Order Books Versus Automated Market Makers
Two ways to make a market in an event contract. One always quotes and may quote badly; the other quotes well when there is interest and not at all otherwise.
There are two mechanisms for making an event market tradeable, and the difference shows up in every price you read from one.
The order book
Participants post bids and offers; the venue matches them. Price comes from what people are willing to do.
Good: tight spreads where there is genuine two-sided interest; price discovery that reflects actual opinion; no pricing at all rather than bad pricing when nobody is interested.
Bad: needs participants. An empty book quotes nothing, and bootstrapping a new market is hard. Maintaining quotes requires active management, which is why most venues offer maker incentives. What market makers do.
The automated market maker
A formula holds a pool of capital and quotes a price derived from it. Trading against the pool moves the price along a curve. Common in early on-chain prediction markets and still used in places.
Good: always quotes. Liquidity can be supplied passively by depositing into the pool. A brand new market is tradeable from the first second.
Bad: the formula does not know anything. It will quote a price that is badly wrong until someone trades it back, and the person who trades it back takes that value from the pool. Liquidity providers therefore lose systematically to informed flow — the event-market version of the problem that affects all automated makers. Spreads are effectively set by the curve rather than by competition, so they are often wider than a functioning book.
Which produces better prices
For markets with real interest, order books, clearly. Competition between makers tightens spreads far below anything a fixed curve provides.
For markets with no interest, an AMM at least gives you a number, and whether that is a feature depends entirely on whether you would rather have a bad price or no price. For reading a forecast, no price is more honest. For being able to trade at all, a bad price is better than nothing.
The hybrid in practice
Some venues run an order book off-chain for speed while settling on-chain, which keeps the price discovery of a book and the custody properties of a contract. How Polymarket's order book works.
Why this matters for reading prices
AMM prices on quiet markets are not forecasts. They are the last state of a curve after whoever last traded. Treat them with the same suspicion as a stale book. Thin market warning signs.
Depth means something different. On a book, depth is resting orders you can count. On an AMM, it is the curve's steepness — how far the price moves per contract. Both answer "what will I pay for size", and the arithmetic is different.
Cross-venue comparison needs care. A book mid and an AMM spot price are not the same kind of number. Comparing them without accounting for the curve overstates what is reachable, on top of all the other reasons a measured gap is not a return. What a gap is worth.
Next
The events category opens with macro release markets.
Educational content only. Nothing here is financial advice, a solicitation, or an offer to trade. Prediction markets can move against you and a position can lose its entire value. Where this guide describes a difference between venues it describes a theoretical cross-market edge in observed prices, which is not a profit: fees, spreads, capital lockup, transfer time and differing settlement rules all come out of it first.
Keep reading
Order Books and Liquidity · 3 min readExit Early or Hold to SettlementTaking a profit before resolution frees capital and pays a spread. Holding pays no spread and ties up capital. The arithmetic that decides it.
Order Books and Liquidity · 3 min readWhat Settlement Does to a Position You Still HoldHolding to settlement means no exit, no spread on the way out, and a payout decided by a rule. The mechanics, and the ways they surprise people.
Order Books and Liquidity · 3 min readThin Market Warning SignsA checklist for spotting a market whose price is not a forecast, before you quote it, compare it or build anything on top of it.
Forecasting and Calibration · 3 min readUpdating on News Without OverreactingNew information should move your forecast by an amount proportional to how surprising it is. Most people move too far on noise and not far enough on signal.