Prediction Markets vs. Sportsbooks: Same Games, Different Machines

A sportsbook banks your bet: it writes the line, takes the other side, and keeps a margin inside the odds. A prediction market matches you with another trader and charges a posted fee for the privilege. Same games, same outcomes, structurally different machines, and every practical difference on this page (cost, exit options, state availability, even your 2026 tax forms) falls out of that one distinction.

The two worlds are also colliding in court and in business. This week alone, a federal judge blocked Minnesota’s exchange ban and New York’s attorney general sued Kalshi in state court; meanwhile DraftKings, the archetypal sportsbook operator, now runs its own CFTC-regulated exchange. Here is the whole comparison, with our own measured numbers where the marketing usually goes.

Who You Are Actually Betting Against

At a sportsbook, your counterparty is the house. The operator prices every line, accepts your bet against its own book, and manages its risk across thousands of customers. That model is why the menu is so deep (the book can manufacture a price for almost anything, from a same-game parlay to a first-basket prop) and why the price includes a margin: the book is paid for making the market and carrying the risk.

On an exchange, your counterparty is another trader. The venue runs an order book, matches buyers with sellers, and posts a fee schedule; it earns nothing from your losses and holds no position against you. As our plain-English explainer covers, every contract is a yes/no question priced 1 to 99 cents that settles at $1 or $0, and until settlement the position trades like any other market. That is also the honest limit of the model: an order book needs two sides, so menus run thinner than a sportsbook’s and exotic combinations are still young.

The exit difference deserves one more sentence. A sportsbook’s cash-out button is an offer the house prices at its own discretion; an exchange exit is a sale to the open market at the going price. Both let you leave early. Only one of them lets you leave at a price the crowd set, and our DFS comparison makes the same point against contest-style products from the other direction.

What Each One Costs, Measured

Sportsbook pricing hides its cost in plain sight. A standard -110/-110 line carries a 4.55% hold, and nothing on the bet slip itemizes it. Because we archive odds continuously, we can do better than the textbook number: across 95,734 pregame snapshots in July 2026, snapshot-weighted median holds on the six-book regulated panel ran from 3.87% on the most liquid big-league moneylines (FanDuel MLB, n=37,813) up into the 6-8% range on spreads and totals in smaller markets. Odds data via The Odds API; the same archive powers our Divergence Board, which puts book prices next to exchange prices on the same games every day.

Exchange pricing itemizes. Takers pay a formula fee that peaks at 50-cent prices ($1.50 to $1.75 per 100 contracts), makers rest orders free on most Kalshi markets and earn rebates on Polymarket US, and our fees guide carries both venues’ full formulas with a calculator. What exchanges lack is the promo economy: no bonus bets, no boosts, no insurance. The sticker is the sticker.

Who Regulates What (and This Week’s Score)

Sports betting is state law. Since the Supreme Court struck down the federal ban in 2018, each state has decided for itself, and the licensed map now covers 39 states plus Washington, D.C., with Missouri the newest arrival (December 2025). A license buys real strings: geofencing at the state line, 21-plus verification in most markets, state taxes, and mandated consumer protections including statewide self-exclusion programs.

Event-contract exchanges are federal creatures. The CFTC has regulated them as derivatives venues since 2004, which is how they reach California and Texas without a single state gambling license, and which is exactly what states are now litigating. The last week captured the whole war in miniature: on July 27, 2026 a federal judge enjoined Minnesota’s first-in-the-nation exchange ban four days before it took effect, and on July 31 New York’s attorney general answered with a state-court suit asking a judge to bar Kalshi’s New York contracts, with the state estimating the damages it seeks around $36 billion. Neither case is decided; both venues keep operating while the courts work. The running per-state record lives in our state-by-state tracker.

Where the Map Splits: State by State

The two licensing models produce opposite maps. California, Texas, and Georgia have no legal sportsbooks at all, yet the federal exchanges operate in all three. Michigan, Massachusetts, Nevada, and Washington run licensed sportsbook markets, yet court orders keep sports event contracts dark in all four. Nine states where the split is sharpest:

New York deserves its own sentence: it is simultaneously one of the largest licensed sportsbook markets in the country and, as of July 31, the state suing Kalshi hardest. Both facts are true at once, which is the whole story of this map.

Where Each One Is Simply Better

Strip out the legal fight and the products still fit different people. Feature by feature, honestly:

• Sportsbooks win on menu depth. Same-game parlays, player props, live betting on every timeout, and a promo economy that pays you to show up. None of that exists at scale on an exchange yet.
• Sportsbooks win on mandated safeguards. Licensed states require self-exclusion programs, deposit limits, and problem-gambling funding. Exchange safeguards vary by venue: Kalshi publishes a toolkit; Polymarket US does not publish one.
• Exchanges win on exits. A position can be sold, trimmed, or hedged at the market price any time; a bet slip waits for the final whistle or the book’s cash-out offer.
• Exchanges win on transparent pricing. The fee is posted, the price is the crowd’s live probability, and there is no promo math to reverse-engineer.
• Exchanges win on reach. They operate in California, Texas, and Georgia today; sportsbooks legally cannot.

What neither side wins: a guarantee. The book’s hold and the exchange’s fee are both real costs, and no structure turns a coin-flip opinion into income.

Taxes: One Settled, One Not

Sportsbook winnings live in a settled, newly-tightened regime. For 2026, Form W-2G reporting applies at a $2,000 threshold (per the IRS W-2G instructions, inflation-adjusted going forward), withholding generally runs 24% on large long-odds payouts, and the One Big Beautiful Bill Act caps gambling-loss deductions at 90% of winnings starting this tax year, so even a break-even year can produce a tax bill.

Exchange winnings are unsettled: the IRS has issued no guidance, and practitioners argue for gambling-income, capital-gains, or Section 1256 treatment, with meaningfully different outcomes. Whether the new 90% loss cap even applies to exchange trading is one of the open questions. Our prediction-market tax guide walks the candidate treatments and the records worth keeping. Nothing on this page is tax advice; a CPA who has seen your numbers beats any article.

Which One Fits You

Match the machine to what you actually want. If you want Tuesday-night entertainment with a deep menu and promos, a licensed book is built for exactly that, and our sportsbook rankings compare the operators (start with the FanDuel review if you want to see how we grade a book). If you want positions you can manage, priced by a crowd instead of a house, our Kalshi vs. Polymarket comparison settles the which-exchange question on verified facts. And if the real question is which one is safer, that argument deserves its own essay: prediction markets or sportsbooks, safer or more dangerous takes it head-on.

Prediction Markets vs. Sportsbooks FAQ

The questions readers actually ask when they compare the two, answered from the sourced facts above.