Hyperliquid TWAP orders explained

What it does

A TWAP — time-weighted average price — takes one order and releases it as a series of small suborders across a window you set.

The point is not the average price. The point is that the book has time to refill between the pieces. A single large market order walks up the book in one move and pays for every level it consumes. The same size spread over an hour meets a fresh book each time, and the fills cluster much closer to the price you saw.

That is the whole mechanism. Everything else on this page is the fine print.

The limits

Two of those matter more than the others.

The interval floor means a long window does not buy you more suborders, it buys you smaller ones. A one-hour TWAP and a six-hour TWAP of the same size fire at a similar cadence; the six-hour version simply cuts the size into more, smaller pieces.

The slippage ceiling is the safety mechanism. If a suborder cannot fill within 3% of the price at that moment, it is skipped rather than filled at any price. The size does not vanish; it carries forward into the remaining suborders.

The fee point people get wrong

A TWAP’s suborders are takers. Every one of them crosses the book, so the whole order pays the taker rate.

So a TWAP does not belong on the list of ways to reduce fees. It belongs on the list of ways to reduce cost, which is a different and usually larger number.

The arithmetic is easy to sanity-check. At the entry tier the taker fee is 0.045% of notional. If market-ordering your size in one go would move the price by more than that — and on any market outside the top handful, in real size, it will — then the impact you avoid is worth more than the maker fee you gave up.

Placing one

1. Check the book can take your size at all

   A TWAP manages impact; it does not create liquidity. If your order is a meaningful fraction of the market’s daily volume, no schedule fixes that. Judging depth comes first.

   You should see volume and open interest large enough that your order is a small share of a day

2. Choose TWAP in the order ticket

   The ticket switches from a price to a duration. There is no limit price on a TWAP; the slippage ceiling is the only price control you get.

   You should see fields for total size and a runtime, instead of a single price

3. Set the runtime against how urgent the trade is

   Longer is gentler on the book and worse on timing risk. A window measured in hours is normal for size in a liquid market. A window measured in days is a different kind of decision, because the market can move further than your saving during it.

   You should see a window long enough for the book to refill, short enough that the thesis still holds

4. Attach reduce-only if this is an exit

   Same rule as every other exit. Without it, a TWAP that overshoots the position keeps going and opens one the other way, piece by piece, which is harder to notice than a single order doing it.

   You should see a TWAP that can only shrink the position

5. Watch the fill progress, not the price

   The useful signal mid-TWAP is the share that has actually executed. A stalled fill count means suborders are being skipped, which tells you the market is moving faster than your ceiling allows.

   You should see a clear view of how much has filled and how much is being skipped

When it is worth it, and when it is not

Worth it: entering or exiting a position that is large relative to the market’s depth; building into a thin market where a single order would visibly move the price; exiting over a session rather than in one print.

Not worth it: any order small enough that it would clear inside the top of the book anyway. Below a few thousand dollars on a major market, a TWAP adds timing risk and saves nothing, because there was no impact to avoid.

The wrong tool: an urgent exit. A TWAP is deliberately slow and can be skipped. If you need to be flat now, close the position and pay the taker fee.

TWAP versus a scale order

They solve the same problem from opposite ends. A TWAP accepts the market’s price to guarantee progress; a scale ladder names its prices and accepts that it may not fill.

If you have a view and time, ladder it. If you have size and a deadline, TWAP it.

Common mistakes

Expecting a fee saving. Covered above. The saving is impact, not fees.

Setting a window longer than the reason for the trade. A multi-day TWAP on a thesis about the next few hours is a slow way to trade something other than what you intended.

Forgetting it is running. A TWAP keeps working after you close the tab. Check the open orders panel before assuming you are done for the day.

Using one to hide size. Every fill is public on-chain, and a run of even suborders at a fixed cadence is more recognisable than a single print, not less.