
That is the grid's fatal flaw in one calculation. In a range, the grid earns $800 a month. In a trend, the same grid loses $900 in a week. The grid does not have a stop-loss because a stop would defeat the purpose. That is also why the grid has no floor.
A grid is a short volatility position. It works until volatility expands.
The terminology — grid, martingale, and DCA
Three terms get used interchangeably. They are not the same.
Grid trading places buy and sell orders at fixed intervals. Each level uses the same lot size. The grid does not double. It does not average down with increasing size. It is a mechanical oscillation harvester.
Martingale doubles the position size after every loss. It is a bet that the next trade will win. In a grid context, a martingale grid doubles the lot size at each lower level. This is a completely different risk profile and it is almost always a path to ruin.
Dollar-cost averaging (DCA) adds to a losing position at fixed intervals with the same lot size. This is the same as a fixed-size grid. Some vendors call it DCA to avoid the word "grid."
The distinction matters because the risks are different. A fixed-size grid has linear risk. A martingale grid has exponential risk. Do not confuse them.
The fatal flaw — trending markets
Grids do not fail in ranges. They fail in trends. And trends always come. The range that a grid harvests is not a permanent state. It is a pause before the next trend. When the pause ends, the grid fills every level in one direction and accumulates a floating loss it cannot recover from.
GRID IN A TREND · THE ACCOUNT KILLER
Twenty-eight candles · price breaks the range and runs · every grid level fills against the move
Price breaks the range and trends. Every grid level below the mid fills. The floating loss compounds.
The grid vendor's lie
Grid vendors show you the equity curve from a ranging period. They do not show you the equity curve from a trending period. The ranging period looks like a smooth, upward-sloping line. The trending period looks like a cliff. Both are the same grid. The only difference is the market state. If a vendor cannot show you the drawdown, they are not showing you the strategy. They are showing you a marketing chart.
Fixed-size vs martingale grids
Not all grids are equal. The lot size progression determines the risk profile.
Grid type
Risk
Survivability
Fixed-size grid
Linear — each level uses the same lot size.
Survivable with a cap on total positions. Drawdown is proportional to the move.
Martingale grid
Exponential — each level doubles the previous lot size.
Almost never survivable. One trend wipes the account.
Anti-martingale grid
Decreasing — each level uses a smaller lot size.
Safest grid type. Drawdown is capped. Profit per range is lower.
Fibonacci grid
Sub-exponential — lot sizes follow a Fibonacci progression.
Less dangerous than martingale but still compounds rapidly. Not recommended.
If you are going to run a grid, run a fixed-size grid with a hard cap on total open positions. No doubling. No Fibonacci. No "recovery" mode. The grid either works at a fixed size or it does not work at all.
A martingale grid is not a strategy. It is a bet that the market will never trend again.
Risk management for grids
If you are going to run a grid, these are the non-negotiable rules.
Grid risk management
01
Cap the total position count. Never let the grid open more than a fixed number of positions. Eight levels is a reasonable maximum. Sixteen is reckless. Unlimited is not a strategy.
02
Use fixed lot sizes. No doubling. No martingale. The same lot on every level. If the grid needs to double to work, it does not work.
03
Size the grid so the maximum drawdown is survivable. If the cap is 8 levels and the grid spacing is 20 pips, the maximum adverse move is 160 pips. Size each level so 160 pips of adverse movement is no more than 10% of the account.
04
Set a hard stop on the whole grid. If price moves beyond the last grid level by more than one grid step, close everything. The grid is dead. Take the loss.
05
Only run the grid in a confirmed range. If the higher timeframe is trending, do not run a grid. Grids and trends are opposite trades.
06
Stop the grid when the range ends. If price breaks the range boundary, the grid is no longer valid. Close it and wait for a new range to form.
07
Never run a grid on a trending pair. GBP/JPY, gold, and indices can move 500 pips in a day. A grid on those instruments is a countdown to ruin.
Safe grid
Levels8 max
Lot sizeFixed
Max drawdown10% of account
StopHard stop at level 9
RegimeRange only
SURVIVABLEDrawdown is capped
Unsafe grid
LevelsUnlimited
Lot sizeDoubling
Max drawdownUnlimited
StopNone
RegimeAny
RUINOne trend ends the account
Regime detection — when to run a grid
The grid's success depends entirely on one thing: is the market ranging or trending? If you cannot answer that question, you cannot run a grid. Here is the decision table.
