How Limit Order Books Actually Work

Most traders learn markets through charts. But a chart is a rear-view mirror: it shows you what prices were. The place where prices are actually made — where every tick originates — is the limit order book. If you trade electronically and you don't understand the book, you are operating machinery you've never seen.

The book is a queue, not a picture

A limit order book is, at its core, a pair of sorted queues. On one side sit the bids: resting limit orders to buy, sorted from the highest price down. On the other side sit the asks (or offers): resting limit orders to sell, sorted from the lowest price up. The highest bid and the lowest ask define the inside market, and the distance between them is the bid-ask spread.

Everything you see on a price chart is the residue of one event, repeated millions of times: an incoming order crossing the spread and matching against a resting order on the other side. When a marketable buy order lifts the best offer, the trade prints at the ask. When a marketable sell hits the best bid, it prints at the bid. Price "moves" when one side of the inside market is fully consumed and the next price level becomes the new best.

This is why microstructure researchers describe price not as a signal that trading reveals, but as an outcome that trading produces (O'Hara, 1995). The book is the production line.

Price-time priority: the rule that governs everything

Nearly every major electronic venue — including CME Group's Globex platform for futures — matches orders under some form of price-time priority (often called FIFO, first-in-first-out). The logic has two layers:

1. Price priority. Better-priced orders trade first. A bid at 5001.00 always executes before a bid at 5000.75, no exceptions.
2. Time priority. Among orders at the same price, the one that arrived first executes first.

The consequences are profound. Queue position is an asset: a limit order that sits first in line at a price level captures the spread when it fills, while an identical order at the back of the same queue may never trade — or worse, may fill only when the market is about to trade through the level (a phenomenon we examine in our article on adverse selection). Sophisticated participants invest heavily in earning and protecting queue position, which is one of the quieter reasons latency matters even for strategies that are not "high-frequency" in any meaningful sense.

Some products use alternative allocation rules — pro-rata matching, common in certain interest-rate and options markets, allocates fills proportionally to order size rather than arrival time. Knowing which allocation algorithm governs your product is not trivia; it changes optimal order-placement behavior entirely.

Depth, liquidity, and what the book actually tells you

Market depth is the quantity resting at each price level beyond the inside market. A book with 500 contracts within two ticks of the mid-price absorbs a large market order with minimal price movement; a thin book gaps. This is the practical meaning of liquidity: not volume traded, but the book's capacity to absorb order flow without repricing.

Two caveats keep practitioners honest. First, displayed depth is not committed depth: resting orders can be canceled in microseconds, and studies of modern markets consistently show that the large majority of submitted orders are canceled rather than filled. Second, not all liquidity is displayed — iceberg orders show only a fraction of their true size, replenishing as slices execute. The visible book is an honest but incomplete map.

The matching engine: where determinism lives

The matching engine is the exchange's central computer that maintains the book and executes the priority rules. Its defining property is determinism: given the same sequence of inbound messages, it produces the same trades, every time. Orders are processed strictly sequentially — there are no ties, only ordering.

For anyone building or operating trading systems, this determinism is the foundation of everything downstream. It is what makes a synchronized audit trail possible: every order acknowledgment, modification, cancellation, and fill can be reconstructed message by message. U.S. regulators lean on exactly this property — CFTC and NFA recordkeeping expectations assume that a firm can reproduce the lifecycle of any order it sent. Infrastructure that timestamps and retains that lifecycle (the approach we take at GIDEON with immutable, exportable execution logs) is simply the firm-side mirror of what the matching engine already does exchange-side.

Why this matters even if you never look at the book

You might trade from signals, from a model, or from a discretionary chart read — and never open a depth-of-market display. The book still governs your outcomes:

• Your fills are queue events. Whether a limit order executes is a function of queue position and order-flow arrival, not of whether price "touched" your level on a chart.
• Your costs are book events. Slippage on a market order is exactly the depth you consumed beyond the inside price.
• Your risk is book-shaped. In stressed conditions, depth evaporates before price moves — the book thins first, then gaps. Systems that monitor only price discover stress one step too late.

The limit order book is the market's true interface. Charts summarize it; the book is it.

References

• Harris, L. (2003). Trading and Exchanges: Market Microstructure for Practitioners. Oxford University Press.
• O'Hara, M. (1995). Market Microstructure Theory. Blackwell.
• Hasbrouck, J. (2007). Empirical Market Microstructure. Oxford University Press.
• Gould, M. et al. (2013). "Limit Order Books." Quantitative Finance, 13(11).