# What Is Slippage in Crypto Trading?

Updated: 24 September 2026 — Venue facts read 18 September 2026 — Session Log staff

Slippage in crypto trading is the gap between the price on screen when an order is sent and the average price it actually fills at. A market order climbs the order book until it is filled, so the thinner the resting orders, the further the fill drifts from the mid price.

## What the figures say

1. The same $100,000 market order on a BTC perpetual cost 0.02 bps against the deepest book on this page and 160.58 bps against the thinnest: $0.20 of drift against $1,605.80.
2. A hundred times the order is not a hundred times the cost. On Aster the same market order cost 0.09 bps at $10,000 and 1.03 bps at $1 million, roughly 11 times the rate per dollar.
3. A $1 million market order met no counterparty at all in the Paradex book we read, and found resting orders for only 52.8% of its size on Aevo.
4. Paradex charges retail takers 0% and still turned $100,000 into $1,605.80 of drift, against $160.00 at the dearest published round trip here.

## Quoted, word for word

> "There is no public order book. The order book displayed on the trading screen is indicative only and does not represent resting orders from other traders." — Extended documentation, quote-driven execution page, 18 September 2026. https://docs.extended.exchange/extended-resources/trading/rfq-execution
> "Market Orders Rejected: The system will reject all Market Orders." — edgeX documentation, Stock Perpetual, 18 September 2026. https://edgex-1.gitbook.io/edgeX-documentation/trading/stock-perpetual
> "Validators are responsible for storing orders in an in-memory orderbook (i.e. off chain and not committed to consensus), gossipping transactions to other validators, and producing new blocks for dYdX Chain through the consensus process." — dYdX documentation, architecture overview, 18 September 2026. https://docs.dydx.xyz/concepts/architecture/overview
> "A Sequencer coordinates first-in–first-out transaction ordering and provides soft finality, executing exchange operations and generating data consumed by other system components." — Lighter documentation, technical architecture, 18 September 2026. https://docs.lighter.xyz/about-lighter/technical-architecture-lighter-core
## Where the gap between the screen and the fill comes from

An order book is a queue of offers at each price. The top bid is the most anyone is willing to
pay at this second, the top ask the least anyone will accept, and the mid price sits halfway
between them. A price feed usually shows the mid. A market order never trades there. It crosses
the spread first, then works down the queue, taking each resting offer in turn until the whole
size has found a counterparty.

So the first part of the bill is the spread. Across the books we read in the session that closed on
22 September 2026, the median spread on a BTC perpetual ran from
0.01 bps to 11.85 bps — the same contract, the same minutes, a
difference of more than two hundred times.

The second part is depth: how much money is genuinely resting near the mid. We count the dollars
posted within ten basis points of the mid on both sides together, because that is the band an
ordinary intraday order works through. Inside it the books here held between
$245.00 and $29.3 million. What the order pays above the mid once it has
worked through that band is the slippage.

## Slippage in crypto trading, measured on ten order books

Every BTC perpetual book below was polled six times an hour until the log closed on 22 September 2026, and each figure is
the median of those readings. Rows run by what a $100,000
market order cost, cheapest first; where two books cost the same, the venues sit in
alphabetical order.

| Venue | Median spread | Resting depth within 10 bps | $100,000 market order | $1 million market order |
|---|---|---|---|---|
| [edgeX](https://pro.edgex.exchange) | 0.04 bps | $29.3 million | 0.02 bps | 0.23 bps |
| [Extended](https://extended.exchange/) | 0.12 bps | $7.8 million | 0.06 bps | 0.77 bps |
| [Hyperliquid](https://hyperliquid.xyz) | 0.12 bps | $7.0 million | 0.06 bps | 0.71 bps |
| [Aster](https://www.asterdex.com) | 0.01 bps | $14.9 million | 0.37 bps | 1.03 bps |
| [ApeX Omni](https://omni.apex.exchange/) | 0.83 bps | $1.7 million | 0.46 bps | 5.16 bps |
| [Lighter](https://lighter.xyz) | 0.01 bps | $16.2 million | 0.71 bps | 1.55 bps |
| EVEDEX | 4.12 bps | $18.5 million | 2.06 bps | 2.75 bps |
| [Aevo](https://www.aevo.xyz/) | 2.55 bps | $1.3 million | 4.06 bps | 7.37 bps |
| [dYdX](https://www.dydx.xyz) | 1.69 bps | $866,108 | 6.93 bps | 10.90 bps |
| [Paradex](https://www.paradex.trade) | 11.85 bps | $245.00 | 160.58 bps | not filled |

Two books charged the same 0.06 bps on $100,000 — Extended and
Hyperliquid alike — and one charged 0.02 bps, twenty cents on that size.
Aevo's figure for $1 million covers the 52.8% of the order that found resting orders in
the median snapshot, not the whole of it. Spreads, depth and impact on this page are ours, from
public order-book endpoints, read from one machine; no account was opened and no order placed.
Publication here is paid for.

