What it is
Ask what is the order book in trading and the answer is simple: it is every limit order that has not yet been filled, sorted by price. The bid side shows where buyers are resting, the ask side shows where sellers are resting, and the size at each price is the depth of market. Together they tell you where liquidity currently sits.
It is a book of intentions, not of results. The trade book, the tape, or a footprint chart shows where size actually traded. The order book shows what is still standing there and can therefore still disappear. To know whether anyone was genuinely willing to transact, read the trades; to know how fast liquidity is stacked or pulled, read the book.
Most order book trading strategies come down to reading changes in that liquidity: size clustering at a key price, the cluster vanishing just before price arrives, depth tilting to one side, or liquidity thinning as price runs. The mechanics are the same in any central limit order book, including gold futures.
What it looks like in order-by-order data
On an order-by-order feed the book stops being a static table and becomes a stream of events. A single anonymous order is inserted at a price, its size is modified, it is cancelled, or it is filled. Line those events up in time and you get the lifecycle of orders rather than one number.
What you can actually observe:
- Liquidity walls: size at one price clearly larger than neighbouring levels on the same side.
- Cancellations: resting size disappearing as price approaches, leaving the wall thin.
- Refills: new resting orders appearing at or near the same price shortly after a print.
- Liquidity retreat: depth pulling back on one side so the top of book goes thin.
Mind the wording. A refill is an inference, not an observation about a person: you see new anonymous orders at the same price, not the same trader restocking. Iceberg orders and other hidden liquidity can only be inferred from repeated refills and how executed volume compares with visible size, never treated as confirmed fact.
Common misconceptions and limits
The first misconception is treating the order book as a trade book. Resting-size imbalance across ten levels is not the same thing as the per-price aggressive buy and sell imbalance on a footprint chart, and the two should not be used interchangeably.
The second is expecting the book to call direction. Resting orders can be pulled in an instant, and a large bid sitting at a price does not mean anyone will transact there. Thin liquidity or a lopsided book describes current conditions, not a forecast.
The third is scope. Public order-level data is anonymous and excludes orders outside the exchange, so hidden liquidity can only be inferred. A snapshot is exactly that: judging the book without its time series is where most over-reading happens.
What the GC research says
Everything below applies only to COMEX gold futures (GC), comes from TradeWhy research, and reflects data through 2026-08-31. Do not carry these numbers over to other products.
- Liquidity walls at key levels (study W16, n=297, ten-level book data from 2024-07 through 2026-08): dividing wall size by the median ten-level size, the probability of a break within 30 minutes is 49 for small walls, 48 for medium walls, and 31 for large walls (1.3 times the median or more). The bigger the wall, the smaller the share that breaks inside that window. It is single-period data with the independent-period review passed.
- Cancel-before-touch signals and post-touch refill ratios: this did not hold up in the GC data, so no stable signal. Treat it as failed verification.
- Ten-level resting-size imbalance used as the primary signal: this did not hold up in the GC data either, failing repeatedly. It only works as a weak secondary input, worth roughly 8 percentage points of marginal improvement.
- Thinning liquidity read as bearish: this did not hold up in the GC data, with sign flips in the holdout sample.
Taken together, what survived review on GC is structural statistics about wall size and subsequent break rates, not directional reads taken straight off book imbalance or thinning depth.
How to see it in TradeWhy
TradeWhy is an AI market-intent analysis and strategy validation tool built for intraday CME Group futures traders, starting with COMEX gold (GC) and CME Nasdaq (NQ). It puts real-time macro information, price response, and L3 order-by-order behaviour side by side to explain what the dominant force is doing right now. Recurring behaviour can be organised into strategies, which enter the strategy portfolio only after historical and forward validation.
For the order book specifically, L3 is the relevant layer: candles are the price result, L2 is a photo of resting size, and L3 records each exchange-visible order from insertion and modification through cancellation to execution. Whether resting orders are genuinely willing to trade, whether liquidity is retreating, whether size keeps being refilled after prints, and whether heavy volume actually moved price all become observable.
TradeWhy explains market behaviour and helps users research and validate strategies. It does not give buy or sell recommendations and makes no promises about returns; the final trading decision is yours. It also does not show trader identities or orders outside the exchange, because that data is not available.
FAQ
What is the difference between the order book and the trade book?
The trade book records orders that already matched, while the order book holds orders still resting. The tape tells you which prices were actually hit; the book tells you where liquidity sits and how fast it can be pulled. Use them together, not as substitutes.
What is the order book in trading?
It is the full set of unfilled limit orders on the bid and ask side, sorted by price, with the resting size at each level. It is a picture of intent and available liquidity, not a record of executed business.
How do you read the order book?
Start with the depth profile and look for size clearly larger than neighbouring levels on the same side, then watch how that size behaves as price approaches: does it hold, get pulled, or get refilled after a print? Read it as a time series of liquidity.
Can the order book give a directional signal?
Not on its own. Resting orders can be cancelled instantly, and a large bid does not mean anyone will transact there. In the GC research, ten-level resting-size imbalance as a primary signal and thinning liquidity as bearish both failed verification; at best they are weak secondary inputs.
Can you spot iceberg orders in the order book?
Only by inference. You see new anonymous orders appearing repeatedly at the same price, or executed volume exceeding visible size. Public order-level data is anonymous and excludes orders outside the exchange, so hidden liquidity cannot be treated as confirmed fact.
Written by AI from TradeWhy's metric definitions, research findings and published product statements, then rule-checked. Research figures come from TradeWhy's study of COMEX gold futures (GC), data through 2026-08-31, and apply to GC only. Analysis only, not investment advice.