Why a prediction market's odds move even without new trades

11 November 2026

Prediction market odds can move even when no one is actively buying or selling in the order book. This happens because market makers - automated or human - continuously adjust the prices they quote based on incoming information, order flow elsewhere, and changes in their own inventory, not just on the trades that actually execute.

In short, the displayed probability is a live quote, not a record of the last trade. It reflects what a market maker is willing to transact at right now. When that quote changes, the odds move, even if nobody takes the other side.

The market maker's job

On-chain prediction markets use automated market makers (AMMs) or on-chain order books with designated market makers. Both types quote two-sided prices: a price at which they will buy a share (the bid) and a price at which they will sell (the ask). The midpoint between these is what you see as the "current" probability.

A market maker's goal is to profit from the spread while keeping their inventory balanced. If they think the true probability of an event is 60%, they will quote a bid just below 0.60 and an ask just above 0.60. The spread - say 0.58/0.62 - is their compensation for providing liquidity.

But their view of the "true" probability is not static. It changes as new information arrives. A market maker who reads a news wire, sees a prediction from another market, or notices a large buy order on a competing venue will adjust their quotes accordingly. The new quote appears immediately, even if no trade has yet occurred on their own platform.

What moves quotes without trades

Several specific factors can shift a market maker's quoted odds without a single execution on your screen.

New information. Any piece of public news - a poll result, a court ruling, a tweet from a prominent figure - can change the expected outcome in the market maker's model. They reprice within milliseconds.

Order flow on other venues. Prediction markets are not isolated. If the same event is traded on Polymarket, Azuro, and a traditional bookmaker, the market maker will watch the aggregated flow. A sustained bid on another venue signals demand that may spill over, so they raise their own asks preemptively.

Inventory drift. Suppose a market maker has accumulated too many YES shares and too few NO shares. They are net long YES. To rebalance, they may lower their ask on YES (making it cheaper to buy) and raise their bid on NO. This rebalancing happens without any single large trade; it is a standing adjustment to their quotes.

Hedging activity. Market makers often hedge their positions in related markets. If a market maker holds a large position in a Trump-REPUBLICAN market and then sees the odds in a separate "Which party wins the presidency?" market move, they will adjust their quotes in the first market to keep their overall book neutral. The hedge is the driver, not a direct trade in that market.

Funding or borrowing costs. In some prediction market designs, liquidity providers can borrow or lend shares. If the cost of borrowing a particular asset rises, the market maker may widen their spread or shift their quotes to reflect that higher carrying cost.

Time decay. For markets with a fixed end date, the value of a share approaches either 0 or 1 as the deadline nears. A market maker's probability estimate compresses toward the extremes even without any news. Events with a 50/50 chance early on may drift toward 58/42 simply because the remaining time is short and the likelihood of a late surprise falls.

How it appears on-chain

On-chain, these quote changes are visible as state updates. When a market maker updates their order, the smart contract records the new price. If you are looking at a front-end dashboard, you see the odds change. The transaction that moved the price was not a trade - it was a quote update.

In an AMM, the mechanism is slightly different. The AMM's price is a function of the reserves in its pool. If no one trades, the price does not change automatically. But a rebalancing mechanism - such as a fee accrual or a rebalancing bot - can trigger a swap that adjusts the pool's ratio. That swap is technically a trade, but it is initiated by the market maker to correct their own position, not by a directional bettor.

Reading the Tape vs. Reading the Book

If you watch only the trade feed, you will miss most of the action. A single large trade might print at 0.60, but the current ask could be 0.62 because the market maker has already repriced upward in anticipation of follow-through buying. Conversely, a market can look stable on the trade tape while the order book is silently thinning or widening.

To understand why odds moved, you have to look at the order book depth, the time of the last quote update, and the market maker's inventory. None of that is visible from the trade history alone.

The Practical Takeaway

Do not assume a price move implies a recent trade. It often reflects a market maker's decision to change their displayed price. That decision is based on private models, external signals, and their own risk tolerance. For a trader, this means:

1. Check the order book spread, not just the last price.
2. Look at the timestamp of the current quote. If it changed but no trade followed, the market maker repriced.
3. Watch for quote updates on both sides of the book. A one-sided move is more significant than a symmetric shift.
4. Be aware that a wide spread can indicate a market maker who is unwilling to take on risk, not necessarily a market with no liquidity.

Odds are a live opinion. They move because the opinion changes, not because money necessarily changed hands. That distinction matters for anyone trying to infer "who is buying what" from a price chart. Sometimes the answer is simply: no one bought anything. The market maker just changed their mind.

Back to prediction markets