Guide

How Prediction Markets Work

A prediction market is an exchange where people buy and sell contracts tied to a future event, such as who wins an election or whether a rate cut happens by a certain date. Each contract is designed to pay $1 if the event occurs and $0 if it does not, so its price, quoted from 0 to 100 cents, can be read as the crowd's estimated probability. A contract trading at 63 cents implies roughly a 63 percent chance. These are probabilities produced by trading, not predictions of certainty or financial advice.

The core mechanism: a $1 binary contract

Most prediction markets run on binary, or yes/no, contracts. The exchange defines a clear question with a fixed resolution date and a rule for deciding the outcome, for example: "Will Candidate A win the election?" Each "Yes" share pays out $1 if the answer turns out to be yes, and nothing if it turns out to be no. "No" shares pay the reverse.

Because the maximum payout is $1, the price can only sit between $0 and $1, usually quoted in cents. Buyers and sellers meet in an order book or are matched by an automated market maker, the same way a stock exchange pairs trades. When more people want to buy "Yes" than sell it, the price rises; when sentiment cools, it falls. When the event finally happens or fails to happen, the market resolves: winning shares are redeemed for $1 each and losing shares expire worthless.

Why the price reads as a probability

The link between price and probability comes from simple expected value. If you believe an event has a 70 percent chance of happening, a "Yes" share is worth 70 cents to you on average: 70 percent of the time it pays $1, and 30 percent of the time it pays nothing, which works out to about $0.70.

If the share were trading at only 50 cents, you would expect to profit by buying it, and so would others, pushing the price up. If it traded at 90 cents, sellers would step in and push it down. The price settles where buyers and sellers roughly balance, which is the level the marketplace collectively treats as the implied odds. This is why analysts often say the price "is" the implied probability. It is the number at which informed money stops finding an obvious edge. In practice, fees, the bid-ask spread, and the small cost of tying up money until resolution mean the reading is an approximation rather than an exact figure.

A worked example

Suppose a market asks whether a central bank will cut interest rates at its next meeting. Early on, with little information, "Yes" trades at 40 cents, an implied 40 percent chance.

Then a weak jobs report is published, making a cut look more likely. Traders who think the real probability is now higher buy "Yes," and the price climbs to 58 cents. A trader who bought at 40 cents can now sell at 58 cents for a gain, without waiting for the meeting, just as a stockholder can sell before a company reports earnings. If the meeting arrives and the bank does cut, the market resolves: every "Yes" share pays $1, so anyone holding from 40 cents earns 60 cents per share, while "No" shares pay nothing. The moving price chart is, in effect, a real-time record of how the crowd's confidence shifted as news arrived.

Who trades, and why the signal can be useful

Participants range from hobbyists and political junkies to professional traders and people hedging real exposure, such as a business worried about a policy change. The mix matters: traders are risking their own money, which rewards being right and punishes wishful thinking, unlike a free opinion poll.

This incentive is why prediction-market prices can aggregate scattered information well. In one long-running academic example, the Iowa Electronic Markets, run by the University of Iowa's Tippie College of Business, were found to be closer to the final result than polls about 74 percent of the time in a study of 964 polls spanning the 1988 to 2004 U.S. presidential elections. Some researchers have since questioned how directly market prices and same-day polls should be compared, so the edge over polls is real but debated rather than absolute. Venues such as Kalshi, a U.S. exchange regulated by the Commodity Futures Trading Commission, and Polymarket, which originated as a crypto-based platform, now run large numbers of markets on politics, economics, and other topics.

The real limitations

Prediction markets are not crystal balls, and several well-documented flaws can distort the price.

Low liquidity. In a market with few traders and little money, a single sizable order can swing the price far more than new information warrants. Thinly traded contracts are noisy and should be read with caution.

Manipulation. Because the price is itself a public signal, a participant may push it to influence perception or to mislead less informed traders. Large "whale" positions have visibly moved prices in real episodes.

Long-shot bias. Research repeatedly finds that very unlikely outcomes tend to be overpriced and heavy favorites underpriced, a pattern called the favorite-longshot bias. A study of Kalshi contracts found that those bought for under 10 cents lost, on average, well over half their value, meaning rare events priced at a few cents tend to win even less often than the price suggests.

The practical takeaway: treat a market price as a useful, money-weighted estimate of probability, best on liquid, heavily traded questions, and never as a guarantee or a recommendation to bet.