Bitsgap: Grid Bots Across Exchanges, and the Range Problem

One screen, several exchanges: what that changes and what it does not

As we understand it, the arrangement is the common one for this kind of platform. Your funds sit at whichever exchange you choose. You give the platform an API key for each, and it places orders within the permissions on that key. What "from one place" adds is a single screen. What it cannot add is a single rulebook.

Every exchange keeps its own trading fee, its own smallest order size and its own history of outages. If you connect two exchanges you are running under two sets of terms, and the platform's own charge is only one line in the total cost of a bot. The first habit worth building is to treat each exchange as a separate account with its own limits, even when one dashboard shows them side by side.

A grid in two sentences, then the arithmetic

A grid bot divides a price range into equal steps and keeps a buy order under the market and a sell order over it. When a buy fills, the bot lists a sell one step higher, and when that sell fills, one round trip is done and the gap between the two prices, less fees, is the gain. Our pages on grid trading and grid bot fees and costs go through the mechanics in full.

Take one step in a made-up grid: buy $100 of a coin at 98, sell it at 100. Before fees the gain is $2.04, a step of about 2.04 percent. A fee is taken on the buy and again on the sell.

Calculated by us on 3 Nov 2026, exact. The step is thin, so a fee that looks small as a percentage is a large share of what the step earns. A coin that trades in a narrow band with a high fee can complete many steps and keep very little of each.

Worked example: the price leaves the range

All numbers are made up. A grid runs from 80 to 120 in steps of 2, starting with price at 100. Half the money, $1,000, sits in cash to fund 10 buy orders of $100 each, at 98 down to 80. The other half is 10 coins bought at 100, held to fill 10 sell orders at 102 up to 120. The account starts at $2,000. To keep the sums readable, fees are left out here.

The grid is a machine for harvesting wobble. On the way down it keeps buying, so by 70 it holds more coins than it started with, bought at falling prices, and it ends further behind than the position it started from. On the way up it sells every coin and sits in cash, missing the rest of the rise. In neither case does it cap a loss by itself. A stop or take-profit level, where a platform offers one, is a separate setting, and it is worth knowing before the start what a bot does outside its range.

A habit for a multi-exchange view: tie the dashboard to the exchange

• Net or gross. Does the profit shown already have exchange fees taken off, or only the platform's?
• Open orders. Do the orders listed on the platform match the open orders on the exchange itself, level by level?
• Totals. Does the platform's balance for each exchange match that exchange's own statement on the same day?
• One exchange at a time. Check each connected exchange separately. A match on one tells you nothing about another.

Judge a first grid by running the smallest amount the exchange allows, in a range you picked by looking at a chart of the past, and then watch what happens as price reaches an edge. That one edge is where most of the lesson sits. For the maths of a grid on your own numbers, use the grid calculator.

Where this page stops

Whether Bitsgap works, or is safe, is not something this page claims, and no other product is set beside it. The sums use invented numbers to show how fees and a range behave. Crypto-assets and leveraged products can lose all the money put into them, and past or simulated results are not a guide to future results.