Before prediction markets widely adopted automated market makers, serious trading ran on order books—the same machinery behind stock exchanges, futures pits, and many regulated event-contract venues today.
This chapter builds the classic mental model: bids, asks, matching, and price discovery without crypto jargon. Regulated U.S. event contracts and large sports exchanges still run this way; crypto-native venues often pair a book with an AMM layer later in the module—but the central limit order book vocabulary is the baseline.
What an order book is
An order book is a live list of intentions to trade. Bids are prices and sizes where buyers are willing to purchase; asks (offers) are prices and sizes where sellers are willing to sell. The exchange does not set one official price. It matches compatible orders when a bid is at least as high as an ask, or when two parties cross voluntarily.
Think of it as an auction running continuously, not once at the closing bell.
Core order types
Limit orders rest at your chosen price until filled or canceled—for example, “buy 2,000 YES at $0.42” on a Fed-rate contract. Market orders fill immediately at the best available prices on the other side of the book, walking through the ladder if necessary. Cancel and replace updates resting liquidity without forcing a trade, which is how market makers tighten quotes after news.
Venues that support immediate-or-cancel or fill-or-kill variants help traders who need size now but want to cap how much of the book they consume. Prediction markets use the same logic as equities; only the instrument label changes (YES shares at 42¢ instead of a stock ticker).
A market sell walks down the bid ladder; a market buy walks up the ask ladder. If your order is larger than size at the best price, you receive partial fills at each level until done or until an IOC/FOK rule stops the remainder. That is why headline “best ask” is only the first step in a large trade.
Matching rules (price–time priority)
Most central limit order books use price–time priority: the best price wins (highest bid, lowest ask), and at the same price the earlier order fills first. That rewards liquidity providers who quote tight markets and makes queue-jumping fair only through a better price.
Suppose two bids sit at $0.50—1,000 shares posted at 10:00:01 and 2,000 shares at 10:00:03. A seller hits the bid with a 1,500-share market order. The first 1,000 fill against the earlier order; the remaining 500 fill against the second. Time priority mattered as much as price.
The spread and mid-price
The best bid is the highest buy price in the book; the best ask is the lowest sell price. The spread is ask minus bid—friction plus uncertainty. The mid is the average of best bid and best ask, often the headline “market price” in media.
If the mid is $0.50 but the bid is $0.48 and the ask is $0.52, you pay 52¢ to buy immediately and receive 48¢ to sell immediately. That four-cent round-trip is a real cost before fees, even when journalists quote 50%.
| Metric | Example | How to read it |
|---|---|---|
| Mid | $0.50 | Roughly 50% implied if the spread is tight |
| Buy (lift ask) | $0.52 | You need a stronger view than 52% to buy YES profitably |
| Sell (hit bid) | $0.48 | You forgo upside above 48¢ when you exit now |
Your profit and loss use bid and ask, not the mid alone—the next chapter applies that directly to event contracts.
How price discovery happens
Price discovery is the process of finding clearing levels where supply meets demand. News arrives; informed traders lift offers or hit bids. Resting liquidity is consumed; market makers repost quotes. A new equilibrium bid and ask cluster forms. No single oracle sets odds—the tape of trades and quotes is the forecast.
On a major election headline, you might see a tight 49¢/51¢ market jump toward 58% within seconds as asks are lifted, then watch makers widen the spread and reload size over the next few minutes until prices converge with related venues. The path matters as much as the destination.
Roles in the ecosystem
Retail traders express views and often cross the spread. Market makers quote both sides, earn the spread, and manage inventory risk. Arbitrageurs align this market with related contracts on other venues or with bundled YES/NO relationships. The exchange runs the matching engine, surveillance, settlement, and rulebook.
Prediction markets fail without market makers on order-book designs. An empty book produces a meaningless mid—and a small aggressive buyer can move the displayed probability cheaply. Depth is not a cosmetic feature; it is what makes the mid informative.
Centralized vs on-chain order books
Traditional centralized CLOBs offer microsecond-to-millisecond matching, low per-trade fees, and regulatory reporting—with finality on the exchange ledger. On-chain CLOBs trade transparency and settlement on the chain for block-time constraints and gas costs. Many crypto prediction products skipped full on-chain books and used AMMs instead; regulated U.S. event contracts increasingly resemble traditional CLOB plus compliance.
For forecasters the practical split is not “crypto vs TradFi” but whether a two-sided quote exists at your size. A fast matcher with no makers is still an empty market. A slower on-chain book with tight human quotes can still be the best place to express a view.
What you see in a UI
A typical ladder shows bid prices and sizes on one side, ask prices and sizes on the other, and last trade plus spread in the center. A depth chart plots cumulative size at each price; steep walls mean moving the market costs more (slippage). If the best ask is $0.52 for 500 shares, then $0.53 for 1,000 more, a 3,500-share buy might average near $0.54 even though the “headline ask” was 52¢ for the first slice.
Hidden liquidity and data habits
Some venues support iceberg or hidden size—displayed quantity smaller than true resting size. Others show only aggregated depth. The lesson for prediction markets is unchanged: simulate your trade against visible depth, and treat last trade as confirmatory, not authoritative, when the spread is wide.
Last trade, mid, and volume-weighted average price can diverge after a single aggressive sweep. Habit: read bid, ask, spread, then depth, then time since last print. That ordering prevents mistaking a one-lot spike for a consensus move.
Link to prediction markets
As in the opening module, price ≈ probability on a standard $1 binary YES contract. Buying YES at the ask near $0.63 means you pay 63¢ for $1 if YES wins. The mid is what headlines cite; your execution is bid or ask. A four-cent spread on a 50¢ contract is a large relative friction compared with liquid large-cap equities—prediction markets are often wider, and that spread is part of the forecast you must clear.
Use limit orders when you are not urgent; use market orders when news dominates and you accept slippage. Before trading, check best bid, best ask, and spread in cents; compare size at the touch to your order; note whether the last trade is stale on a live event; and glance at depth within a few cents of the mid.
Co-location and speed (brief)
Professional market makers on regulated venues sometimes use co-located servers to minimize latency. Retail forecasters do not need microsecond optimization, but you should know that the touch you see may already be stale relative to maker models when news is breaking. That is another reason limit orders fail to fill at the old mid during spikes.
On election night, treat the book as a live sensor: widening spread often means makers fear adverse selection; vanishing depth at the touch may mean a headline is still being digested; sudden convergence across venues means arbitrage capital arrived. The ladder is not only execution infrastructure—it is a readable uncertainty signal.
Opening and closing auctions
Some venues run opening auctions to set the first mid without a burst of market orders, and closing procedures near resolution to reduce manipulation into settlement. If your market uses them, continuous-book skills still apply between auction windows—you simply respect discontinuous times when quotes reset.
Practice exercise (mental)
Pick a live contract, write down best bid, best ask, spread, and size at touch without looking at mid. Mid should be derivable; if you cannot derive it, the UI trained you wrong. That thirty-second habit pays off on every later chapter. Repeat it after every major headline before you trust a screenshot someone sent you.
Key ideas
Order books match bids and asks, not opinions in a vacuum. Spread is the cost of immediacy; depth is the cost of size. Price–time priority is the default fairness rule. Market makers and arbitrageurs make the book usable for forecasters; without them, mids mislead.
Next: how bid, ask, spread, and depth look on prediction-market contracts specifically.