What is a prediction market?
A prediction market turns a question into something you can buy and sell. Will the Fed cut rates in December? Will a team win tonight?
Each question becomes a contract that pays $1 if the answer is yes. Traders buy and sell it until the result is known.
The idea is old. The University of Iowa has run election markets since 1988, through the Iowa Electronic Markets.
Today the largest in the US are Polymarket US and Kalshi, both regulated by the CFTC. Brokerages and sportsbooks now offer them too.
How a prediction market contract works
Every contract has two sides, Yes and No. Their prices add up to about $1.
Buy Yes at 40¢, and if the event happens you get $1: a 60¢ gain before fees. If it does not, you lose your 40¢.
You do not have to wait for the result. You can sell to another trader at the current price whenever there is a buyer.
Here is a live market, priced as of this page loading.
- The question. Will David Lisnard win the 2027 French presidential election? trades at 11¢ for Yes, so traders give it about a 11% chance.
- Buy 100 Yes contracts for $11.00, before fees.
- If it happens, they pay $100.00: a $89.00 gain. If not, you lose $11.00.
- Or sell early to another trader, at whatever the market pays then.
Why the price is a probability
The payout is fixed at $1, so the price is the only thing that moves. Nobody sensible pays 80¢ for a coin flip.
That makes the price a forecast. A contract at 62¢ says traders, on balance, put the chance at about 62%.
Converting takes no math: cents are percentage points. Our implied probability guide compares that with sportsbook odds, and the betting odds calculator converts any price for you.
Prices move when news does. A jobs report, an injury or a poll can reprice a market within seconds.
What you can trade on prediction markets
Politics: elections, primaries and confirmation votes, from the White House to a mayor's race.
Economics: Fed rate decisions, inflation, jobs reports and GDP.
Sports: game winners, spreads, totals and player props across the major leagues.
Crypto, weather and culture: Bitcoin price ranges, daily high temperatures, award shows and box office.
Our live odds page shows what is trading now, across venues.
Regulated exchanges vs crypto prediction markets
Prediction markets come in three kinds, and the difference decides your protections.
Regulated exchanges, like Polymarket US and Kalshi, take dollars, check your identity and answer to the CFTC.
Crypto markets, like Polymarket's international site, run on stablecoins and self-custody wallets. They settle through an oracle and sit outside US regulation.
Play-money markets use points instead of cash. Nothing is at stake, which makes them good practice and poor forecasts of anything costly.
Prediction markets vs sports betting
They look alike, and some courts treat sports contracts as bets. The mechanics still differ in several ways.
Who sets the price: a sportsbook sets its odds, while on an exchange traders do.
Who takes the other side: the sportsbook itself, or another trader.
How the operator earns: a sportsbook builds a margin into its odds, while an exchange charges a published fee.
Who regulates it: a state gaming regulator for a sportsbook, and the CFTC for an exchange.
Who regulates prediction markets?
In the US, the CFTC. It designates exchanges as contract markets and supervises how contracts list, trade and settle.
That federal status lets apps offer contracts nationwide. It is also contested: several states say sports contracts are unlicensed betting.
Three federal appeals courts have ruled on that question, and they split. Our guide to whether prediction markets are legal tracks each case.
Outside the US, rules vary by country. Some treat these contracts as gambling, and some block the apps outright.
How accurate are prediction markets?
Often more accurate than polls, though not infallible. The best long-run test comes from the Iowa Electronic Markets.
Researchers compared its forecasts with 964 polls across five presidential elections, from 1988 to 2004.
The market was closer to the final result 74% of the time. It beat the polls in every election when forecasting more than 100 days out.
Accuracy here means calibration. Of all contracts trading at 70¢, about 70% should resolve Yes, and 30% should not.
So a 70¢ favorite that loses is the market working, not failing. Thin markets, and markets open to insiders, are less reliable.
A short history of prediction markets
For decades, US prediction markets were small experiments run by universities. Real-money markets at scale are a 2020s story, and a contested one.
The risks of prediction markets
You can lose your whole stake. A contract that resolves No pays nothing.
Thin markets move easily. One large order can swing the price, and you may not find a buyer when you want out.
Rules decide payouts. A market settles on its named source and wording, which may not match the headline.
Insider trading is a crime here too. In 2026, prosecutors charged a soldier and a Google engineer over trades made on secret information.
And the law is unsettled. A state order can close a category of contracts where you live.
Where you can trade prediction markets
In the US, you have more than a dozen regulated options. They range from exchanges like Polymarket US and Kalshi to apps inside Robinhood, Coinbase and DraftKings.
They differ on fees, markets, funding and which states they serve. Our app rankings score each one on the same six criteria.
These are the top five. For the mechanics of one app, see how Kalshi works.
| # | App | Score | Best for |
|---|---|---|---|
| 1 | Kalshi | 4.9 / 5 | all six market categories, traded on the exchange itself |
| 2 | Gemini Predictions | 4.7 / 5 | existing Gemini crypto customers |
| 3 | Polymarket US | 4.7 / 5 | trading Polymarket from the US, on a CFTC-regulated exchange |
| 4 | OG | 4.6 / 5 | trading directly on the exchange behind Crypto.com’s markets |
| 5 | Interactive Brokers | 4.4 / 5 | comparing ForecastEx, Kalshi and CME prices in one account |