A 10 USDT BTC long grid at 50x showed a liquidation price of 60,261.4, right inside its own 33,000-90,000 range. How BYDFi futures grid settings get you there

Look at the red box in the image before anything else. BYDFi's grid form is telling the user their bot would be liquidated before price ever reaches the bottom of the grid they just drew. That one warning explains most of what can go wrong with a futures grid, so this post decodes it, then walks through how BYDFi's futures grid works.

The image is a real BYDFi web screen from BYDFi's own Futures Grid help article: the BTCUSDT perpetual chart (header price 77,783.6, chart dated Sep 2, 2026) with the Futures Grid panel on the right. Settings shown: Long, price range 33,000 to 90,000, 2 grids, arithmetic, 10 USDT investment, 50X leverage. The estimate block reads 0.0035 BTC order quantity per grid, 99.75 USDT (997.50%) estimated profit per grid, estimated long liquidation price 60,261.4 USDT, and a risk reminder that the liquidation price is within the grid range. These are BYDFi's capture values, not live prices.

Decoding that screen

• BYDFi grids are arithmetic: every gap is equal. (90,000 - 33,000) / 2 = 28,500 between levels.
• One grid step on 0.0035 BTC: 0.0035 x 28,500 = 99.75 USDT. That is the profit-per-grid figure shown.
• 99.75 / 10 USDT invested = 997.5%. The big percentage comes from 50x and a very wide step, not from anything special.
• The estimated long liquidation price assumes every long in the grid gets filled. At 60,261.4 it sits well above the 33,000 floor, so a drop into the lower half of the range would liquidate the bot first.

BYDFi's own fix list when this warning shows: add more investment, narrow the range, or lower the leverage.

The three directions

• Neutral: opens nothing at start. Shorts as price rises, longs as it falls. Built for sideways ranges.
• Long: only longs. Adds on dips, closes on rises. For a market grinding up with chop.
• Short: only shorts. Adds on rises, closes on drops. For a market grinding down.

Long and short grids open starting positions when activated. BYDFi's example: ETHUSDT long grid, 2,000 to 3,000, 10 grids, activated at 2,750. Three grid levels sit above price (2,800, 2,900, 3,000), so the bot opens 3 longs and closes them one by one as ETH climbs through those levels. The same short grid activated at 2,650 opens 7 shorts for the 7 levels below price (2,600 down to 2,000).

Settings that matter

• Price range: too wide and orders fill rarely; too narrow and a one-way move can liquidate the bot. Upper must be above lower.
• Number of grids: more grids = more fills but smaller profit each and more capital needed. Overly dense grids can see profits eaten by fees.
• Investment: sets the size of each order. It comes out of your Spot Wallet, not your futures wallet.
• Leverage: 1x to 50x.
• Optional: activation price (bot waits until price reaches it, must be inside the range), upper and lower stop prices (bot stops when hit, usable as take-profit or stop-loss).

House rules worth knowing

• Minimum 10 USDT to start, no strategy creation or management fees, normal futures fee rate of your main account applies
• Up to 30 grid bots running at once
• Max position size is capped at Risk Limit Tier 1 for that pair
• Every bot order is a limit order
• Bots run in Hedge Mode and Cross Margin Mode, USDT-M futures only
• You can add funds anytime, but can't withdraw below the original investment
• A bot stops on its own if it is liquidated, if the pair is delisted, or if a stop price is hit. Leftover funds go back to the Spot Wallet

Why grids lose money, per BYDFi

Wrong direction for the trend, stopping the bot before it reaches profitable levels, and grids so tight that fees eat the profit.

Before you press Create Bot

• Is the estimated liquidation price outside your range? If not, change range, leverage or size first.
• Does the direction match the market you actually expect: range, grind up, or grind down?
• Are the grid steps wide enough to clear fees on each round trip?
• Have you set a lower stop price you can live with?