Grid Bot Settings That Survive Fees: Range, Grid Count, Spot vs Futures, and Infinity Grids

What are good grid bot settings? A range that matches recent realized volatility, spacing wide enough to clear round-trip fees, and a grid count you can actually fund at your order size. Spot grids risk holding a falling coin; futures grids add funding and liquidation; infinity grids remove the ceiling but keep the downside.

Every knob on a grid bot setup screen is a risk decision wearing a profit dial's clothes. Binance's spot grid bot accepts anywhere from 2 to 500 grids, which turns "how many grids should I use?" into a live question — because more grids means thinner spacing, and thin spacing is how a bot loses money on trades that technically work. If you are new to the mechanics, start with our explainer, what a grid trading bot is and when it loses money. This page then walks each setting in the order you should decide it.

Setting 1: The range

The range is the only setting that determines whether the strategy applies. Everything else is calibration inside it.

How to pick it. Anchor to recent realized volatility, not to a round number and not to hindsight. A workable starting frame: look at the asset's observed high-low envelope over a period comparable to how long you plan to run the bot (days for a swing-timeframe grid, weeks for longer), then widen it. A range calibrated exactly to last month's action gets punctured by one normal breakout.

What happens when price exits — in each direction:

• Breaks below the lower bound: the bot has bought every leg down. It is now holding full inventory at a loss, its cash is deployed, and its sell ladder sits above a price that may not return. This is the failure mode that dominates.
• Rallies above the upper bound: the bot sold out at each rung and sits in cash. The loss is opportunity, not principal — it captured the early part of the move and watched the rest.

Set exit bounds before you start. A range without a planned exit is not a range; it is a hope. Decide in advance at what condition you shut the bot down — a hard price level, a drawdown threshold, a time stop. Our kill switch architecture article covers the mechanics of standing a bot down, and our guardrails for autonomous trading agents covers exposure caps — both apply to a small exchange-native grid exactly as much as to a custom system.

Setting 2: Grid count and spacing — the fee floor

Grid count and spacing are one decision expressed two ways: count divides the range, spacing is what's left over.

spacing = (upper bound − lower bound) ÷ number of grids

Here "number of grids" means the number of intervals the range is divided into, the convention Binance's own grid parameter documentation uses. If a platform counts price levels instead, the number of intervals is levels − 1. For example, a hypothetical range of 60,000 to 66,000 with 20 grids gives (66,000 − 60,000) ÷ 20 = 300 per grid.

Arithmetic spacing puts equal price gaps between rungs; geometric spacing puts equal percentage gaps. On anything volatile, geometric is the better default — a 300 gap means something entirely different at 20,000 (1.5%) than at 65,000 (about 0.46%), and percentage spacing keeps the economics of each cycle uniform.

The fee floor. Every cycle pays two fills, two fees:

cycle profit = spacing − round-trip fees (+ slippage)

If spacing is tighter than the round-trip fee, the bot loses money on every successful trade. A hypothetical illustration with round numbers: suppose a venue charges 0.1% per fill and your order size is $10,000 — that is $10 per fill, $20 per round trip, or 0.2% of the order. Each rung's spacing therefore has to be wider than 0.2% of price (about $20 of gross capture on a $10,000 order) just to break even. A grid with 0.15% spacing would capture $15 and pay $20, losing $5 on every completed cycle even though the trade "worked."

Fee levels differ widely by venue and tier. For reference, Hyperliquid's fee schedule lists a base-tier perpetuals maker fee of 0.015%, so a round trip where both fills are maker orders costs about 0.030% (2 × 0.015%), before slippage and funding — a far lower floor, which is why fine spacing can make sense on low-fee maker-only venues and rarely does on taker-heavy ones. Fee schedules depend on rolling volume and tier discounts, so always re-check the exchange's own current fee page at setup time. For a breakdown of the other cost layers around a bot (subscriptions, spread), see our guide to what AI crypto trading bot fees actually cost.

Why 500 grids is a ceiling, not a target. A platform allowing 500 grids is a feature limit, not a recommendation. Doubling grid count halves spacing and halves your fee headroom per cycle. The right count is set by the fee floor from below and by your capital from above — each rung needs an order size you can fund, and a 100-rung grid with dust-sized orders just fragments the same capital across more fees.

Setting 3: Order size and capital split

The last knob decides what the bot is holding at the worst moment. At the bottom of the range, a fully-filled grid is 100% inventory; at the top, it is 100% cash. Size orders so that the maximum-inventory state is a position you would have chosen to hold on purpose. If the fully-deployed state would be oversized for your risk tolerance, the grid is too big — shrink the capital, not the conviction.

