Best Grid Trading Bot? What to Check

What a grid bot's settings actually are

Before you compare bots, know what you are comparing. A grid bot is a small set of numbers plus a rule. The bot draws price lines, puts buy orders below the market and sell orders above, and replaces each order when it fills. The settings below are all it has.

• Range: the lowest and highest price the grid covers. Outside it, the bot has no orders. This is the biggest risk setting.
• Number of grids: how many steps the range is split into. More grids mean smaller gaps, more fills and more fees.
• Spacing type: arithmetic (equal price gaps) or geometric (equal percent gaps). Which exist varies by platform.
• Amount per grid: the size of each order, which sets the money at work and the size of the pile if price falls.
• Stops and triggers: an optional stop below the range, a stop above it, or a start price. Some bots have none.
• Leverage and direction (futures only): these bring a liquidation price into play. See grid bot settings for each setting in more detail.

Two bots with the same settings behave the same on the same prices, apart from fees, fills and downtime. So the difference between bots is mostly in the costs, the safeguards and what is shown to you, not in some secret method.

The checklist: questions to ask any grid bot

Use the table as a list of questions to answer from the bot's own documentation and your own test, not from a promise. If you cannot find the answer, treat it as unknown.

Not every bot can answer every row. A bot that answers most of them in writing, before you deposit anything, tells you something about how it is run. A bot that answers none of them tells you something as well.

A worked example: what fees do to a grid with many small trades

This is a made-up example, not any real bot. The range is 100 to 120 in 20 grids with a step of 1, so each grid buys at one line and sells at the next. Each grid trades 10 units. The fee is 0.1 percent of the traded value on every fill, and each completed grid has two fills. Everything below was calculated by us on 30 Oct 2026 and is exact.

Each grid grosses 10.00. The fee is about 2.01 at the bottom and 2.39 at the top, since a higher price means a bigger traded value. One full cycle of all 20 grids grosses 200.00, pays 44.00 in fees and nets 156.00. Fees take 22.0 percent of the gross. That is before any loss from price leaving the range.

The same grid at three fee rates

The gross never changes. Only the fee does. Fee schedules vary by venue and by account, so use the numbers from your own, and ask whether a fee is charged on both sides.

The same fee on a narrower grid

Now keep 20 grids and the 0.1 percent fee, but squeeze the lines to a step of 0.20, about 0.2 percent apart. The round trip fee is about 0.2 percent, so it nearly equals the gap.

A bot that advertises "more grids" or "more trades" is not showing you a better result. More fills means more fees, and the gap between the lines must be wider than the round trip fee for a completed grid to earn anything. The grid calculator lets you try your own range, and grid bot fees and costs goes through the arithmetic in full.

Red flags

• A promised or implied return. No grid can know what price will do. A stated monthly or yearly return is a claim about the market, not about the bot.
• Fees that are hard to find. If you cannot work out the cost per fill, and any profit or subscription shares, you cannot work out the result.
• No stated risks. Every grid has a range, and every range has an edge. A page that never mentions what happens at the edge is leaving out the main point.
• Stats that show only wins. Look for realised and unrealised results, the drawdown and the period covered. See are grid bots profitable for what evidence would count.
• Pressure to deposit or to raise leverage. Higher leverage shortens the distance to liquidation, and the seller of the bot is not the one who bears it.
• Withdrawal rights in the keys. A trading bot needs permission to trade. It does not need permission to move money out.

What a score or a review cannot tell you

A ranking, a star rating or a review can describe how a bot looked to someone else, on their market, over their dates, with their settings. It cannot tell you:

• what price will do next, which decides whether a grid is a good fit;
• whether the reviewer's range is a range you can afford to be wrong about;
• what your fees will be, since they vary by account and venue;
• how the bot behaves in an event it has not yet met, such as a fast gap through the range;
• whether a review was independent or paid for.

Treat a review as a source of questions, then go back to the checklist. Test the settings you plan to use on a demo or very small size first, and read the stats the way the example above reads them: gross, fees, open loss, then net.