Grid Trading Strategy Settings

The five settings

A grid bot places buy orders below the current price and sell orders above it, at fixed steps. When a buy fills, it places a sell one step higher. When that sells, it earns the gap between the two, minus fees. The whole machine is described by five choices. This page covers how they interact. For how a single grid cycle works, see grid trading strategy, and for the bot as a whole, grid trading bot.

• Range: the lowest and highest price the grid covers. Outside it, the bot stops trading.
• Number of grids: how many price steps the range is cut into.
• Spacing type: arithmetic spacing makes every step the same size in price. Geometric spacing makes every step the same percentage, so steps are larger at higher prices.
• Stop: a price at which the bot sells what it holds and shuts down. It may be below the range, above it or absent.
• Investment: the money the bot may use. It is divided between the grids.

Number of grids against spacing

The spacing is the range divided by the number of grids. A made-up example: the range is 80 to 120, the price is about 100, the investment is 1,000, and the round trip fee (buy and sell together) is 0.2 percent. That is 0.1 percent a side, a made-up rate. Rules and rates vary by venue, so use your own.

Calculated by us on 29 Oct 2026, exact for the stated inputs. Going from 10 to 80 grids cuts the net gain per cycle from 3.80 percent to 0.30 percent. The bot completes cycles far more often, but each one earns very little, and a round trip cost of 0.5 percent instead of 0.2 would turn the 80-grid line to zero. Fewer grids earn more per cycle and complete rarely. More grids look busier, and the page on grid bot fees and costs shows where busy becomes losing.

Range against spacing

Now keep 20 grids and the same 1,000 and move the range. Each order is 50. The same round trip fee applies.

A narrow range of 95 to 105 pays little per cycle (0.30 percent) and the price leaves it after a 5 percent fall. A wide range of 60 to 140 pays 3.80 percent a cycle and tolerates a 40 percent fall, but price must travel four times as far to complete one cycle, so it completes fewer. There is no free setting: the same 20 grids buy either cover or frequency. Nothing in the arithmetic says which is right, because that depends on how far the price will actually move, which nobody knows.

The stop level

Use the range 80 to 120 with 10 grids, so the spacing is 4. The price starts at 100 and falls. The bot has spent 500 on five buy orders of 100 each, at 96, 92, 88, 84 and 80. That buys 5.71 units at an average of 87.64. These are the losses on those holdings if the bot exits at different prices, with fees ignored:

• Stop at 78, just under the range: loss 54.97.
• Stop at 70: loss 100.62.
• Stop at 60: loss 157.67.
• No stop, and price reaches 50: unrealised loss 214.73, which is about 43 percent of the 500 put into the buys.

A tighter stop costs less when the trend continues. It also fires on a dip that would have recovered, and then the bot is gone and the recovery is missed. A stop that is far away behaves like no stop. Calculated by us on 29 Oct 2026, exact. A stop below the range also interacts with leverage if the grid runs on borrowed exposure: see futures grid liquidation, where the forced close can come before the stop does.

Spacing type

With arithmetic spacing every step is the same size in price, so the percentage gained per cycle is larger at lower prices and smaller at higher ones. In the 80 to 120 range with 10 grids, a step of 4 is 5.0 percent at a price of 80 and 3.3 percent at 120. Geometric spacing keeps the percentage the same at every level. The two give nearly the same result in a narrow range, and drift apart in a wide one. Neither is better in itself: arithmetic puts more grids where price is low, geometric spreads them evenly in percent.

Investment

Investment divided by the number of grids is the money in each order. Doubling it doubles the gain per cycle and doubles the loss for each unit of fall. It does nothing to the shape of the risk. It also sets how much of the account sits in a position that is stuck when price leaves the range.

How the settings interact

• More grids and a wider range can leave the spacing unchanged. The bot then has more orders, a bigger spread of cover and the same money cut into smaller pieces.
• A tight stop with a narrow range turns the grid into a small, quickly closed bet.
• Many grids plus a high fee rate push the net per cycle toward zero, whatever else is set.
• Leverage on top changes the meaning of the range, since the forced close may sit inside it.

How to check a set of settings

1. Work out the spacing in percent, and subtract the round trip fee and an allowance for spread. If the answer is not clearly positive, the bot loses on every cycle.
2. Work out the fall from the starting price to the lower bound, and compare it with moves the asset has made over the period the bot will run.
3. Work out the loss on everything the bot would hold at the lower bound, and at the stop if there is one.
4. Run the numbers in the grid calculator and compare the profit per cycle with that loss. The ratio of the two is the real shape of the bet.

What it does not tell you

The tables assume the price moves through the grid in an orderly way and every order fills at its price. Real fills have slippage, gaps skip levels and costs vary. The tables say what a setting does to the arithmetic. They say nothing about which direction price will go, and that is what decides how a grid ends.