A book is the most honest object on an exchange: every line is an order somebody has actually committed to. It is also the most over-interpreted, because a commitment can be withdrawn in a millisecond and most of them are.
One side lists every price at which somebody has agreed to sell, and how much. The other lists every price at which somebody has agreed to buy, and how much. They are sorted so that the best offers face each other in the middle: the lowest price anyone will sell at, and the highest price anyone will buy at.
The distance between those two numbers is the spread. It is not a fee and nobody charges it, but it is a real cost: if you buy at the best ask and immediately sell at the best bid, you are down by the spread before anything has happened. On a liquid pair it is a rounding error. On a thin one it can exceed everything a venue charges you in a month.
Every line in the book is an obligation, not a prediction. It says what somebody will do at a price, not what they think the price will be.
A market order does not execute at one price. It executes against the book, line by line, starting at the best available and walking outwards until the whole quantity is filled. If the top line does not hold enough, the rest comes from the next line, then the next, at progressively worse prices.
The average of those prices is what you actually paid. The gap between it and the price you saw on screen is slippage, and it is entirely a function of your size relative to the depth sitting there. Two traders sending the same order one second apart can get different average prices for no reason other than that one of them was larger.
This is also the practical reason to prefer a limit order when the size is meaningful. A limit order cannot slip: it fills at your price or it does not fill. What it can do instead is fill partially, which is a different problem and a much more manageable one.
Four readings are extremely common and none of them is supported by what the object actually is.
A visible wall of size at a price looks like support, and it is treated as such by a great many people watching the same screen. It can also be withdrawn instantly by whoever placed it, and orders placed specifically to be seen and then pulled are a well-documented behaviour on every venue that has ever existed. A wall tells you somebody wants that order to be seen. Nothing more.
More size on the bid than on the ask feels bullish. It is a snapshot of resting intentions, taken at an instant, in a book that turns over many times a minute. The correlation between visible imbalance and the next move is weak, unstable, and thoroughly arbitraged on any liquid pair.
Iceberg orders display a fraction of their real size. Some venues support them explicitly, and on those the visible book systematically understates the true depth. You are reading an incomplete object and there is no way to know by how much.
There is no participant identity in a book. A thousand small orders and one large one broken into a thousand pieces look identical, and the difference between them matters enormously.
Three uses that hold up, and they are all about cost rather than direction.
Before sending an order, look at the cumulative column and find the level at which your quantity would be filled. That number is your realistic entry price, not the one at the top. If it is materially worse, the order is too large for the moment and should be split or made passive.
The same position can be cheap on one pair and expensive on another, purely because of depth. A pair whose book empties two levels down is not a pair to run size on, whatever its chart looks like.
A book that thins out and a spread that widens are the two most reliable signals that execution has become expensive. They usually happen together, they usually happen before volatility is obvious on a chart, and they are a good reason to postpone a market order by a few seconds.
The object is the same everywhere, but four features of crypto change how it behaves, and each one has a practical consequence for what your order costs.
There is no opening auction and no closing print. Depth follows the working hours of the regions where the largest participants sit, which means the same pair can be comfortable to trade at one hour and expensive at another, with no news and no chart pattern to explain it. A trader who always works late is systematically paying more slippage than one who does not, and has usually never noticed.
The minimum price increment and the minimum quantity are set per pair by each venue. On a pair with a coarse tick the spread cannot go below one tick however much competition there is; on a pair with a fine tick the book can look deep while every level holds almost nothing. Reading the number of levels without reading the tick tells you very little.
Much of what sits in a crypto book is placed by programs that cancel and replace continuously, and that withdraw entirely when their own risk limits are hit. This is why depth evaporates in seconds during a violent move: it was never a commitment to be there, it was a commitment to be there while conditions held.
The same asset quoted against a stablecoin and against a currency has two separate books, usually with very different depth. The stablecoin book is normally the deeper of the two by a wide margin, which is the practical reason most size trades there even when the trader thinks in a national currency.
None of these four is visible on a chart. All four decide what a given order costs, which is why the book is worth reading even by people who never trade from it.
The book on your screen belongs to one venue. Every other venue trading the same pair has its own, with its own depth, its own spread and its own best prices, and none of them is connected to the others. There is no consolidated book in crypto the way there is in some regulated equity markets, and nothing obliges two venues to show the same price at the same instant.
Most of the time arbitrage keeps them close, because a persistent gap is free money for whoever closes it. But close is not identical, and the gap widens exactly when it matters: during fast moves, when the arbitrage capital is already committed elsewhere and the transfers needed to close a gap take longer than the gap lasts.
Two practical consequences follow. The first is that a price quoted on an aggregator is a summary of several books and not a price you can trade at; the price you can trade at is the one in the book in front of you. The second is that depth on a pair is not a property of the asset, it is a property of the venue: the same pair can be liquid in one place and thin in another on the same afternoon.
When somebody says a price, ask where. The question sounds pedantic until the first time the answer costs you money.
Next to the book, most terminals show the tape: the trades that actually happened, with their size and their side. The book shows intentions, the tape shows facts, and reading them together is more informative than either alone.
A book with heavy bids and a tape full of sells tells you the resting orders are being consumed rather than respected. A thin book with a quiet tape tells you nothing is happening and the spread you see is what you will pay. Neither of these is a signal to trade; both are information about what a trade would cost right now.
No. Nobody charges it and it does not go to the venue. It is the distance between the best buy price and the best sell price, and it is a real cost only in the sense that entering and exiting immediately costs you that distance. Trading fees are charged on top of it, separately, by the venue.
Because a market order walks the book. The displayed price is the best available line; if your quantity exceeds what that line holds, the remainder fills at the next lines, which are worse. The average of all the fills is your price. This is slippage, and it grows with your size relative to the depth.
It means somebody has placed an order there and is willing to be seen doing it. Resting orders can be cancelled instantly and frequently are. Treating a visible wall as a floor or a ceiling is one of the most common and most expensive readings of a book.
It is better against slippage and worse against certainty. A limit order fills at your price or not at all, so it cannot slip, but it can leave you unfilled while the move happens without you. Which risk matters more depends entirely on the strategy, and neither answer is universally right.
Treat it as a summary, not as something you can trade against. An aggregator adds up books from venues that do not share order flow, so the total it shows is not available to any single order. The depth that decides your fill is the depth of the one book your order will reach.
Because depth is bought rather than inherited. A venue attracts resting orders with its fee structure, its reliability and the size of the flow already there, and those advantages compound. It is why liquidity concentrates and why a pair can be comfortable in one place and unusable in another on the same afternoon.
It means a cheaper trade, which is not the same thing. Depth reduces slippage and nothing else. It does not make a direction more likely to be right, and a very liquid pair can move violently in the time it takes to read this sentence.
Not on its own. A tight spread at the top of the book tells you what one unit costs, not what your size costs. A venue can show an attractive spread on a single line and hold almost nothing behind it, in which case a real order walks straight through and pays far more than the spread suggested. Read the spread and the depth together, or read neither.
Because the people providing those resting orders withdraw them when the risk of being filled at a stale price rises. Depth is thinnest exactly when the market moves fastest, which is why slippage is worst at the moment most traders feel most compelled to use a market order.