## Why the cost grows faster than the order, and where it stops filling

The first slice of an order meets the offers nearest the mid. Every further slice reaches a
price one step further out, so the average fill price walks away from the mid as the order
grows, and cost per dollar rises with size instead of holding steady.

Our readings show it plainly. On Paradex a $10,000 market order cost
113.92 bps and a $100,000 order 160.58 bps: ten times
the order at 1.4 times the rate. On Aster, a far deeper book, the same
progression ran 0.09 bps, then 0.37 bps, then
1.03 bps — a hundredfold increase in size for about 11 times the rate.
On edgeX it ran 0.02 bps, 0.02 bps and
0.23 bps.

Past the end of the book there is nothing left to buy. A $1 million market order found no
counterparty at all in the Paradex book we read, which held a median $245.00
within ten basis points, and the snapshots record no fill for it. On Aevo the same order met
resting offers for 52.8% of its size and the rest would have waited or been cancelled. A venue
can also refuse the order outright: edgeX rejects every market order on its stock perpetuals
while the underlying market is shut
([edgeX documentation, Stock Perpetual](https://edgex-1.gitbook.io/edgeX-documentation/trading/stock-perpetual),
read 18 September 2026).

## The fee is known before the order, the fill only after

A taker fee is a rate in a published schedule. At 0.080%, the dearest crossing on
this desk, a $10,000 position opened and closed costs $16.00, and that figure is
knowable before the order is sent. Slippage is knowable only afterwards, because it depends on
what happened to be resting in the book at the moment the order arrived.

Which of the two dominates depends entirely on size. At $10,000 the fee is usually the larger
number: $10.00 of taker fees against 2.94 bps of drift, which
on that size is $2.94. At $100,000 the order changes. Paradex
charges retail takers 0% and its book still turned a $100,000 order into
$1,605.80 of drift, against $160.00 of fees at the dearest published rate
here. A schedule that reads as free is not the same thing as a cheap fill.

Funding is a third bill, charged for holding a position rather than opening one, and outside
this page.

## What each of these books does not tell you

Depth and impact are one measurement in one window. Each venue also documents limits that decide
whether an order fills at all; every line below comes from its own documentation, terms or
incident notice, read on 18 September 2026.

- **edgeX** — stock perpetuals reject all market orders while the underlying market is closed,
  and auto-deleveraging can take part of a position that is winning and close it at the
  bankruptcy price.
- **EVEDEX** — positions are written to Arbitrum in accumulated batches rather than trade by
  trade (EVEDEX trading terms and documentation, 18 September 2026), and it lists 52 perpetual
  markets against hundreds on the largest venues, with no spot market.
- **Extended** — 278 of its 325 active markets are quote-driven, where the
  displayed book is indicative only, and a market switches to reduce-only once its open interest
  reaches the cap.
- **Hyperliquid** — markets created through permissionless deployment are shown in the app
  without being reviewed by the operator, and its community vault absorbed about $4 million in
  March 2025.
- **ApeX Omni** — the shares and commodities perpetuals close at weekends and on holidays, and
  positions cannot be opened, closed or modified while they are shut; the published audits cover
  the rollup infrastructure rather than its own trading system.
- **Aster** — the chain's core contracts and node infrastructure are not open-sourced and phase
  one supports no external validators; a mispriced perpetual forced liquidations in September
  2025, which the venue reimbursed.
- **Lighter** — the free account is deliberately the slower one, at 300 ms of taker latency
  against 140 ms on the paid tier, and every order passes through a single operator-run
  sequencer, which was down for four and a half hours in October 2025.
- **Aevo** — the rollup is sequenced by one operator posting batches to Ethereum every hour, and
  a withdrawal to Ethereum costs 25 USDC and takes roughly three hours.
- **dYdX** — the chain halted for hours in the crash of 10 October 2025 and stale oracle prices
  followed the restart; transfers and withdrawals are gated for 50 blocks after any outage of
  five minutes or more.
- **Paradex** — the zero-fee retail profile carries a speed bump and lower rate limits, and the
  chain was rewound in January 2026 after a maintenance bug had liquidated positions that
  should have stood.

