What Are Prediction Markets? Types and Examples

Search Fintrev

Search our guides, reviews and glossary.

A prediction market is an exchange where people trade contracts whose payoff depends on the outcome of a future event, such as an election, a central bank decision, an economic data release or a sports result. A typical contract pays a fixed amount, often $1, if the event happens and nothing if it does not. Its trading price therefore works as the crowd’s estimate of the probability: a contract trading at 62 cents implies the market collectively sees about a 62% chance of that outcome.

How Prediction Markets Work

A prediction market starts with a question that will have a verifiable answer by a set date: “Will Candidate A win the election?” or “Will the monthly inflation reading come in above a given level?” The operator lists contracts on each possible outcome and defines how and when the market will be settled.

Traders who think an outcome is more likely than the current price implies buy it. Traders who think it is less likely sell it, or buy the opposite outcome. As news arrives and the settlement date approaches, buying and selling push the price up or down. When the event resolves, contracts on the winning outcome pay out in full and the rest expire worthless.

This resembles a futures market, where traders bid a contract’s price up or down based on what they expect an underlying asset to be worth on a future date. The difference is the underlying: here it is the outcome of an event rather than the price of a commodity or a stock index.

Worked Example: Reading and Trading a Yes/No Contract

Take a hypothetical contract that pays $1 if Team X wins a championship and $0 if it does not.

• Current price: $0.33
• Implied probability: $0.33 / $1.00 = 33%

A trader who believes Team X’s true chance is closer to 45% buys 100 contracts for $33 (before fees).

• If Team X wins, the contracts settle at $1 each, so the trader receives $100. Profit is $100 − $33 = $67, a return of about 203% on the amount staked.
• If Team X loses, the contracts settle at $0 and the trader loses the full $33.

The trader does not have to hold until settlement. If Team X wins several games and the price rises to $0.55, the trader can sell the 100 contracts for $55 and lock in a $22 gain before the final result is known. That ability to exit early is a key feature separating an exchange-traded contract from a ticket that can only be cashed after the event.

For the buyer, expected value can be written as:

Expected profit per contract = (your probability × payout) − price paid − fees

At a 45% belief, the expected profit per contract is (0.45 × $1) − $0.33 = $0.12 before fees. At a 33% belief, it is zero. Trading only makes sense when your estimate differs from the market’s, and your estimate is right more often than the price assumes.

Interpreting Prices When Outcomes Do Not Add Up to 100%

In a market with several outcomes (for example, four candidates), the prices of all outcomes should in theory sum to $1. In practice, the sum can drift above or below $1 because of the bid-ask spread, fees and the timing of trades. A simple adjustment divides each price by the total.

Hypothetical example: Yes trades at $0.60 and No at $0.44, a total of $1.04. The normalized probability for Yes is 0.60 / 1.04 ≈ 57.7%. The extra four cents is a cost to traders, not information about the event.

Contract Types

• Binary (winner-take-all) contracts pay a fixed amount if a specific outcome happens and nothing otherwise. Most mainstream event contracts work this way.
• Vote-share or index contracts pay an amount proportional to a measured value. On the Iowa Electronic Markets, vote-share contracts pay according to a candidate’s share of the vote, so a contract on a candidate who wins 52% of the vote would settle at $0.52 per $1 contract. The price then estimates the expected vote share rather than the probability of winning.
• Multi-outcome markets list a separate contract for each possible result, such as each candidate in a primary or each range for an economic reading.

The contract type changes what the price means. A 52-cent winner-take-all contract signals a 52% chance of winning; a 52-cent vote-share contract signals an expected 52% of the vote, which could correspond to a much higher chance of winning.

Models of Prediction Markets

Operators use several mechanisms to match traders and set prices.

Continuous Double Auction

A continuous double auction matches buyers with sellers through an order book, much as a stock exchange does. Buyers post bids, sellers post asks, and a trade happens when the two meet. The operator keeps a ledger of every trade and pays out to whoever holds the contracts at settlement. This model works well when many traders are active, but in a thinly traded market there may be no one willing to take the other side at a reasonable price.

Automated Market Makers

An automated market maker solves the liquidity problem by having the operator, or an algorithm, stand ready to trade at all times. The operator acts as the counterparty to every trade, much like the house in a casino, and adjusts prices according to how much has been bought on each outcome. Economist Robin Hanson developed one widely cited design, the logarithmic market scoring rule, which moves prices in response to each purchase so that the market always quotes a probability.

Play-Money Markets

Real money gives traders an incentive to be accurate, but it can run into gambling laws in some jurisdictions. Play-money markets use virtual tokens instead, sometimes offering prizes to the participants who accumulate the most. They can operate where real-money betting is restricted and carry little financial risk for participants, though the weaker incentive may affect how carefully people trade.

Blockchain-Based Prediction Markets

Crypto prediction markets use blockchain technology and smart contracts to run markets without a single controlling operator. The smart contract holds funds and pays out automatically, and the outcome is often determined through a voting or reporting system among token holders. Augur was one of the first decentralized prediction markets. It drew criticism after users created markets on the deaths of public figures, prompting warnings that such platforms could become “assassination markets.”

