Sportsbook vs prediction market arbitrage: how the math works, what the fees take, and the five ways a locked profit comes unlocked

Q: What is arbitrage between a sportsbook and a prediction market?

It is backing one side of a game at a sportsbook and buying the opposite outcome on an exchange such as Kalshi or Polymarket so that, after fees, the total cost is below the guaranteed payout. It exists when the implied probabilities of the two positions, fees included, add up to less than 100 percent.

Q: How do you calculate a sportsbook and exchange arbitrage?

Convert the sportsbook price to an implied probability (1 divided by decimal odds), add the exchange price per contract plus the fee per contract, and check the sum is under 1. At +120 (decimal 2.20) and 52 cents on Kalshi, the sum is 45.45 percent plus 52 percent plus a 1.75-cent fee, or 99.2 percent, a margin of about 0.8 percent.

Q: Is sportsbook and prediction market arbitrage risk-free?

No. The second leg can move before it fills, the two venues can settle the same game differently, sportsbooks can void bets at prices they consider an error, and they can limit stakes. In the worked example a one-cent move on the exchange turns a profit of about 80 cents into a loss of 20 cents.

Q: Is arbitrage betting legal?

Placing offsetting positions is not itself an offence in US states with licensed sportsbooks, but sportsbook terms allow operators to limit or close accounts, and prediction market availability varies by state. This is not legal advice.

Guide · Updated October 27, 2026 · 18+ · Not legal, tax or financial advice · By BetG8 Editorial · Sports

Sportsbook vs prediction market arbitrage, explained — Two venues, two prices, one game. The profit lives in the gap between them, minus everything that can go wrong.

Short answer: Sportsbook and prediction market arbitrage means betting one side of a game at a sportsbook and buying the other side as an event contract on an exchange such as Kalshi or Polymarket, so that the combined cost is below the payout whichever side wins. The test is one line: convert the sportsbook odds to an implied probability, add the exchange price and its fee per contract, and see whether the total is under 100 percent. In our example, +120 at a sportsbook against a 52-cent contract on Kalshi adds up to 99.2 percent after Kalshi's fee, a locked margin of about 80 cents on $99.20. That margin disappears if the exchange price moves by a single cent, and it can be voided, limited or settled differently before it is ever collected. This guide is arithmetic, not a recommendation to bet or trade.

LIVE MARKETPro Football 2027 ChampionTrade on Polymarket →

Gaps between sportsbooks and exchanges are real and easy to find, because the two kinds of venue price the same game in different ways. A sportsbook sets both sides and builds in a margin. An exchange matches traders and charges a fee, and its price moves with whoever is trading at the moment. When the two disagree enough, the arithmetic says you can own both outcomes for less than they pay. What the arithmetic leaves out is most of this guide.

The formula

A two-way arbitrage needs two positions that cannot both lose. At a sportsbook with decimal odds d, the implied probability is 1 ÷ d, so +120 (decimal 2.20) implies 45.45 percent and −110 (decimal 1.909) implies 52.38 percent. On an exchange, a contract that pays $1 if an outcome happens costs its price c, which already reads as a probability, plus a fee f per contract. The condition for a locked profit is:

1 ÷ d + c + f < 1

To lock the same payout on both legs, size them together. For a payout of $100, buy 100 contracts on the exchange and stake 100 ÷ d at the sportsbook. The total outlay is (100 ÷ d) + 100 × (c + f), and the profit in either outcome is $100 minus that outlay. Our implied probability guide covers the conversions from American and fractional odds.

A worked example, fees included

Suppose a sportsbook offers Team A at +120, and on Kalshi the contract on Team B winning, which is the same as Team A not winning in a game that cannot draw, trades at 52 cents. Kalshi's taker fee is 0.07 × contracts × price × (1 − price), rounded up to the cent, under its published schedule; Polymarket's sports category uses a rate of 0.05 in the same formula.

Step	Calculation	Result
Sportsbook implied probability	1 ÷ 2.20	45.45%
Exchange price	52 cents	52.00%
Gross gap	100% − 97.45%	2.55%
Kalshi fee on 100 contracts	0.07 × 100 × 0.52 × 0.48, rounded up	$1.75
Exchange leg cost	$52.00 + $1.75	$53.75
Sportsbook stake for a $100 return	$100 ÷ 2.20	$45.45 (returns $99.99)
Total outlay	$45.45 + $53.75	$99.20
Profit if Team A wins / Team B wins	$99.99 − $99.20 / $100 − $99.20	$0.79 / $0.80 (about 0.8%)
Same trade on Polymarket sports	fee 0.05 × 100 × 0.52 × 0.48 = $1.25	profit about $1.30 (1.3%)
Kalshi price moves to 53 cents first	$45.45 + $53.00 + $1.75	loss of $0.20
Sportsbook line moves to +110 first	$47.62 + $53.75	loss of about $1.37

Two things stand out. The fee consumed more than two-thirds of the 2.55-point gross gap on Kalshi, which is why a gap that looks generous on a screen is often thin in practice. And the trade is fragile in both directions: a one-cent move on the exchange, or a ten-cent move in the American line, wipes it out. You can run other prices through our sportsbook vs exchange price comparison and the fee calculator.

