Why do grid bots lose money?

How a grid bot makes money in the first place

A grid bot splits a price range into levels and places a buy order below the price and a sell order above it at each level. Every time price falls to a level and rises back to the next one, the bot completes a small round trip: it buys low and sells one step higher.

The income comes from swings, not from a forecast. A sideways market with many back-and-forth moves is the bot's best case. A market that moves in one direction is its weak spot, and almost every grid bot loss starts from there.

1. Price trends out of the range

In a falling market, a long or neutral grid keeps buying at every level on the way down. These buys are not closed by sells, because price never comes back up to them. The bot ends up holding a growing position bought at higher prices: inventory.

Grid profit and open loss are not symmetric. Each round trip earns one step; the open position loses on every unit for the whole distance price has fallen.

A short grid has the same problem in a rising market. A neutral grid is exposed on both sides: whichever way price leaves the range, it holds the losing side.

What limits it: choose the range from recent volatility, not from a hope; decide in advance what you will do if price leaves the range; and compare a directional setup with a neutral one before choosing.

2. Steps are too small, so fees eat the profit

Each round trip pays a fee twice: on the buy and on the sell. If the distance between levels is small, most of the step goes to fees. Many tight levels look busy, but the net result can be close to zero, or negative with taker fees.

What limits it: keep the step at least several times the round-trip fee. The grid bot calculator shows the net profit per grid and the share that goes to fees.

3. Leverage turns a drawdown into a liquidation

On futures, a grid bot uses leverage. The worst moment is when every order on the losing side has filled: the position is at its largest and its average price is far from the current price. If the liquidation price at that moment is inside the range, a normal move to the edge of the range can close the whole position.

A liquidation is final: the bot cannot wait for price to come back, and the loss is realised at the worst point.

What limits it: check the worst-case liquidation price (all orders on one side filled) before starting, and keep it well outside the range. Lower leverage moves it further away. Terms: leverage and liquidation in the glossary.

4. Funding quietly erodes long-running positions

Perpetual futures charge or pay a funding fee at regular intervals, usually every 8 hours and more often on some pairs. A bot that holds inventory for days pays (or receives) funding on the full position size, not on the margin.

What limits it: look at the current funding direction for the pair, and count funding separately when you judge a bot's result. See funding fee in the glossary.

5. Trailing settings drag the grid into a worse position

Trailing (range shifting) moves the grid when price leaves the range. Trailing in the direction of the trend can help; trailing against the position can hurt. On a long grid, trailing down keeps buying as the grid follows a falling price, so the long inventory grows. If the decline continues, losses grow faster than with trailing off.

6. The test looked better than reality

Many people set up a bot after a good-looking backtest. Backtests are useful but can be optimistic: a fill may be counted when price only touched a level, thin order books are treated as deep, fees and funding are simplified, and the tested period may simply have been sideways.

What limits it: test with realistic rules and over different market conditions. See how to test a grid bot.

7. There is no exit plan

Without a rule written before the start, the decision to stop is made in the worst state: during a drawdown, under stress. Typical outcomes are closing at the bottom or holding a losing position far too long.

What limits it: before starting, write down when the bot stops: a time limit, a maximum inventory, a price outside the range, or a loss you accept. Then follow it.

Checklist before you start a grid bot

• Range: based on recent volatility; what happens if price leaves it on either side?
• Step: several times the round-trip fee; check the net profit per grid.
• Liquidation: worst-case liquidation price well outside the range.
• Funding: current rate and direction for the pair.
• Direction: compare long or short with a neutral setup.
• Exit rule: written down before the start.
• Test: run the same settings virtually on live prices first, then start small.

A grid bot is not a way to avoid risk; it is a way to take a specific risk (a range) in exchange for small, frequent profits. Knowing where the losses come from is what makes that trade-off a choice instead of a surprise.