Look at the book. To buy right now you have to meet the lowest price a seller is asking; to sell right now you have to take the highest price a buyer is bidding. Those two numbers are never the same, and the gap between them is the spread. It appears on no fee schedule and on no statement, yet the moment you choose to trade immediately, you have already crossed it once.
This is not a piece about order tactics. It costs out one thing: why the gap exists, how it differs from fees and from slippage, which routes make it expensive, and a way to measure what you actually paid that needs no market data at all. Every price and percentage below is an assumed figure used to demonstrate the method, not a reading from any account or any live book.
The gap between bid and ask
An order book has two sides. The highest resting buy order is the bid; the lowest resting sell order is the ask. The ask always sits above the bid, because if it did not, those two orders would already have traded with each other. The distance between them is the spread.
With some assumed numbers:
| Book (assumed) | Value | What it means for you |
|---|---|---|
| Best bid | 100.00 | Sell right now and this is your price |
| Best ask | 100.05 | Buy right now and this is your price |
| Spread, absolute | 0.05 | The distance between them |
| Spread, relative | about 0.05% | Divided by the mid price |
Assumed figures for illustration. A real book moves continuously and no fixed number describes it.
The last two rows are the point. That 0.05% is not money anyone collects from you as a charge. It is the drop you accept in exchange for not waiting. Buy at 100.05, sell straight back at 100.00, and the trip has cost you 0.05 without a single decision going wrong — before any fee is counted. Worth noting what that example is not: 0.05 on a 100 price is a figure chosen to keep the arithmetic readable. On a major pair it is usually far narrower, and on a thin one it can be many times wider.
Why the gap is there at all
The spread is not a toll the venue installs. It is what someone earns for standing ready to trade with you at any moment. A resting order carries two risks: the price may move against it while it waits, and whoever comes along to take it may know something the poster does not. Both have to be paid for, and the payment shows up as a bid set slightly lower and an ask set slightly higher.
So the direction never varies: the side that takes pays the spread, the side that rests earns it. Wikipedia’s treatment of the bid–ask spread puts it the same way, framing the gap as a measure of liquidity and of transaction cost, and as the price of immediacy specifically.
One useful consequence follows. The more participants and the denser the resting orders, the narrower the gap. Thin pairs, quiet hours, and orders large enough to eat through several levels all widen it noticeably. It is why the same amount of money can cross a major pair without your feeling anything, and lose a visible slice on a pair that barely trades. If you want to see this rather than take it on trust, open the book on two pairs of the same coin side by side and read the top level on each.
Spread, fee, slippage: stop merging them
These three get used interchangeably, and some explainers state flatly that the spread is a fixed charge set by the platform. In foreign exchange and CFDs that is partly fair — those venues often do mark up the quote, and by an amount they choose. Carried over to a crypto order book it stops being true. The spread there is made of other people’s resting orders, and the venue is not taking a percentage out of it.
| Spread | Fee | Slippage | |
|---|---|---|---|
| Set by | Resting orders in the market | The venue’s schedule | Book depth at the moment you send |
| Occurs | Whenever you trade immediately | On every fill | Only when the fill misses your expected price |
| Direction | Always against you | Always an outflow | Can go either way |
| Visible on a statement | No | Yes | No |
| Size depends on | Liquidity, hour, pair | Your tier and discount settings | Order size against the book |
One line to keep them apart: the spread is what you pay to cross the gap, slippage is what you pay for widening it yourself, and the fee is what the venue charges on top. Small orders meet the first and not the third; only orders big enough to clear several levels bring slippage into it. Fees are independent of both, and saving on them is a separate exercise, laid out in the fee guide.
There is a practical corollary, and the counting matters: a round trip pays two fees but one spread. Look again at the book above. Buying at 100.05 costs you half the gap against the 100.025 midpoint; selling at 100.00 costs you the other half. Out and back is one whole spread, not two — while the fee is charged separately on each of the two fills. Getting this backwards doubles the number you are trying to measure. The fee half of that arithmetic is in what a round trip really costs.
Where the spread costs you most
Ordered from cheapest to most expensive, by how wide the gap tends to be and whether you get a choice about crossing it:
- Resting a limit order and waiting. You are on the earning side. You do not pay the spread; you are, in effect, paid it. The cost is that you may not get filled.
- Taking the market price on a major pair. The gap is narrow and a small order barely registers it. Fees dominate here.
- Taking the market price on a thin pair. The gap can be an order of magnitude wider, and a modestly sized order clears several levels, so spread and slippage arrive together.
- Quote-based conversion. Usually no fee line on screen, because the cost is built into the quoted price. Not seeing it is not the same as not paying it.
- Peer-to-peer and over-the-counter. The price comes from a counterparty, and its distance from the mid can exceed any book spread. It usually goes by the name premium.
The same gap is trivial for one trader and expensive for another
The spread is charged per crossing, so what it means to you depends almost entirely on how often you cross. Put three habits side by side with assumed figures — 0.05% of spread per side, 0.075% of fee per side, 10,000 units moving each way:
| Habit (assumed) | Round trips per year | Yearly cost of crossing | As a share of fees paid |
|---|---|---|---|
| Buy and leave it alone | 1 | about 10 units | about 67% |
| Rebalance monthly | 12 | about 120 units | about 67% |
| In and out most weeks | 150 | about 1,500 units | about 67% |
Assumed values throughout. Real spreads move with the pair and the hour; real rates depend on your tier and discount settings.
