- Prediction markets are exchanges where you buy and sell contracts that pay $1 if an event happens and nothing if it doesn't. The price is the market's probability: 30¢ means about a 30% chance.
- Fees and the spread come off every trade. On Kalshi, 100 contracts at 50¢ cost $1.75 in taker fees, and 100 at 90¢ cost $0.63.
- Prices are good forecasts on average, not guarantees. Research on Polymarket finds only 3.14% of accounts statistically skilled, and the top 1% of profitable users earned 76.5% of profits.
- The CFTC says regulated prediction markets can operate in all 50 states, but individual states still contest some sports contracts.
Prediction markets are exchanges for trading on the outcome of future events: an election, a Federal Reserve decision, a championship, tomorrow's high temperature. Each question becomes a contract, and the price of that contract tells you what the people trading it think the odds are. This guide defines the term, explains how contracts pay out, how to read a price, what trading costs and where the risks are. It is an explainer, not financial advice.
What are prediction markets and how do they work?
Every market asks a yes-or-no question with a fixed resolution rule, such as "Will the Fed cut rates at its next meeting?" You can buy a Yes contract or a No contract. When the event resolves, the winning side pays $1 per contract and the losing side pays nothing.
Because one side always pays $1, the Yes and No prices add up to about $1. If Yes trades at 30¢, No trades at about 70¢. Questions with many possible answers, like "Who will win the World Cup?", are a set of yes/no contracts, one per answer, grouped into a single event.
There is no house setting the odds. Like a stock exchange, the platform matches buyers with sellers, and the price moves as traders react to news. You can sell before the event resolves to lock in a gain or cut a loss. You don't have to hold to the end.
The CFTC, the US regulator, describes the same thing in its investor guidance: event contracts are often yes-no contracts with only two possible outcomes, and traders use them either to hedge real-world risk or to speculate (see sources).
An example of a prediction market
Take a market asking whether a central bank will hold interest rates at its next meeting. Suppose Yes trades at 80¢. A trader who buys 100 Yes contracts pays $80. If rates are held, the contracts pay $100 and the profit is $20 before fees. If rates change, the $80 is lost. The same market has No contracts at about 20¢, so a trader who thinks a change is likelier than 20% might buy No instead.
How prediction markets pay out
Payout is mechanical. When the market resolves under its written rule, each winning contract is credited $1 and each losing contract is worth zero. Profit per contract is $1 minus what you paid, less fees. If you sold earlier, you keep the difference between your sale and purchase price instead. Fees differ by venue: Kalshi charges takers 7% × contracts × price × (1 − price), rounded up to the next cent, while Polymarket's taker fee uses the same shape with a rate that depends on category (for example 4% for politics and 5% for sports, as of October 13, 2026) and charges makers nothing. Fee documentation is in the sources.
What prediction market prices mean
A contract's price in cents reads as an implied probability. If Yes costs 30¢, the market as a whole is pricing the event at roughly a 30% chance. You make money by disagreeing with the market and being right: if you think the true chance is higher than 30%, Yes is cheap; if you think it is lower, No is cheap.
The table below shows what one contract and a 100-contract position mean at a few prices. Fees use the taker formula from Kalshi's fee schedule, effective July 7, 2026.
| Yes price | Implied chance | Profit per contract if Yes | Loss per contract if No | Kalshi taker fee, 100 contracts |
|---|---|---|---|---|
| 10¢ | 10% | 90¢ | 10¢ | $0.63 |
| 30¢ | 30% | 70¢ | 30¢ | $1.47 |
| 50¢ | 50% | 50¢ | 50¢ | $1.75 |
| 90¢ | 90% | 10¢ | 90¢ | $0.63 |
Two things stand out. The fee is largest at 50¢, where the outcome is most uncertain, and smallest near the extremes. And a cheap contract is not a bargain by default: a 10¢ contract loses 10¢ nine times out of ten if the price is right. You can run any price and size through the fee and break-even calculator, and how much edge the fees take at each price works through the break-even maths.
For a fuller walkthrough of one price, see what a 72-cent contract means.
Are prediction markets accurate?
On average, reasonably, according to the research. The standard survey of the field, by Justin Wolfers and Eric Zitzewitz in the Journal of Economic Perspectives, found that market forecasts are typically fairly accurate and outperform most moderately sophisticated benchmarks (see sources). Three things drive that:
- Money at stake. A trader who is wrong pays for it, which discourages wishful thinking.
