Prediction Markets vs Sportsbooks: Exchange, Bookmaker and Who Holds the Risk

The core difference between a prediction market and a sportsbook is who takes the other side of your bet. At a sportsbook, the bookmaker is your counterparty. It sets a price, carries the risk, and earns the margin between the true probability and the odds it offers. At a prediction market, you trade with other users through an order book. The venue matches orders and earns fees; it does not need you to lose.

That difference explains the costs, the limits, the exit options and the regulatory fight.

Who carries the risk, and what you pay

The bookmaker carries the risk; the exchange matches it

A sportsbook builds a book. It offers odds on every outcome, adjusts them as money comes in, and tries to keep a margin regardless of the result. Pinnacle describes a betting margin as the difference between the true probability of an event and the odds offered, and says its own average margin is about 2% against an industry average of around 6%.Pinnacle

A simple example: if two outcomes are equally likely and a bookmaker prices both at -110, each side implies a 52.38% chance. The two probabilities add up to 104.76%, not 100%. The extra 4.76% is the theoretical margin. A bettor who backs both sides with $110 each spends $220 and collects $210 on the winner, a loss of $10 before any other cost.

A prediction market does not need that margin. Polymarket's order book matches buyers and sellers rather than taking a position.CFTC explainer The venue charges a taker fee in some categories.Polymarket fees The price is whatever the two sides agree on, and the venue's revenue comes from fees rather than from the outcome.

What you pay, explicitly and implicitly

Sportsbook costs are embedded in the odds. You see -110 or 2.50, not a line item called "margin," which makes comparison difficult.

Prediction market costs are more explicit but not always simpler. On Polymarket, the taker fee is calculated as shares × fee rate × price × (1 - price). In politics, the rate is 0.04. Buy 100 shares at 52 cents and the fee is 100 × 0.04 × 0.52 × 0.48 = about $1.00. Your all-in cost is $53 for a position that pays $100 if the event happens, so the break-even probability is 53%, not 52%.Polymarket fees

A sportsbook at -110 requires a 52.38% win rate to break even. On that comparison, the prediction market is not automatically cheaper. It may be cheaper on a different market, at a different price, or for a maker who avoids taker fees, but the only honest comparison is the all-in cost of the exact trade.

Betting exchanges use a third model: Betfair charges commission on net winnings, not the stake, with a rate that depends on the market and the customer's discount rate. The formula is net winnings × market base rate × (100% - discount rate).Betfair commission It rewards consistent winners and penalizes churn less than a per-trade fee, but it is still a cost.

Limits and account restrictions

This is where the structural difference becomes personal. A sportsbook can decline your business. It can set a maximum stake, reduce your stake factor, restrict you from certain markets, or close your account for commercial reasons. The UK Gambling Commission's 2025 review of the largest online real-event betting operators found that 643,779 of 14,923,840 active accounts, or 4.31%, had some form of restriction. Stake-factor reductions were the most common, affecting 62.17% of restricted accounts.UK Gambling Commission

The same review found that 46.78% of restricted customers were in lifetime profit, compared with 25.42% of active customers overall. That does not prove every restriction targets winners, but it shows why successful bettors often find their maximum stake cut. A bookmaker's margin is easier to protect if it can manage the customers who beat it.

An exchange does not have the same incentive. It earns from matching, not from your losses, so its constraint is usually the available liquidity rather than a house decision about your account. That is a different kind of limit: not whether you are allowed to bet, but whether anyone is willing to take the other side at that price.

The practical consequence is that a prediction market can be better for a consistently profitable trader and worse for someone who needs to bet $50 on a game five minutes before kickoff. Liquidity, not permission, becomes the constraint.

Hedging and exiting

An exchange lets you trade out. If you buy Yes at 40 cents and the price rises to 60 cents, you can sell your shares for a profit before the event. You can also take the other side through a No share. That is fundamentally different from waiting for a bookmaker to settle the bet.

A sportsbook may offer cash-out, but the price is set by the bookmaker, can be suspended, and is not a tradable order book. If you want to hedge, you generally need another bookmaker or an exchange. That matters for anyone managing a portfolio rather than placing single bets.

Prediction markets also allow more granular hedging across related outcomes: buying No on one outcome can be similar to buying Yes on the others. Sportsbooks offer related bets, but at the bookmaker's price, not a live order book.

Regulation is not settled

In the United States, sports betting is regulated by states, while event contracts are regulated by the CFTC as derivatives. The line between the two is being litigated. In September 2026, the Sixth Circuit ruled that Ohio and Tennessee could apply their gambling laws to Kalshi's sports-related event contracts, finding that Kalshi had not shown the contracts were swaps within the CFTC's exclusive jurisdiction. The Ninth Circuit reached a similar result for Nevada, while the Third Circuit had earlier upheld CFTC exclusivity. The CFTC has sued multiple states, and the issue is widely expected to reach the Supreme Court.CNBC

That affects access, consumer protection and taxes, not just price. The CFTC says event contracts are federally regulated; several states say sports contracts are gambling. Until the Supreme Court or Congress resolves it, legal status can depend on where you live and what the contract is tied to.

Outside the United States, licensing and tax treatment vary by jurisdiction, and a prediction market operated offshore may sit outside the local gambling framework. That changes both recourse and tax treatment. This is not tax advice, but it is a reason to check the venue's licensing and your local rules before assuming the sportsbook treatment applies.

Where each is better

Sportsbooks are better for simplicity, small stakes, mainstream sports, promotions and immediate settlement. They offer fixed odds and a regulated counterparty; the cost is the margin and the possibility of being limited.

Prediction markets are better for politics, economics and events where the question is not a standard sports bet, for traders who want to exit before resolution, and for anyone who values a transparent order book over a posted price. They can be cheaper at the extremes or for makers, and they do not have a house taking the other side of your bet. The costs are liquidity, fees, resolution risk and legal uncertainty.

Neither is universally superior. The right comparison is the all-in cost of the specific trade, the size you can execute, and the recourse you have if something goes wrong.

The risk section

A sportsbook bet carries counterparty risk: the bookmaker must pay if you win, and it controls the maximum stake and the cash-out price. An exchange carries market risk, liquidity risk, resolution risk and, on crypto-native venues, smart-contract and custody risk. The absence of a house does not mean the absence of a conflict; the venue still writes the rules, operates the matching engine and decides how disputes are handled.

Prediction markets are also not automatically fairer. A thin book can be manipulated, a resolution rule can be ambiguous, and a platform can restrict jurisdiction or freeze withdrawals. The structural advantage is transparency, not a guarantee.

If you are choosing between the two, start with the trade you want to make, then compare the executable price after fees, the depth for your size, the exit mechanism and the legal route. The label on the venue is less important than those four numbers.