## What a day trader can do about it

None of this is a reason to avoid market orders; it is a reason to size them against the book
rather than against the balance. Five things follow from the readings above.

Read the depth column before the fee column once a position passes about $50,000, because that
is where the fill starts to outweigh the schedule. Split a large order rather than sending it in
one piece: the book refills between slices. Use a limit order when the entry can wait, and take
queue risk instead of price risk. Check that the book on screen holds real resting orders. And
check the hours, because a perpetual on a share or a currency can reject market orders outside
the session of its underlying market.

Two pages on this desk carry the numbers behind this one. The venues
[ordered by the book we measured](/best-crypto-exchange-for-day-trading) sets out depth, spread
and impact side by side, and the [fee schedules compared](/day-trading-fees-compared) do the
same for the published rates. Every contract named here is a leveraged perpetual, and a fill
that goes against a position at a triple-digit multiple costs far more than either bill.

## FAQ

### Is slippage the same as the spread?

No. The spread is the distance between the top bid and the top ask before the order is sent; slippage is everything the order pays beyond the mid once it has worked through the queue. The spread is part of it. On the books read to 22 September 2026 spreads ran from 0.01 bps to 11.85 bps.

### How do you calculate slippage on a trade?

Take the average price the order filled at, subtract the mid price at the moment it was sent, and divide by that mid price. The answer is usually quoted in basis points, where one basis point is a hundredth of a percent. At 6.93 bps a $100,000 order gives up $69.30.

### What is a reasonable slippage tolerance for crypto?

It depends on the book, not on a rule of thumb. A tolerance set tighter than the book can deliver cancels the order; one set far wider than it invites a poor fill. Our readings on a BTC perpetual ran from 0.02 bps to 160.58 bps on the same $100,000 order.

### Do limit orders have slippage?

A limit order cannot fill worse than its limit price, so it carries no price slippage. What it carries instead is queue risk: it may fill in part, or not at all, and the market may leave without it. On several venues here a resting order also earns the maker rate rather than paying the taker rate.

### Why is slippage worse on a decentralised exchange?

Often it is not the design but the size of the book. The thinnest book on this page held $245.00 within ten basis points while the deepest held $29.3 million, and both settle on a chain. Pool-priced venues are a separate case, since they quote from a pool rather than from resting orders.

### Can slippage work in your favour?

Yes, and it is called positive slippage. If the price moves toward the order between the moment it is sent and the moment it is matched, the fill lands better than the quote. It is the same mechanism running the other way, and it is not something to plan a session around.

### How much slippage is normal on Bitcoin perpetuals?

On a deep book, almost none at retail size: a $10,000 order cost 0.02 bps on the deepest book on this page, read in the session ending 22 September 2026. The same order cost 113.92 bps on the thinnest. Normal is a property of the venue and the size, not of Bitcoin.

### Does slippage get worse when the market is volatile?

Usually, because market makers widen their quotes and pull size when prices move quickly, so the band near the mid empties out. Our medians come from readings taken in one calm window — 52 behind the shortest run here and 128 behind the longest — and they do not describe a crash.

### Does order size change how much slippage you pay?

Yes, and not in proportion. On Aster the same market order cost 0.09 bps at $10,000, 0.37 bps at $100,000 and 1.03 bps at $1 million: a hundredfold jump in size for roughly 11 times the rate per dollar. Bigger orders reach further into the queue.

### Do stop-loss orders suffer slippage?

A stop usually triggers a market order, so it pays whatever the book holds at that second, and stops tend to trigger when books are thin. On a venue that rejects market orders outside session hours, a stop on a share perpetual may also sit unfilled until the underlying market reopens.

### How can you reduce slippage when day trading crypto?

Size the order against the resting depth rather than the account, split large orders so the book refills between slices, use limit orders whenever the entry can wait, and check whether the displayed book holds real resting orders. Our depth column runs from $245.00 to $29.3 million.

### Is slippage included in the trading fee?

No. A fee is a published rate the venue charges on the notional; slippage is paid to whoever was resting in the book, and no schedule lists it. The Lighter default account and the Paradex retail profile both post 0%, and their books cost quite different amounts to cross.

## About this entry

This site is paid placement. Every figure on it carries the date it was read and the source it came from, terms are quoted rather than paraphrased, and a figure that changes moves the date with it. Corrections: editorial@coinlen.com.

Session Log staff, 24 September 2026

This site is paid placement.