How grid trading on Hyperliquid works

Grid trading on Hyperliquid means placing orders across a defined price range so the account can buy lower and sell higher as price moves through the grid.

Anello’s grid bots run on Hyperliquid spot markets and are built around exact previews, USD-per-order sizing, and Hyperliquid-specific precision rules. This article explains the mechanics end to end.

The core idea

A grid has a lower price, an upper price, a number of levels, and an order size. The bot uses those settings to create a ladder of prices: buy orders rest on the rungs below the current price, sell orders on the rungs above. When a buy fills, the bot places a sell one rung higher; when that sell fills, the round trip has captured the gap between the two rungs, and the cycle repeats.

The strategy is mechanical. It does not predict direction; it defines behavior if price moves within the range. In a ranging market, repeated movement between levels produces many small realized trades. No single trade is dramatic. The output is the accumulation.

A worked example

Suppose BTC trades around $100,000 and you configure a grid from $90,000 to $110,000 with 21 levels and $100 per order.

• The spacing ratio is (110,000 / 90,000) ^ (1/20) ≈ 1.0100, so each rung sits about 1% above the one below it.
• Each order is sized as $100 of BTC at that rung’s price: the buy near $99,000 is for about 0.00101 BTC, the buy near $90,000 for about 0.00111 BTC. Dollar value per order is constant; the base amount varies with price.
• Price dips to $99,000 → that buy fills → the bot posts a sell one rung up near $100,000. Price recovers → the sell fills → that round trip realized roughly 1% on $100, minus fees.

A choppy week can cross the middle rungs dozens of times. Each crossing in both directions is another cycle.

Why geometric spacing

Anello spaces levels geometrically: each rung is the previous rung multiplied by a constant ratio, rather than separated by a fixed dollar amount. Fixed-dollar spacing makes rungs at the top of a range proportionally tighter than at the bottom ($1,000 is 1.1% at $90k but only 0.9% at $110k), so identical fills would earn different percentages. Geometric spacing keeps the percentage move, and therefore the profit per round trip, consistent across the whole range. For an asset class quoted in percentage moves, that is the right default, and it is the only spacing Anello supports.

The preview is the contract

Before a bot is created, Anello shows every order it will place: exact prices (rounded to Hyperliquid’s tick rules) and exact sizes (rounded to the market’s size precision). What you see in the preview is what the bot places. No drift, no recalculation behind your back. If the preview doesn’t look right, nothing has happened yet; adjust and preview again.

Two sizing rules are enforced up front so orders don’t fail later:

• Each order must be worth at least $20 at creation (Hyperliquid rejects orders under $10 notional, and the buffer keeps rounded orders safely above that floor).
• Level spacing must be at least 0.3%, so rungs stay distinct after price rounding and fees have room to be covered.

How the bot runs

Once placed, the grid maintains itself. The bot listens to Hyperliquid’s real-time feed: when an order fills, the replacement on the opposite side is placed automatically and the fill is accounted, including across disconnects and restarts. You don’t need to do anything for fills; that loop is the product.

The exception is orders you cancel by hand on Hyperliquid. The bot doesn’t silently override a deliberate cancel. Restoring those rungs is a user-approved action called repair, covered in Repair manually canceled grid orders.

Adjusting a running grid

A live grid can be changed without stopping and recreating it:

• Extend the range. Add rungs below the bottom or above the top at the grid’s existing spacing. Existing orders don’t move; only the new rungs are placed.
• Add funds. Increase the USD size per order. The bot re-places the grid at the larger size on the same price levels, keeping the bot’s history and profit ledger intact.
• Compounding. Off by default: every order stays at your fixed USD size and realized profit accumulates in your balance. Switched on, each re-post carries the full proceeds of the fill it follows, so gains stay inside the grid and order sizes grow over time. You can flip it on a running bot; the next fill follows the new setting.

Common failure modes

A grid can underperform or lose money in predictable ways:

• Price exits below the range. The rungs filled buys all the way down, so the bot holds inventory bought above the current price, and the grid stops trading below its bottom rung. This is the main risk. The unrealized loss resolves only if price returns.
• A sustained trend instead of a range. Down-trends create the inventory problem above; up-trends sell the inventory early and leave the grid watching the rest of the rally from the sidelines.
• Spacing too tight relative to fees. If the gap between rungs barely exceeds the round-trip fee cost, the grid works for the fee schedule, not for you. See fees and risks for the actual numbers.

The practical defenses: a range wide enough to contain realistic movement, order sizes that leave a full-range fill comfortable, spacing comfortably above fee cost, and notifications so you notice when the bot needs attention.

Because Anello grids trade spot, the inventory a grid holds is owned outright. There is no leverage, no funding drag, and no liquidation price. The drawdown risk is real; the force-closure risk is not. For the full risk picture, including the perp mechanics that apply to positions you take on the terminal, read Risks of grid & perp trading.