How a grid bot earns per-grid profit, why fees and grid count decide whether it pays, and why breakouts and trends out of the range are the real risk.
Strategy Templates · Updated 2026-10-08
A grid bot splits a price range into evenly spaced levels and places a buy order below the current price and a sell order above it at each level. When price falls one step, the bot buys; when price rises back one step, it sells what it just bought. Each completed buy-then-sell round is one grid profit. The bot does not need to predict direction. It needs price to keep moving back and forth through its levels.
Four settings define the bot. The lower and upper bound set the range. The grid count sets how many levels sit inside it. Spacing follows from those: in an arithmetic grid each step is a fixed amount, (upper − lower) ÷ grids; in a geometric grid each step is a fixed percentage, with the ratio (upper ÷ lower)^(1 ÷ grids). Finally, the investment is split evenly, so each grid gets investment ÷ grids of quote currency.
The profit on one round is not simply the step size. You pay a fee on the buy and again on the sell, so the per-grid profit is (sell ÷ buy) × (1 − fee)² − 1. This formula is the core of everything below: a grid bot is a machine for collecting many small, fee-sensitive gains.
This uses the example on the site's grid bot profit calculator. Range 50,000 to 70,000 USDT, 20 arithmetic grids, 10,000 USDT invested, a 0.1% fee per trade, started at 60,000. The step is (70,000 − 50,000) ÷ 20 = 1,000 USDT, and each grid is funded with 10,000 ÷ 20 = 500 USDT.
At the bottom of the range, a round buys at 50,000 and sells at 51,000. Gross, that is 51,000 ÷ 50,000 = 1.02, a 2.00% move. After fees: 1.02 × (0.999)² = 1.02 × 0.998001 = 1.017961, so the net profit is about 1.80%, roughly 9 USDT on a 500 USDT grid. At the top of the range, a round buys at 69,000 and sells at 70,000. Gross, 70,000 ÷ 69,000 = 1.014493, a 1.45% move; net, 1.014493 × 0.998001 = 1.012465, about 1.25%. That is why the calculator reports 1.25% to 1.80% per grid: in an arithmetic grid, the same 1,000 USDT step is a smaller percentage at higher prices.
Now change only the fee. At 0.04% per trade, the fee Live Paper Trading charges on paper fills, (0.9996)² = 0.99920016. The bottom grid becomes 1.02 × 0.99920016 = 1.019184, about 1.92%, and the top grid 1.014493 × 0.99920016 = 1.013681, about 1.37%. Cutting the fee changed per-grid profit by about 0.12 percentage points, which is small when each step is 1,000 USDT.
Now change the grid count instead. With 100 grids in the same range, the step shrinks to 200 USDT. At the bottom, 50,200 ÷ 50,000 = 1.004 gross; net at a 0.1% fee, 1.004 × 0.998001 = 1.001993, about 0.20%. At the top, 70,000 ÷ 69,800 = 1.002865 gross; net, 1.002865 × 0.998001 = 1.000860, about 0.09%. Fees now eat roughly half the bottom grid's 0.40% gross move and about 70% of the top grid's 0.29%. Push it to 200 grids (a 100 USDT step) and the top grid's net result is about −0.06%, a loss on every completed round. More grids means more trades, but past a point it means paying the exchange more than you earn.
Grid bots do best in a ranging market: price oscillates inside the range many times, each oscillation completes rounds, and the bot banks a small profit on each. The more round trips, the more profit, as long as each round clears its fees.
They do worst when price leaves the range. The same calculator example shows both exits. If price falls straight to 50,000, every buy order fills on the way down and none of the sells do, so the bot ends up holding coins bought at an average above 50,000. It is worth 8,757.86 USDT, a 12.4% loss. Simply holding from 60,000 would be worth 8,325.00, a 16.75% loss, so the grid lost less, but it still lost. If price rises straight to 70,000, the bot sells its coins one level at a time and ends with 10,447.42 USDT, while holding would be worth 11,655.00. In a strong uptrend the grid gives away most of the move.
The underlying issue is inventory. A grid bot is always holding some coins and some cash, and the mix shifts against you: as price falls it accumulates coins, and as price rises it sells them. In a trend that never comes back, the bot ends up fully in coins at the bottom of a falling market or fully in cash at the top of a rising one. Completed rounds show profit, while the open inventory carries an unrealized loss that can be larger than all the grid profit collected so far. If price leaves the range and the bot stops trading, that open position can sit there for a long time.
The Live Paper Trading record lists two grid bot templates on BTC/USDT in its delayed snapshot as of 7 October 2026. Both cover a performance period from 9 September 2026 to 21 September 2026, and both show a last update of 21 September, so as of the 7 October snapshot their records were not current. One shows a total return of +0.015% with a maximum drawdown of −0.077%; the other shows −0.014% with a maximum drawdown of −0.113%. Both show a win rate of about 78.9%, and Sharpe is displayed as 0.00, so it tells you nothing here.
That pattern is typical of a grid. A high share of winning trades sits next to a total return close to zero, because each win is small and the open inventory marked at the current price offsets it. When you read a grid bot's record, start with the dates and the last update time, then compare total return with maximum drawdown, and treat the win rate as a count of completed rounds rather than evidence of profit. Twelve days in a quiet market say little about what happens when price breaks out.
The calculator's exit cases assume price moves straight to a bound with no bounces. Real paths are messier, and a range that bounces a lot on the way down will collect some grid profit that partly offsets the inventory loss. The examples also ignore slippage, funding on perpetual futures, and leverage, all of which change results. A few weeks of paper results do not show how a template behaves through a crash or a long trend, and paper fills are not real executions. Neither the calculator nor the league tells you which range to pick or when a market will stop ranging.
A grid bot trades many small, fee-sensitive gains for exposure to large moves out of its range. Check the per-grid profit after fees before choosing a grid count, know what the bot is worth at each bound, and size the position so a break below the range is survivable. This is education, not investment advice.
Educational content, not investment advice. Guide content and league rankings are independent of any exchange or sponsor relationship.