Last updated: July 13, 2026. This article is educational analysis, not betting advice. Betting involves risk; only adults 21+ should wager, and only with money they can afford to lose.
Sportsbooks and prediction exchanges price the same games, but they don't always agree. When a book says a team is 44% to win and Kalshi says 48%, someone is wrong — and occasionally the gap is wide enough that you can take both sides and, if both legs fill and settle as expected, lock in a small margin regardless of who wins. That's a book-vs-exchange arbitrage.
This article explains why those gaps open, walks through a real worked example including the exchange fee, and — because this is EdgeFinder — spends just as much time on the ways it goes wrong. Arbitrage between books and exchanges is a real inefficiency, but it rewards careful people and punishes sloppy ones.
Why books and exchanges price the same game differently
A sportsbook sets a line, adds a margin (the vig), and adjusts based on the money it takes. A prediction exchange like Kalshi doesn't set prices at all — traders do, by posting bids and offers on event contracts that pay $1 if the outcome happens. Two different pricing mechanisms, two different crowds, two different speeds. Gaps open for a few structural reasons:
- Different user bases. Sportsbook lines are shaped by recreational betting flow — favorites, home teams, big names get overbet, and books shade lines accordingly. Exchange prices are set by whoever shows up to trade, which on smaller markets can be a thin crowd with its own biases.
- Slower repricing on the exchange. Books employ traders and feed-driven algorithms that move lines within seconds of news — a scratched starter, a lineup change, sharp money elsewhere. On an exchange, the price only moves when someone updates their orders. On a busy NFL market that's near-instant; on a Tuesday MLB total it can lag by minutes. Stale exchange quotes are where most gaps live.
- Fee structures push prices apart. A book's vig is baked into its odds. Kalshi charges an explicit trading fee on top of the contract price, so a Kalshi quote that looks better than a book's odds may not be after fees. (EdgeFinder shows Kalshi prices net of the trading fee for exactly this reason.)
- Different limits and different customers being limited. Books restrict winning customers; exchanges generally don't. Some sharp flow that would normally correct book lines migrates to exchanges instead, and vice versa. See why sportsbooks limit winners.
Most of the time the gap is smaller than the combined cost of the book's vig and the exchange's fee, and there's no trade. Occasionally it isn't.
A worked example, fee included
Say the Guardians are playing the Yankees. A sportsbook is offering the Guardians at +125 on the moneyline. On Kalshi, the "Will the Guardians win?" market has YES trading at 48 cents — which means you can buy NO at 52 cents, a position that pays $1 per contract if the Guardians lose.
Convert both to implied probabilities. At +125, a $100 bet returns $225, so the book is pricing the Guardians at 100 ÷ 225 = 44.4%. Kalshi's crowd is pricing them at 48%. Bet the Guardians at the book, buy NO on Kalshi, and the two positions together cost less than they pay out:
| Leg | Position | Cost | Pays if Guardians win | Pays if Guardians lose |
|---|---|---|---|---|
| Sportsbook | Guardians ML +125, $100 stake | $100.00 | $225.00 | $0 |
| Kalshi | 225 NO contracts at $0.52 | $117.00 | $0 | $225.00 |
| Kalshi fee | Taker fee (see below) | $3.94 | — | — |
| Total | $220.94 | $225.00 | $225.00 |
The Kalshi leg is sized at 225 contracts so that both outcomes pay exactly $225. Either way, you collect $225 on $220.94 committed — a margin of $4.06, or about 1.8% — provided both legs fill at these prices and both settle on the same outcome. The risk list below covers every way that proviso can fail.
The fee math
Don't skip this part. As of July 2026, Kalshi's standard taker fee is 0.07 × contracts × price × (1 − price), rounded up to the next cent, charged when your order fills against the order book (resting limit orders that get filled pay a lower maker rate on many markets). For our 225 NO contracts at 52 cents:
0.07 × 225 × 0.52 × 0.48 = $3.9312, which rounds up to $3.94.
