Contents
What a grid bot is, in plain wordsWhere the money in a grid actually comes fromLong, short and neutral gridsThe four settings that decide everythingWhen the grid breaks: a trend that keeps goingWhy tight grids lose their profit to feesWhat to expect, and how to try it without stress
What a grid bot is, in plain words
Imagine the price of Ethereum bouncing between 2,800 and 3,200 USDT for a week. If you wanted to profit from that, you would have to sit by your phone buying every dip and selling every bounce, day and night. A grid bot does exactly that job for you: you give it a price range and a number of levels, and it spreads limit orders evenly across the range and keeps re-placing them as they fill.
There is no intelligence in it. It is software inside the exchange that places and replaces orders according to the rules you typed in, and it keeps doing that until you stop it.
On futures the bot isn't buying coins and keeping them. It opens and closes parts of a position with the money the exchange holds as your deposit, so if the price runs far enough against you the exchange can close the position itself — the same liquidation that applies to any manual trade. The leverage you give the bot decides how far "far enough" is.
Where the money in a grid actually comes from
The profit is the gap between two neighbouring rungs of the ladder, nothing more. Say ETH is around 3,000 USDT (all numbers here are illustrative), you set the range 2,800–3,200 and ask for 20 levels. The step between levels is 20 USDT, and each grid order is worth 50 USDT of position. The bot buys at 2,980, and if the price comes back to 3,000 it sells that piece again. Whichever rung it happens on, one closed pair on a 50 USDT order with a 20 USDT step gives you about 0.34 USDT before fees.
A third of a dollar does not sound like much, and on its own it isn't. The point of a grid is repetition: on a choppy day the price can cross the same few levels ten or twenty times, and each crossing closes another pair. Your result depends on how much the price wiggles inside your range, not on where it finishes. A sideways week with wide daily swings is the best thing that can happen to a grid bot.
Long, short and neutral grids
On futures you usually pick one of three modes. A long grid only works the buy side: it opens or adds to a long position at the lower rungs and takes profit at the higher ones, so you end up holding a long while the price sits near the bottom of your range. A short grid is the mirror image — it sells into strength and buys back lower, holding a short, which is how people run grids on a coin they think is slowly drifting down.
A neutral grid does not take a big position at the start. It leans long if the price drops below where you launched it and leans short if the price rises above, so it stays close to flat as long as the price keeps coming back to the middle. Exchanges label these modes in slightly different ways, so read the short description next to each mode in the app before you confirm anything.
The four settings that decide everything
- The range — the top and bottom price of your ladder. Too narrow and the price escapes in a day; too wide and the steps become large and rarely fill. Most people look at the last few weeks on the chart and pick the band the price has actually been living in.
- The number of grids — this sets the step size. More levels means smaller steps, more frequent fills and smaller profit per fill; fewer levels means rarer but bigger pairs.
- Leverage and the margin you give the bot. This decides the size of each order and how far the price can move against you before the exchange closes the position. A grid that survives a 15% move at 2x can be wiped out by the same move at 10x.
- Trigger and stop prices. A trigger price makes the bot wait and only start when the price reaches a level you like. A stop price stops the bot when the market leaves your range. Most exchanges then ask what to do with the position the bot has built — check which option you choose, because if you keep it, you are left with a normal trade and no ladder behind it.
There is also a floor on the size of each order: on Binance, for example, an ETH futures order starts at 20 USDT of position and most altcoins at 5 USDT, so a small deposit simply cannot be split into fifty levels. The quickest way to feel how these settings interact is to open a Binance Futures account, go to the bot section and fill in the form without starting anything. Before you press start, look at what the form shows you — step size, number of orders, the margin it will use and the liquidation price if it gives one. If something is not shown, check it in the app.
The bot has no opinion about the price. You pick the coin, the range and the direction, and it works that ladder until you stop it. If your range is wrong, it will carry out a wrong plan very efficiently.
When the grid breaks: a trend that keeps going
Here is the catch. A long grid buys every step down, so a steady fall fills your whole ladder and leaves you holding the full position at an average price somewhere inside the range. The closed pairs still show a tidy little profit, while the open position sits in a much bigger loss. Nothing has malfunctioned — the bot did exactly what you asked, and the market simply did not come back.
This gap between the "grid profit" number and your actual balance confuses almost every beginner, and it gets wider once you add trading fees and the funding payments that come with holding a perpetual position — small amounts that longs and shorts exchange every few hours, quietly accumulating while the bot holds its position overnight. The full breakdown lives in my grid bot shows profit but my total is negative.
The protection is boring and it works: modest leverage, a stop price below your range, and money you can afford to lose. A grid with no stop and 20x leverage in a falling market is the classic way to find out that a bot can lose everything.
Why tight grids lose their profit to fees
Every fill costs you a fee, and a grid fills constantly. Keep the same 2,800–3,200 range but ask for 100 levels and your step becomes 4 USDT — about 0.13% of the price. On Binance's base fee level a limit order that waits in the order book costs 0.02% and an order that takes the price immediately costs 0.05%, charged on entry and on exit separately. In the best case that eats roughly a third of each tiny pair; in the worse case it eats almost all of it.
So the step has to be clearly bigger than the round-trip fee, or you are running a machine that trades beautifully and earns nothing. Fee levels differ by exchange and by your volume, so check your own fee line in the app rather than assuming — how futures fees are charged walks through the arithmetic.
What to expect, and how to try it without stress
A grid bot is not a money machine and nobody can tell you a percentage it will return. In a calm, swinging market it collects small amounts steadily; in a strong trend it hands most of that back and more. Treat the first one as an experiment: a small amount, a coin you follow, a range you can defend in one sentence, and a stop price you set before you press start.
The honest way to learn the rhythm is to run one small grid for a week and watch it daily — where it filled, what the open position looks like, how much went to fees. When you are ready for the buttons themselves, launching your first bot step by step covers the setup flow from the beginning. Availability differs from country to country, so check during sign-up whether futures and bots are offered where you live.
Is a grid bot profitable?
It can be, in a market that keeps swinging inside your range, and it loses money when the price leaves that range and keeps going. Nobody can promise you a return, because the result depends entirely on how the price behaves during the days your bot is running. Judge it by your account balance, not by the "grid profit" line.
What are good grid bot settings for a beginner?
A range taken from the price band of the last few weeks, a step clearly larger than the round-trip fee, low leverage, and a stop price under the range. Fewer levels with a sensible step beat a hundred levels that fill all day and feed the fee counter.
Does a grid bot work in a strong trend?
Badly, if the trend goes against the direction of your grid. A long grid in a falling market buys all the way down and ends up holding the whole position at a loss, while the closed pairs show only a small gain. A grid pointed in the same direction as a gentle trend does better, but a fast one-way move still runs past your top rung and stops earning.
How is a futures grid different from a spot grid?
On spot you own the coins and the worst case is a bag you hold for a long time. On futures the bot works with the deposit the exchange holds, so leverage, funding and liquidation all enter the picture — more can go wrong, and faster. The same range and the same step can be a patient spot strategy or a liquidation, depending on the leverage you set.
Can I stop a grid bot whenever I want?
Yes — you can stop it at any time. On most exchanges you then choose what to do with the open position the bot has built up; check which option you are picking, because if you keep the position you now have an ordinary manual trade with no ladder behind it. Decide right then whether it needs a stop-loss.
Do I need experience to run one?
You need enough to choose a range and accept the loss if the price leaves it — that is a trading decision, not a technical one. The bot handles the clicking, not the judgement. If you cannot explain in one sentence why the price should keep bouncing inside your range, you are not ready to fund that grid yet.