Spot grid vs futures grid

Spot grid

• Holds: the actual asset.
• Leverage: none in a standard spot grid.
• Funding and liquidation: none. The worst case is holding a depreciated asset.
• Direction: primarily long exposure.
• Costs per cycle: the exchange's spot fee tier.

Futures grid

• Holds: perpetual contracts, supported by margin.
• Leverage: available. Pionex states its futures grid supports up to 100× depending on market and setup.
• Funding: paid or received on the open position (settled every 8 hours on Pionex).
• Liquidation: yes. A leveraged grid in a trend can be wiped out, not just drawn down.
• Direction: long, short, or neutral modes.
• Costs per cycle: the futures fee tier, plus funding on the open position.

Pionex's futures grid documentation describes long, short, and neutral modes, a funding rate settled every 8 hours, and liquidation triggered off the mark price. The honest framing: a futures grid is a spot grid plus three new ways to lose — funding drag on a position held in the wrong regime, leverage multiplying the trend-loss, and a liquidation price that converts a bad week into a zero. Even a neutral-mode grid is not liquidation-proof, because filled orders build exposure. We don't recommend leveraged grids.

A first grid belongs in spot. The spot grid's worst case — you hold the asset — is a position you can survive long enough to think about.

The infinity grid bot

Pionex.US's help center describes its Infinity Grid as a version of its grid bot where you never have to worry about price exceeding the grid range: you set a lowest price and a profit-per-grid percentage rather than an upper bound. In its worked example, the bot keeps the value of the coins it holds roughly constant, selling small portions as price rises and buying more as it falls — and it keeps doing so as long as price stays above the lowest price you set.

Note the asymmetry, because the marketing won't point it out: removing the upper bound fixes the upside failure (a bounded grid sitting in cash above a breakout) and leaves the downside failure intact. A falling market still has the bot buying on the way down, and a break below your lowest price ends the setup. "Infinity" describes the ceiling, not the safety. Features and parameters differ by platform and by region, so read the venue's own documentation before relying on any of this.

Settings checklist before you start

1. Paper-trade first. Run the exact configuration in simulation or a test environment long enough to see at least one adverse move. Our case for paper trading before going live applies to grid bots as much as to reasoning systems.
2. Define the max loss before funding. A number, written down: the total drawdown at which you shut it off, not a feeling.
3. Write the exit rule. Price level, time stop, or regime condition — and how you'll enforce it when the bot is happily cycling while the range dies.
4. Scope the API key. Trade-only, no withdrawal permission. See our API key safety checklist.
5. Check the fee floor. Confirm current maker/taker fees on the exchange's own fee page and verify spacing clears the round trip with headroom.
6. Have a kill switch. Know, before launch, the two clicks that stop the bot — and that they work when you need them.
7. Label your expectations. If you cannot state what market condition this bot needs to work, you don't have settings; you have a slot machine.

FAQ

What are the best grid bot settings? There is no universal set. The right settings are a function of the asset's current volatility, the venue's fee schedule, and your capital — which is why the fee floor and range-fit matter more than any copied "best settings" screenshot. Anyone selling fixed settings is selling yesterday's regime.

How many grids should I use? As few as clear the fee floor comfortably, and no more than your capital funds at a meaningful order size. If spacing after fees leaves little headroom over the round-trip cost, you have too many grids. A platform's maximum is a cap, not a target.

Spot grid vs futures grid — which should I use? Spot for a first grid: no leverage, no funding, no liquidation, worst case you hold the asset. Futures grids add funding drag and liquidation risk and belong only with a stated reason for the leverage — which most setups don't have.

What is an infinity grid bot? A grid bot with no upper price bound: per Pionex.US's documentation you set a lowest price and a profit-per-grid, and the bot keeps trading as price climbs. It remains fully exposed on the downside.

What leverage should I use on a futures grid? We won't give a leverage recommendation — leverage is the setting that converts a grid's worst case from a drawdown into a liquidation. If the strategy only makes sense at 20×, the strategy doesn't make sense.

How does a grid bot calculator work? The typical inputs are range, grid count, order size, and fee rate; the outputs are spacing, capital required, and profit per cycle after fees. The critical line in any such calculation is spacing minus round-trip fees — if a calculator doesn't show the fee-adjusted number, it is showing you marketing. We don't publish a calculator tool; the arithmetic above is the whole thing.

---

Risk disclaimer: This article is for information only and is not financial advice. Grid trading loses money in strong trends in both directions, and leveraged futures grids can be liquidated. No APR, ROI, or profit figures are published here because none can be honestly promised, and all worked numbers are hypothetical illustrations, not backtests or results. We do not recommend using leverage. Exchange limits and fees change; verify them on the venue's own pages. Trading involves substantial risk of loss; do your own research and never deploy capital you cannot afford to lose.

---

If you want an outside read on how your site and content actually surface in AI answer engines — the same question behind this cluster — run a free scan →