Why the gaps exist

A sportsbook quoting −110 on both sides of a spread is asking for 104.76 percent of implied probability, a margin of about 4.5 percent that our vig guide breaks down. To stay balanced it shades lines toward the side its customers favour, and it does not always move as fast as an exchange order book. An exchange has no built-in margin on the price, but its customers are a different crowd, its book can be thin away from the most popular games, and a single large order can push a price several cents. Gaps open when one venue reacts to news before the other, when a sportsbook leans a line against its own customers' bias, or when an exchange book is briefly one-sided. They close fast, which is the first problem.

The five ways a locked profit comes unlocked

Risk	Where it comes from	What it can cost
Leg risk	The second order fills at a worse price or not at all	A one-cent move turns the example from +$0.80 to −$0.20; an unfilled leg leaves a full one-sided bet
Settlement mismatch	The venues use different rules for overtime, postponements, retirements or official data	Both legs can lose, or one can be refunded while the other settles
Voided bets	Sportsbook house rules allow voiding at prices judged an obvious error	The sportsbook leg is cancelled and the exchange leg stands alone
Stake limits	Sportsbooks may cap what a profitable customer can bet	The trade cannot be sized; the profit per trade becomes trivial
Costs outside the formula	Deposit and withdrawal time, capital tied up until settlement, and tax treatment	Small margins can turn negative over a year

Settlement rules. Each exchange contract settles under its own written terms, and a sportsbook settles under house rules approved by a state regulator; they do not always agree on what happens if a game is suspended, moved or decided in overtime. Read both before treating two positions as opposites. Our guide to how prediction markets resolve covers the exchange side.

Voids. FanDuel's rules allow it to void bets at odds "materially different from those available in the general betting market at the time the bet was made", and DraftKings and bet365 have similar obvious-error clauses, Covers reported. A price that is far enough off the market to create an arbitrage is precisely the kind of price those clauses describe. Voids are not rare: a WCPO investigation found Ohio sportsbooks voided more than $77 million of wagers in their first 13 months, about 0.91 percent of the $8.5 billion handled.

Limits. Data the Massachusetts sportsbooks gave their regulator in 2025 showed 0.64 percent of bettors limited, with 12.7 percent of limited bettors capped below 1 percent of the default maximum stake; on September 30, 2025 the Gaming Commission moved to require operators to tell customers when they are limited. Exchanges do not limit winning traders in the same way, but their books have finite depth at any given price.

How to check a quoted arbitrage

Arbitrage services and social posts quote gaps before costs. Before believing one, run four checks. First, convert both prices to probabilities and add the exchange fee for your order size, not the headline price. Second, confirm the two positions are true opposites under both venues' rules, including draws, overtime and postponement. Third, check the depth: the exchange price shown may be good for only a few contracts, and the sportsbook may cap the stake. Fourth, decide which leg to place first; the one most likely to move should usually go first, and the plan should include what happens if the second never fills. A gap that survives all four is usually well under 2 percent before taxes. The expected value guide shows why a small negative surprise on a large number of trades matters more than one good trade.

Where both venues are available

An arbitrage needs both legs to be legal where you are. Licensed sportsbooks operate only in states that have authorised them, and prediction market sports contracts are unavailable in some states under court orders or settlements, including Illinois, Michigan and Nevada as of October 2026. Check our 50-state Kalshi tracker before assuming an exchange leg is open to you. Sportsbook terms also generally allow an operator to close or restrict an account at its discretion, which is a commercial risk rather than a legal one.

What this guide does not cover

This guide covers two-way arbitrage on games that cannot end in a draw, using taker orders. It does not cover three-way markets, middles, promotional or bonus arbitrage, maker rebates, or the tax treatment of winnings on each leg, which can differ between sportsbook wagers and exchange contracts and depends on your situation; our Kalshi tax guide covers the exchange side. Fee figures are from Kalshi's fee schedule effective July 7, 2026 and Polymarket's fee documentation as of October 2026; some Kalshi series use different multipliers and both schedules change. Nothing here is legal, tax, financial or betting advice, and nothing here suggests that any specific opportunity exists or is profitable.

Frequently asked questions

What is arbitrage between a sportsbook and a prediction market?