Notice what the fee did: without it, this arb returns $8.00 on $217.00 — about 3.7%. With it, the edge falls to 1.8%. The fee roughly halved the profit. The fee is largest near 50 cents and shrinks toward the extremes, so arbs on heavy favorites and longshots lose less to fees than coin-flip markets. Many "arbs" that look real on raw prices are negative after the fee — always compute it before you place either leg. Our free arbitrage and no-vig calculators handle the two-sided stake sizing; the fee formula and current schedule are in Kalshi fees explained. Fee schedules change — verify the current one at kalshi.com before trading. Polymarket's US exchange also introduced taker fees on sports markets in 2026 (with reduced or rebated pricing for resting limit orders), so the same lesson applies there: price the fee first.
The honest risk list
A true arbitrage has no risk once both legs are filled at the expected prices and both sides settle on the same outcome. Every word of that sentence can fail.
1. Settlement-rule mismatches
This is the sneaky one. Your two positions are only a hedge if the book and the exchange grade the same event the same way. They don't always:
- Shortened and postponed games. Baseball is the classic minefield. Books and exchanges have different rules for rain-shortened games, suspended games that resume the next day, and postponements — one side may grade a shortened game as official while the other voids and refunds. If one leg voids and the other stands, you're not hedged anymore; you're just betting.
- Totals and overtime. Confirm whether each side's market includes overtime or extra innings. A total that settles "regulation only" on one venue and "including OT" on the other is two different bets.
- Void windows. Books often void bets if a game isn't completed within a set window (commonly around 24–48 hours); exchange contracts settle per their own written rules, which may allow a longer window or settle on the rescheduled game.
The fix is unglamorous: read the market rules on the exchange contract page and the house rules at the book before placing either leg. Every Kalshi contract publishes its exact settlement terms.
2. Fill risk and leg risk
Exchange order books can be thin, especially outside marquee games. The 52-cent NO quote might only have 60 contracts behind it; buying 225 could walk the price up to 54 or 55 cents and erase the edge. Worse is leg risk: you place the book bet first, then the exchange price moves before you complete the hedge. Now you hold a one-sided bet you never intended. Mitigations: check depth before you start, fill the thinner (exchange) leg first, and size to what the order book actually shows — not the top-of-book quote.
3. Stake sizing and capital drag
A 1.8% edge means you need meaningful capital deployed to earn meaningful dollars, and that capital is locked in two places until settlement. Mis-sizing the legs — easy to do when one side is in American odds and the other in cents — turns a lock into a lean. Round-number stake limits at books, minimum tick sizes on the exchange, and withdrawal timelines all add friction. And books notice arbitrage-shaped betting: accounts that consistently beat the closing line get limited, which shrinks your future capacity.
4. Regulatory ground that's still shifting
As of July 2026, Kalshi is a CFTC-regulated exchange, and in April 2026 the Third Circuit held that the CFTC likely has exclusive jurisdiction over its sports event contracts. But several states are still contesting access — sports contracts are restricted or contested in a handful of jurisdictions, and litigation is ongoing. Availability can change on short notice. Check your jurisdiction and the exchange's own eligibility rules before assuming both legs are available to you.
An inefficiency that rewards care
Book-vs-exchange gaps are real, and they exist for durable structural reasons: different crowds, different repricing speeds, different fee models. But the edge per trade is small, the fee eats a large fraction of it, and every risk above is a way to convert a lock into a loss. The people who make this work treat it like operations, not gambling: rules read in advance, fees computed before entry, legs sized precisely, depth checked.
If you want to see how exchange and book prices actually compare in practice, EdgeFinder displays de-vigged fair prices from dozens of sportsbooks next to Kalshi and Polymarket quotes — with Kalshi shown net of the trading fee — and logs every model pick, win or lose, on our public record page. Start with the arbitrage basics article if two-sided betting is new to you, and use the free calculators to check every number yourself. That's the whole point: nothing here should be taken on faith.