The last column is the finding: it does not change with frequency. Which means the spread and the fee are the same order of magnitude. One of them is published, and you can lower it with tiers and discount settings. The other is not published, and the only levers are which pair you trade and which side of the book you sit on. Plenty of people work hard on the first and have never once measured the second.
The first row carries a quieter conclusion. If the plan was always to hold, the spread is paid twice in total and is not worth agonising over; what actually mattered was whether the entry was taken during the thinnest hour of the week. The last row inverts it: for that trader the first move is not a fee tier at all, it is cutting the trips that were never needed, because each one removed takes a spread and a fee with it. To see the base you are working against, the annual savings estimator sizes the fee half first.
Measuring what you actually paid
This needs no market data. Take a small amount, move it the way you normally would, move it straight back, and look at exactly one number: how much less you are holding than when you started.
One spread, because you crossed half of it going out and half coming back. Two fees, because each fill is charged on its own.
Walk it through with assumed numbers. Start with 1,000 units, go out and come back, end with 997.5: a loss of 2.5 units, or 0.25%. If your rate is 0.075% per side, the two fills account for 0.15%. Subtract, and 0.10% is left. That remainder is what the price took — the spread, plus any slippage, plus whatever the market did while you were between the two fills.
Which is the honest limit of the method, and it is worth stating plainly rather than burying: the remainder is an upper bound on the spread, not a measurement of it. Four things live in there and this test cannot separate them:
- The spread — the thing you were trying to size.
- Slippage, if either order was large enough to clear more than the top level.
- Ordinary price movement between the two fills. Seconds are enough on a volatile pair, and this one can land either way — it can make a wide spread look narrow.
- Rounding and any markup inside a quote, if the route was a conversion rather than the book.
Two things make the estimate tighter. Turn the trip around fast, so price movement has less room to contaminate it. And keep the order small enough that it fills at the top level, which removes slippage. Do that and the remainder is mostly spread. If you want the spread itself rather than an upper bound on it, there is only one way: read the bid and the ask off the book before you send the order and take the difference. The round trip answers a different question — what did this route cost me end to end — and that is usually the question worth asking.
Where the method really earns its keep is comparison rather than measurement. Run the same amount through two routes at the same time, out and back, and keep whichever leaves you holding more. That sidesteps every difference in wording — zero fee, no commission, free conversion — without needing to understand anyone’s fee structure, and the contaminants roughly cancel because both routes face the same market over the same minute.
To get the fee side right first, the fee reconciliation helper does that half. If the loss is an order of magnitude away from anything fee-shaped, you are not looking at a fee problem at all, and the reverse lookup is identifying an unfamiliar charge.
What narrows it
- Trade where the gap is narrower. The same coin often has several pairs. Picking the one with real depth usually beats any technique.
- Rest a limit order when you can afford to wait. This is the only move that flips the spread from a cost into something you collect, and it tends to come with the better fill rate as well; the mechanics are in maker versus taker.
- Break large orders up and avoid dead hours. This targets the slippage half, though in practice the two show up together.
- Make fewer trips. The spread is charged per crossing, so cutting unnecessary round trips beats shaving the cost of each one.
Sequence matters: see what you are paying in total before deciding which lever to pull. Putting list rate, discount and rebate side by side is quickest with the cost stacking comparator.
Two beliefs worth dropping
“The venue sets the spread, so I should shop for a venue with a low one.” On a crypto order book the gap is set mostly by that pair’s liquidity, and two pairs on the same venue can differ by a factor of tens. Changing pairs usually does more than changing venues.
“It says zero fees, so this route is free.” Zero fees means the fee row is zero. It says nothing about markup inside the price. The test is still the round trip: count what you hold at each end and ignore the wording on screen. Both beliefs share a root, which is treating the spread as something a venue charges you rather than something the market quotes at you — and the fix for both is to check the book rather than the marketing.
Common questions
- Is the spread just another name for the fee?
- No. A fee is charged by the venue on the value of a fill, appears on the schedule and shows up on your statement. The spread is the distance between the bid and the ask, formed by other traders’ resting orders, with no percentage taken by the venue and no row on any statement. A round trip meets both — two fees and one whole spread — and only their sum is the real hurdle.
- Is the spread a fixed number?
- No. It tracks liquidity: narrower on major pairs and in active hours, visibly wider on thin pairs, in quiet hours, and when an order is large enough to clear several levels of the book. Any hard number you read is a snapshot of one moment.
- If I always use limit orders, do I avoid the spread entirely?
- If your order is the one being taken, you are on the earning side and are effectively paid the spread rather than paying it. The trade-off is that a resting order is not guaranteed to fill; if the price never comes back to your level, it simply sits there.
- Conversion shows no fee. Is it cheaper than the order book?
- Not necessarily. Quote-based conversion normally builds its cost into the price, so an empty fee row proves nothing. The reliable comparison is to run the same amount through both routes at the same time, out and back, and compare what is left — not to compare the rates shown on screen.
- How do I find out what the spread actually cost me?
- Use the round trip: move a small amount out, move it straight back, work out how much less you hold, then subtract the two fees your tier implies. Note the counting — a round trip pays two fees but crosses only one whole spread. What is left is an upper bound rather than a clean measurement, because slippage, ordinary price movement between the two fills and any markup inside a quote are all sitting in the same remainder. Keep the order small and the turnaround fast and it is mostly spread. For the spread itself, read the bid and ask off the book before you send.