This article explains how prediction markets work. It is not investment advice and does not recommend trading or betting on any platform.
A prediction market is a place to trade contracts whose payout depends on whether a specific future event happens. Unlike a stock, whose value reflects a company's ongoing earnings, a prediction-market contract has a fixed, binary result: it pays $1 if the event occurs and $0 if it does not.
That simple structure has an elegant consequence. If a "yes" contract is trading at 67 cents and pays $1 when the event happens, buyers should bid it up toward what they truly believe the odds are. In practice, the price tends to settle near the crowd's collective estimate of the probability — so a contract at 67 cents reads as roughly a 67% implied chance. This is the core appeal of prediction markets as forecasting tools: the price is a live, numerical probability.
How a contract works
Every question on a platform like Kalshi or Polymarket is framed as a yes/no proposition — "Will the Federal Reserve cut rates before September?" Traders can take either side, and the "yes" and "no" prices always add up to about $1, minus fees. When the event resolves, winning shares pay $1 each and losing shares pay nothing. A trader's profit or loss is just the gap between the price they paid and that final settlement.
Polymarket and Kalshi: two different animals
The two best-known platforms operate under very different rules.
Kalshi is a U.S.-regulated exchange licensed by the Commodity Futures Trading Commission (CFTC), the federal agency that oversees derivatives. Its event contracts are treated as regulated financial instruments, which has been central to the legal fights over whether such markets are lawful nationwide rather than subject to a patchwork of state gambling laws.
Polymarket is built on blockchain technology and is not available to U.S. residents, operating in a more lightly regulated international space. As the consulting firm KPMG has noted, Polymarket has moved to acquire a CFTC-registered entity in a bid to build a formal regulatory foothold in the United States.
How they differ from sports betting and ordinary markets
Prediction markets sit in an awkward middle ground. Functionally they resemble a sportsbook: put money on an outcome, collect if you are right. But legally they can be different. Sports-betting apps operate under state gaming licenses, whereas CFTC-regulated event contracts are federal financial instruments — a distinction that has been heavily litigated as platforms expand into sports and entertainment questions.
They also differ from conventional derivatives. A futures contract on crude oil can hedge a real commercial exposure and settles on a continuously moving price. A prediction-market contract settles on a binary yes/no event. U.S. rules generally require event contracts to be tied to outcomes with a genuine economic consequence — a standard regulators are still working to define.
Do they actually forecast well?
The evidence is genuinely mixed. The strongest historical case comes from the Iowa Electronic Markets, a long-running academic platform, which outperformed opinion polls in studies of U.S. presidential elections, beating poll-based forecasts a clear majority of the time across several races.
But newer research complicates the picture. Some studies find that simple aggregated surveys of informed people can match or beat market prices, and that a market's accuracy depends heavily on how many people are trading it. When liquidity is thin — few participants, small amounts of money — prices are easier to push around, and ordinary behavioral biases like overconfidence and herd behavior can distort them. A prediction market is a crowd estimate, not an oracle.
The regulatory and integrity questions
Oversight is still being written. The Congressional Research Service, the U.S. Congress's nonpartisan research arm, has laid out the open policy questions: which event contracts should be allowed, how to treat contracts on elections and sports, and how to police manipulation. The CFTC has shifted its stance more than once, and fresh rulemaking has been widely anticipated.
Manipulation is a live concern, especially on decentralized platforms where there is no licensed intermediary to monitor for wash trading or coordinated price-moving. Recent allegations that some Polymarket promoters staged fake bets in marketing videos — now the subject of a request for a CFTC investigation — underline how integrity questions extend beyond the trading itself to how these platforms are sold.
Jurisdiction adds a final layer. Prediction markets are legal for U.S. users on CFTC-licensed venues like Kalshi but remain restricted or prohibited in many other countries, and Polymarket bars U.S. residents outright.
The bottom line
Prediction markets offer a genuinely useful mechanism: they distill scattered opinions into a single, real-time probability. Their record is better than chance, and sometimes better than polls — but they are experimental financial infrastructure, vulnerable to thin trading and manipulation, and operating under rules that regulators are still drafting. They are best read as one more signal to weigh, not a crystal ball.
