What are prediction markets?

The questions prediction markets ask are as concrete as they sound: will inflation come in above 3.0% in September, will this hurricane make landfall in Florida, will a named company ship a product before year end. Each one is a written question with a fixed dollar attached to the answer, which makes the price the market's response to something the rest of finance usually leaves implicit: how likely is this, really? That single property — a live, money-backed probability on a plain-language question — is what makes prediction market trading a different instrument from stocks, options, or sports odds, and it is why economists, journalists, and risk desks all read these prices.

The contract, defined

The tradable object is called an event contract: a written question, a written rule for how it resolves, a named source that decides, and a deadline. You take the YES side or the NO side. Nothing about the contract is open to interpretation after the fact — the rule is published before anyone trades it, which is exactly why reading the rule is the first real skill on these venues. Two traders can agree completely about the world and still disagree about a contract, because the contract is about the rule, not the vibe of the headline.

YES and NO are two sides of the same dollar. If YES trades at 62¢, NO trades at 38¢ — the two prices are complementary by construction, because exactly one of them will be worth $1 at the end. Buying NO at 38¢ is how you express "I think this won't happen"; there is no separate short-selling mechanism to learn.

Why the price is a probability

A contract that pays $1 and costs 62¢ pays you 38¢ if you are right and costs you 62¢ if you are wrong. Break-even is exactly the point where those two outcomes cancel — which happens when the true probability is 62%. Above that, buying is a positive-expectation trade; below it, it isn't. So the price is the market's probability, expressed in cents. Traders read 62¢ as "62%" without thinking about it.

The one honest correction is fees. Exchanges charge a trading fee that is largest for coin-flip contracts near 50¢ and small at the extremes. Two things about how Kalshi charges it matter more to a reader than the rate itself: the fee lands once per order, not once per contract, and it rounds up to a whole cent. On a large order that rounding disappears into the total; on a single contract it dominates. Small orders therefore pay a proportionally higher rate, which is a real argument against testing an edge one contract at a time. The rate itself is Kalshi's to publish, and the current schedule lives on their site.

One trade, all the way through

Suppose a contract on a data release trades at 62¢, and your own work — the release calendar, the components already published, how the number has behaved after similar inputs — puts the real chance at 70%. Every figure below is arithmetic on those two numbers.

Where the prices come from

On a regulated exchange you are trading against other traders through a central limit order book — a public list of the prices people are willing to buy and sell at. The exchange matches buyers with sellers and takes no side of its own. That is the structural difference from a sportsbook, which is your counterparty and prices to make a margin whichever way the event goes. It also means the usual market realities apply: the price you see is the best resting offer, not a guaranteed fill at your size, and thin markets can move several cents when a real order arrives.

Are these prices any good?

Broadly, prices on liquid markets aggregate information well, for a boring reason: a trader who knows something can act on it, and doing so moves the price. Nobody has to be persuaded, and no editor has to approve. But "well" is not "perfectly", and two documented biases are worth carrying with you:

• Long-shot pricing. Very unlikely outcomes often trade a little richer than they deserve, and near-certainties a little cheaper, because a 3¢ lottery ticket is more fun to hold than a 97¢ contract that ties up capital for weeks.
• Thin markets drift. Away from the headline questions, a market may have traded twice today. Its price is one person's opinion with a wide spread around it, not a crowd's verdict.
• Resolution risk is real risk. Most arguments about these markets are arguments about the settlement rule, not the world — a revised statistic, an ambiguous headline, a source that publishes late. Read the rule before you read the chart.

Where the edge actually comes from

Not from having opinions faster. On event contracts, the durable edges are the unglamorous ones: reading the settlement rule more carefully than the crowd; knowing the base rate — how often this kind of thing happens in a normal year — before you look at the price; relative value, where you compare related contracts to each other rather than to your own forecast; and sizing, which decides whether a real edge survives a run of bad luck. Our guide to Kelly-criterion sizing covers that last one, and the strategy guide covers the first three with examples.

It is also worth knowing what these markets are not good for. They are a poor vehicle for expressing a view over years, because the capital is locked until settlement and the fee is paid whether you are early or right. They reward people who write down what would change their mind, and punish people who buy a headline at 88¢ and discover that the remaining 12¢ was the whole story.

Where you can trade them in the US

Event contracts are regulated instruments in the United States, listed on exchanges designated by the Commodity Futures Trading Commission — Kalshi being the most established. That designation is what separates these venues from offshore sites, and it comes with rulebooks, customer fund segregation, and a regulator with jurisdiction. Our sober look at whether Kalshi is legit covers what that protects and what it doesn't, and the Kalshi vs Polymarket comparison covers the venue choice.

The mechanics are the easy part. The hard part is doing the work every time: reading the rule, writing down a probability you can defend, checking the fee-adjusted break-even, sizing to your edge rather than your enthusiasm, and knowing in advance what would make you close. That loop is what Quantreno runs on prediction markets — an AI desk that researches the question, states its reasoning, sizes the position server-side, and hands you a recommendation to confirm or pass on. It is a preset for how to work, not personal advice, and no order is ever placed without you.