How Prediction Markets Are Used

The main purpose of a prediction market is to collect and combine beliefs about an uncertain outcome. Each trader acts on private information and judgment, and the resulting price aggregates those views into a single number. This makes prediction markets a form of crowdsourcing that relies on the “wisdom of crowds.”

Common uses include:

• Forecasting elections. Political markets are the best-known application. PredictIt is a long-running platform focused on political outcomes.
• Economic and policy questions. Contracts can be tied to inflation readings, interest rate decisions or other official data releases.
• Sports and entertainment. Markets have been listed on game results, championships and even box office receipts.
• Business forecasting. Some companies have run internal markets in which employees trade on quarterly sales, product launch dates or project deadlines.
• Research. Academic markets such as the Iowa Electronic Markets let economists study how markets process information.

Prediction markets also connect to ordinary investing. An investor who expects an election to move stock prices can only express that view indirectly, by buying or selling assets expected to gain or lose from the result. A prediction market lets the same investor trade the election outcome itself, and some participants use event contracts to offset a specific risk in their portfolio.

Prediction Markets vs. Related Concepts

Prediction markets vs. financial markets. Stocks and bonds have no fixed end value; their price reflects expected future cash flows. A binary event contract has a capped payoff and a fixed settlement date, so the most it can ever be worth is its payout amount. Prices in both markets move on information, but an event contract’s price maps directly onto a probability.

Prediction markets vs. sportsbooks. A traditional sportsbook sets the odds and takes the other side of every bet, keeping a margin built into those odds. On an exchange-style prediction market, traders set prices among themselves and the operator typically earns fees. The contract can also be resold before the event. Whether a given product counts as gambling under the law is a separate legal question covered below.

Prediction markets vs. binary options. Binary options use the same all-or-nothing payoff structure: the price rises or falls as the likelihood of an outcome changes. Prediction markets apply that structure to real-world events rather than only to asset prices.

Prediction markets vs. polls. A poll asks people what they intend to do or believe. A prediction market asks them to commit money to what they expect will happen. A poll is a snapshot of current opinion; a market price is meant to incorporate expectations about how that opinion and other factors will change by the settlement date.

Regulation of Prediction Markets in the U.S.

In the United States, real-money event contracts fall under the Commodity Futures Trading Commission (CFTC), which treats them as derivatives. Kalshi operates as a federally regulated exchange offering contracts on elections, economic data, weather and sports. Online sportsbook operators such as DraftKings and FanDuel have also moved into prediction-market products.

The Iowa Electronic Markets operates under no-action letters from regulators, which state that they will not take enforcement action. That relief rests on the market’s academic purpose, its small scale and the fact that it is not run for profit.

The boundary between federal commodities regulation and state gambling law remains contested, especially for sports contracts. States and courts continue to debate whether those contracts are derivatives under CFTC oversight or wagers subject to state gambling rules. Outside the U.S., treatment varies by country, and a platform available in one jurisdiction may be prohibited in another.

Example: The Iowa Electronic Markets

The Iowa Electronic Markets, run by faculty at the University of Iowa’s Tippie College of Business since 1988, is the oldest online prediction market. Participants trade with real money but under small position limits, because the market exists for research and teaching rather than profit. Its election markets list both winner-take-all and vote-share contracts. Over multiple presidential elections, research on the IEM has found that its prices forecast results with greater long-run accuracy than traditional opinion polls, though not in every case, and it remains the standard reference in academic studies of how well prediction markets work.

Limitations of Prediction Markets

A price is not a certainty. A 70% contract will resolve “No” roughly three times out of ten if the market is well calibrated. A losing outcome at a high price does not mean the market was wrong.

Crowds need to be large. The wisdom of crowds depends on many independent participants. Many markets have too few traders, or too little money at stake, for prices to carry much information. In a thin market, a single large order can move the price sharply.

Manipulation and bias. Traders with large balances can push a price to create a false impression of momentum. Participants may also bet on outcomes they want rather than outcomes they expect, which can skew prices on emotionally charged questions.

Costs reduce returns. Trading fees, withdrawal fees and the bid-ask spread all eat into payoffs. A trade with a small edge in the expected-value formula can turn negative once costs are included.

Settlement risk. Every contract depends on its resolution rules. Ambiguous wording, disputed sources or a platform’s decision to void a market can override what traders expected. Contracts touching on a person’s death have been cancelled because regulated platforms cannot settle on that basis, and decentralized markets depend on their reporting system to determine outcomes honestly.

Ethical concerns. Markets on violence, deaths or disasters raise questions beyond legality, as the Augur controversy showed. Some critics also worry that markets on events a participant can influence could create incentives to bring those events about.

Capital is locked up. Money in a contract that settles months away cannot earn interest or be used elsewhere, which can make long-dated contracts trade at prices that differ from the true probability.

FAQs

The Bottom Line

Prediction markets matter to anyone who reads event odds or wants to trade an outcome directly, such as an investor hedging an election risk. A price is an estimate, so a trade makes sense only when your own probability differs from the market's and is more often right, after fees, spreads and settlement rules are counted. Thin trading, manipulation and unsettled U.S. law on sports contracts can all weaken what a price tells you.