What Is a Prediction Market?
A prediction market is an exchange where you buy and sell contracts on whether something will happen. Each contract settles at $1 if it happens and $0 if it doesn’t. The price you pay in between is the market’s estimate of how likely it is.
Buy a contract at 30¢ and you risk 30¢ to make 70¢. That price is also a probability: roughly a 30% chance.
The one-minute version
- You trade against other people, not against a house. Someone else holds the other side of your contract.
- Price = probability. 30¢ means about 30%. Prices run from 1¢ to 99¢.
- Both sides always exist. A contract can’t have a buyer without a seller, so there is no such thing as a side the venue declined to offer.
- You can sell before it settles. Unlike a bet slip, a contract can be traded out of at the current price.
- The cost is a spread plus a fee, not vig folded into the odds.
- They are regulated as exchanges, not as sportsbooks — in the US, by the CFTC.
Reading Yes/No as Over/Under
Exchanges label the two sides Yes and No. Sportsbooks label them Over and Under. They mean the same thing:
| Exchange contract | Sportsbook equivalent |
|---|---|
| Yes on “1+ home runs” | Over 0.5 home runs |
| No on “1+ home runs” | Under 0.5 home runs |
| Yes on “20+ points” | Over 19.5 points |
| No on “20+ points” | Under 19.5 points |
Exchanges quote whole-number milestones (“1+”, “20+”); books use half-points to avoid ties. The two describe the same outcome.
The prices are linked. If Yes is trading at 21¢, No costs about 79¢ — the two must add to roughly $1, because exactly one of them pays.
Where the cost is
A sportsbook charges you through vig: margin folded into the odds. You never see it as a line item. An exchange charges in two visible pieces instead.
The bid-ask spread. The best price to buy is always a little above the best price to sell. That gap is the first cost, and it’s the exchange equivalent of a book’s margin.
A trading fee, published and charged separately. Both US venues use the same shape, as of August 2026:
| Venue | Taker fee per contract | Fee at 50¢ | Fee at 90¢ |
|---|---|---|---|
| Kalshi | 0.07 × price × (1 − price) | ~1.75¢ | ~0.63¢ |
| Polymarket US | 0.06 × price × (1 − price) | ~1.50¢ | ~0.54¢ |
The shape matters more than the rate. price × (1 − price) is
largest at 50¢ and shrinks toward both ends, so a coin-flip market is where the fee
bites hardest and a lopsided one is where it nearly disappears. Both venues also charge
less — or pay a rebate — if you post an order and wait rather than taking an
existing price.
Who is on the other side
“Peer to peer” suggests another fan took the opposite view. Usually it’s a firm. Both venues run market maker programs: professional traders commit to quoting both sides continuously in exchange for reduced fees and higher position limits. Kalshi operates a central limit order book as a CFTC-regulated exchange with designated market makers under defined quoting obligations.
That’s why quotes exist at 3am on a Tuesday, and it’s also why you should assume the other side of your trade is someone whose full-time job is pricing these markets.
The two main US venues
| Kalshi | Polymarket US | |
|---|---|---|
| Structure | CFTC-regulated exchange, central limit order book | CFTC-regulated US exchange |
| Taker fee | 0.07 × p × (1−p) | 0.06 × p × (1−p) |
| Sports coverage | Game lines and player props across several leagues | Game lines and some player props |
| Combo props (PRA and similar) | No | No |
There are two different Polymarkets. The CFTC-regulated US app and the global site are separate exchanges with separate order books, separate market ids and different liquidity. A price on one tells you very little about the other. If you read a spread or a volume figure about “Polymarket”, check which one it describes.
How they differ from a sportsbook
- Both sides always exist. A book chooses which sides to post and often lists only the Over on props like home runs. An exchange can’t — the Under is there by construction.
- Resolution rules are their own. These are contract terms, not sportsbook grading rules. A player prop may settle to a fair-market price rather than voiding if the player never appears. Read the rules text before assuming it behaves like the equivalent bet.
- Depth is thinner. A tight spread tells you the price of the next small trade, not that size is waiting there. Books will usually take a larger bet at their posted number.
- You can exit early. A contract can be sold back into the market before the event settles.
- No combo markets. Bundles like Points + Rebounds + Assists are a sportsbook product. Exchanges list the single stats.
- Availability varies by state and keeps changing. Access to sports event contracts has shifted repeatedly through 2026. Check the exchange rather than assuming.
Common questions
Is a prediction market the same as sports betting?
Economically it’s similar — you risk money on an uncertain outcome. Legally and structurally it isn’t: these are exchanges regulated as derivatives venues rather than sportsbooks, you trade against other participants instead of a house, and you can sell a position before it settles.
Are the odds better?
Often, and the advantage is concentrated in a predictable place: lopsided markets. That falls straight out of the fee curve above.
Because the fee scales with price × (1 − price), it is at its
largest on a near coin flip and at its smallest out at the edges of the board. A
sportsbook’s margin does not work that way — if anything, books tend to hold
more on longshots and on the heavy-favorite side of a prop, not less. The two cost
structures diverge most exactly where the price is furthest from 50¢.
So the exchange edge is biggest on:
- Unders on low-numbered props. “Under 0.5 home runs” is a heavy favorite — it usually prices somewhere in the 80s or 90s. Many books don’t list that side at all, and where they do, the juice on a big favorite eats a large share of a small return.
- Longshots. The other end of the same board, for the same reason.
- Anything a book posts one-sided. If there is no Under to compare against, the book’s margin on that market is whatever it wants it to be.
Near 50¢ — a coin-flip moneyline, a well-balanced total — the gap narrows, and a sharp sportsbook price can be competitive.
Two caveats before treating this as a rule. Liquidity varies enormously: the same exchange can be tight on a marquee market and wide on a thin one, and a smaller league is generally worse than a major one. And a quoted spread tells you the price of the next small trade, not that size is available there. The measured comparison, with the numbers and the date they were taken, is in prediction markets and the missing Under.
What does the price actually mean?
It’s the market’s estimate of probability, net of fees. A contract at 65¢ implies roughly a 65% chance. That makes prices directly comparable to a hit rate, which is a different kind of number: one is what the market expects now, the other is how often something has happened before.
Can I lose more than I put in?
No. A contract’s worst case is settling at $0, so the most you can lose is what you paid for it.
Bettorly is an analytics tool — we don’t take bets and don’t sell picks. Historical frequencies never guarantee a future result. 21+. Gambling problem? Call or text 1-800-GAMBLER.
Try the parlay analyzer →