- Constant updating. Prices move within minutes of news, where a poll is a snapshot.
- Many sources of knowledge. The price reflects everyone trading, not one forecaster.
"Accurate on average" still allows any single market to be wrong. Thinly traded markets can be moved by a few large orders, so a cheap contract in a thin market deserves extra scepticism.
Do people make money on prediction markets?
Some do, but gains are concentrated. A CEPR study of 588M trades and $67B of volume on Polymarket found that the top 1% of profitable users took 76.5% of profits. A separate study on skill versus luck found only 3.14% of accounts statistically skilled. Trading is a zero-sum exchange before fees, so every profit is someone else's loss, and fees tilt it further. Our analysis of who makes money on Polymarket breaks down the maker-versus-directional split among 200 profitable wallets. Aggregated data like this is described on our methodology page.
Is a prediction market just gambling?
The experience can feel similar, and some governments and regulators treat betting-style products as gambling. The structure differs: you trade against other people on an exchange rather than against a bookmaker, and prices come from supply and demand. The CFTC notes that regulated platforms do not take a side against their customers and must follow rules against manipulation and insider trading. The risk is still real. The CFTC notes that traders who predict wrongly lose their investment, so only trade money you can afford to lose.
Are prediction markets legal in the US?
The CFTC says prediction markets have been regulated by it since 2004, and that such markets can operate in all 50 states under federal law (see sources). That has not ended the argument: some states have challenged sports event contracts in court, and the rules are still moving, so what you can trade can differ by state and product. The "illegal" label you sometimes see usually refers to unregistered offshore venues or to these state disputes. Unregistered entities operating outside the US can leave you with little or no protection, according to the CFTC. Our PredictIt review shows one venue that runs under a no-action letter instead of exchange registration.
Where can you trade prediction markets?
The main venues differ in regulation, who can sign up, which events they list and what they charge. Kalshi is regulated in the United States by the Commodity Futures Trading Commission as an exchange, and lists contracts on economics, politics, weather, culture and sports. Polymarket is the best-known crypto-based venue. Other brokers and exchanges have added event contracts too.
Which one fits depends on where you live and what you want to trade. Our platform comparison lays out fees, availability and market coverage side by side, Kalshi vs Polymarket goes deeper on the two largest, and the live markets page shows what is trading right now.
Where this can go wrong
- You can lose your whole stake. A contract that resolves against you pays nothing.
- Fees and spreads add up. Buying at the ask and selling at the bid costs the spread on top of fees. That matters most for frequent traders.
- Read the resolution rules. Markets settle on their exact wording and named data source, which may not match the headline.
- Liquidity varies. In a quiet market you may not be able to sell at a fair price before the event.
- Volume can be inflated. A Columbia study estimated that about 25% of Polymarket volume from 2022 to 2025 was wash trading, so headline volume overstates real interest.
- The rules are still moving. State challenges and court cases are ongoing. Check what is available where you live.
Frequently asked questions
What is an example of a prediction market?
A market on whether the Fed will leave rates unchanged at its next meeting. A Yes contract at 80¢ pays $1 if rates are held and nothing otherwise, so the market is pricing the hold at about 80%. Election, weather, sports and economic-data markets work the same way.
Is a prediction market just gambling?
It can feel similar, but the structure differs. You trade against other people on an exchange rather than against a bookmaker, prices are set by supply and demand, and regulated venues are overseen as financial exchanges. The risk is real either way: only trade money you can afford to lose.
Are prediction markets usually correct?
Research finds their forecasts are typically fairly accurate on average. Any single market can still be wrong, especially a thinly traded one.
How do prediction markets pay out?
Each winning contract pays $1 when the market resolves under its rules, and each losing contract pays nothing. Your profit is $1 minus the price you paid, less fees. You can also sell before resolution and keep the price difference.
Can you actually make money on prediction markets?
Some traders do, but profits are concentrated. Academic work on Polymarket finds that the top 1% of profitable users earn most of the profits and that only a small share of accounts show statistically significant skill. See who actually makes money on Polymarket.
Why are prediction markets illegal in some places?
Many governments treat betting-style products as gambling and restrict them. In the US, the CFTC regulates registered exchanges and says they can operate in all 50 states, but some states contest sports contracts, so availability varies by state and product. Our comparison page covers availability by platform.
Getting started
Start small, trade questions you understand, read each market's resolution rules, and work out fees before you buy. If you want to see what's trading, the politics markets and sports markets pages list current prices.