Market state
Grid verdict
What to do
Confirmed range
Run the grid.
The grid's ideal environment. Cap the levels. Size for survivability.
Trending
Do not run a grid.
Trade the trend with a breakout or trend-following strategy instead.
Range after a long trend
Wait.
The range is likely a pause before the trend resumes. Do not grid it.
Range before a news event
Close the grid.
News can break the range in seconds. The grid will fill every level before you can react.
Range with expanding volatility
Reduce or close.
Expanding volatility is the grid's enemy. The range is about to break.
Range with contracting volatility
Run the grid.
Contraction is the grid's friend. The range is likely to hold.
Volatility contraction is the grid trader's best friend. When a range is getting tighter — the highs are getting lower and the lows are getting higher — the grid works best. When a range is getting wider, the grid is about to fail.
GRID ENVIRONMENTS · CONTRACTION VS EXPANSION
Two panels · left: contracting range (grid works) · right: expanding range (grid fails)
Left: contracting range, grid works. Right: expanding range, grid fails. The direction of volatility is the signal.
When this fails
When this fails
The range breaks. Every range ends. When the range breaks, the grid fills every level in one direction. The floating loss compounds. If the grid has no hard stop, the account is at risk of total loss. The grid must be turned off when the range ends.
A news event hits. A single NFP or central bank decision can move price 200 pips in seconds. The grid fills every level before you can react. The floating loss is immediate and large. Close the grid before major news events.
The instrument is volatile. Grids on GBP/JPY, gold, and indices are a countdown to ruin. These instruments can move 500 pips in a day. A grid cannot survive that. Only run grids on low-volatility pairs.
The grid is uncapped. A grid with unlimited levels has unlimited theoretical risk. There is no price level at which the grid stops opening new positions. If the market trends far enough, the grid will blow the account. Always cap the grid.
The grid is a martingale. A martingale grid doubles the position size at each level. The risk is exponential. One trend wipes the account. Never run a martingale grid.
If you remember nothing else: a grid is a short volatility position. It profits when volatility falls and loses when volatility rises. The range is the machine. The trend is the killer.
In one box
A grid is orders at fixed intervals. Buy below, sell above, harvest oscillation.
The grid works in ranges. Every round trip books a profit.
The grid fails in trends. Every level fills against the move. Floating loss compounds.
Fixed-size grids are not martingale. Do not confuse them. The risks are different.
Always cap the grid. Eight levels maximum. Unlimited levels is not a strategy.
Always set a hard stop. If price moves beyond the last level, close the grid.
Only run grids in confirmed ranges. If the higher timeframe is trending, do not grid.
Close the grid before news. A single release can break the range in seconds.
Never run a grid on volatile instruments. GBP/JPY, gold, indices — these are not grid instruments.
Volatility contraction is the grid's friend. Volatility expansion is the grid's enemy.
See it in practice. Our free trading journal lets you tag trades by strategy — grid, breakout, trend — so you can see which approach produces the best risk-adjusted return over time. Log the regime, the strategy, the drawdown, and the outcome. The data tells you which approach your edge lives in.
5 questions · immediate feedback · retake any time
Question 01 of 05
What is the fundamental flaw of grid trading?
Correct: B. The grid has no stop-loss by design. In a range, that is fine. In a trend, every level fills against the move and the floating loss compounds. The grid has no floor.
Question 02 of 05
What is the difference between a fixed-size grid and a martingale grid?
Correct: C. A fixed-size grid uses the same lot size on every level. A martingale grid doubles the lot size at each level. The martingale risk is exponential and almost never survivable.
Question 03 of 05
When does a grid work best?
Correct: A. The grid works in a confirmed range, especially when volatility is contracting. When volatility expands, the range is about to break and the grid is about to fail.
Question 04 of 05
What is the single most important risk management rule for a grid?
Correct: D. The grid must have a maximum number of levels and a hard stop. Without a cap, the grid has unlimited risk. With a cap, the maximum drawdown is known and survivable.
Question 05 of 05
What should you do when a range breaks while running a grid?
Correct: B. When the range breaks, the grid is no longer valid. Close the grid and take the loss. Do not add levels. Do not average down. Do not wait. The grid is dead.