104.1% vs 100%: Prediction Markets vs Sportsbooks, Who Takes More?
On January 17, 2026, Lines.com had FanDuel pricing Bills–Broncos at Bills −100 and Broncos −118. Those two prices imply 50% and 54.1%, which adds up to 104.1%. Kalshi had the same game at $0.48 and $0.52, which adds up to exactly 100% before its disclosed fee. That 4.1-point gap is the bookie's cut, sitting right on the screen. It's real, but it isn't the whole story, because Kalshi still collects a fee. It just bills you for it separately instead of building it into the price.
The 4.1 Points FanDuel Hopes You Never Add Up
Start with 104.1, because that number is the whole business model.
Every American price converts to an implied probability, and you can check any game yourself in about ten seconds:
- Negative odds: odds ÷ (odds + 100). Broncos −118 becomes 118 ÷ 218 = 54.1%.
- Even money: −100 becomes 100 ÷ 200 = 50%.
- Positive odds: 100 ÷ (odds + 100).
Add both sides together. In a fair market the two sides total 100%. FanDuel's Bills–Broncos prices totaled 104.1%, and the extra 4.1 points is the margin you pay just to have an opinion. Kalshi's 48¢ plus 52¢ came to 100¢ flat. If you want the full toolkit for working out break-even rates, here's how to run the odds math on any price.
The difference comes from two different business models. Sports Illustrated puts it plainly: a sportsbook takes the other side of your bet, so it profits by setting odds that favor the house over time. A prediction market collects trading-volume fees whether you win or lose.
According to California Prediction Markets, sportsbooks build a 4–6% vig into their odds, while prediction markets charge an explicit, disclosed fee. The house didn't leave. It just started sending an itemized bill.
Hold Went From 7.5% to 10.2%, and Prediction Markets Noticed
In 2021, US sportsbooks kept 7.5% of every dollar bet. By 2025 they kept 10.2%.
Track360 puts the 2025 figures at $166.9 billion in handle and $16.96 billion in gross gaming revenue. Two companies collect most of it: FanDuel holds roughly 37–44% of online GGR and DraftKings roughly 34–36%, so together they control about three-quarters of the market.
Where does a 10.2% hold come from when straight lines carry 4–6%? It comes from same-game parlays and props, where the margins are widest. That climb is the cost of those bets, and the bettors who play them pay it. Parlays are a tax on hope, and 2025 was a record collection year.
Bettors have started to notice. ESPN reports that the American Gaming Association projects $29.5 billion will be bet on the 2026 NFL season with domestic sportsbooks, up just 0.3%, and the AGA names prediction markets as a major factor. From last September through May, sportsbook handle grew 4%, compared with 14% over the same stretch the year before.
The other side of the ledger is growing fast. Rotowire expects NFL trading volume on prediction markets to double this season, potentially reaching $36.8 billion. Eilers & Krejcik Gaming found August NFL prediction-market volume running 4.6 times higher than the previous August, per ESPN.
Where Each One Actually Beats the Other
Neither platform wins every bet, and Action Network's breakdown shows where each one does.
Use the exchange for:
- Heavily traded sides and totals
- Super Bowl and division futures
- Getting out of a position early, before the game ends
Use the sportsbook for:
- Passing-yard and anytime-touchdown props
- Live betting
Here's the catch on exchanges: California Prediction Markets notes that on thin markets with wide spreads, a prediction market can be the worse price. A Week 14 matchup nobody's trading is not Bills–Broncos in January.
The fee structure matters too. FOX Sports reports that Kalshi and Polymarket both use probability-weighted fees that are cheapest on contracts priced away from 50¢. That means a coin-flip NFL side, the exact bet most weekend players make, is where the fee bites hardest. So run the full math on a 48¢ contract instead of assuming it beats −100.
One more trap: FOX Sports also notes that Plus500 isn't its own prediction market. It offers Kalshi's sports contracts at Kalshi's prices, so checking both counts as one quote, not two.
That rules out both popular pitches. "No house means you can't lose" ignores the fee and the thin-market spreads. "It's a scam" ignores a 48¢/52¢ market that beat FanDuel by 4.1 points before fees. Fees add up the same way whichever slogan you prefer.
The Boring Edge That Works in Both Places
On Bills–Broncos, the same opinion on the same game cost 4.1 fewer implied points on one screen than the other. You didn't need better handicapping to get that, just a second tab.
That's why I keep saying line shopping earns more than handicapping for 90% of bettors. Before every side, total and future this season, compare two numbers:
- The sportsbook's combined implied probability (anything over 100% is the margin).
- The exchange's contract price plus its fee at that price level.
Take whichever is cheaper. Do it every time, including the weeks you feel sure about your pick.
Bet singles. A single only has to beat one margin. A parlay stacks a margin on every leg, and that stacking is what pushed industry hold to 10.2%. Three legs means three cuts.
Set a fixed weekend bankroll and don't add to it. Pick the amount on Thursday, not after the late window goes sideways on Sunday. Gambling is entertainment with a price tag, and the goal is to pay less of it to the margin, not to turn it into income. Shopping lowers what you pay per bet, and the bankroll caps what you can pay in total.
DraftKings — Bet $5, Get $200 in Bonus Bets (code PBDK200) A second sportsbook tab is the cheapest line-shopping upgrade there is, and this one opens with a $5 single, not a parlay. 21+ only, T&Cs apply. Bet Now
Before You Deposit: Know Whose Rules You're Under
The legal ground under prediction markets moved twice in 2026.
In April 2026, the Third Circuit sided with Kalshi in its case against New Jersey, finding that its sports contracts fall under federal commodities law. A few months later, Action Network reports, a federal court in Utah reached the opposite practical result and allowed the state to enforce its gambling laws against Kalshi. Where you live decides which ruling you're living under, so check before you fund anything.
The age gap matters too. ESPN reports that prediction markets let customers trade from 18, while regulated sportsbooks require 21 in most states. AGA CEO Bill Miller has accused prediction markets of "misleading" consumers, refusing to follow state and tribal sports betting laws, and avoiding gaming taxes. Track360 reports the AGA's estimate that sports event contracts diverted $500 million-plus in potential state tax revenue over the trailing year, because prediction-market operators pay no state gaming taxes.
Your next step is simple:
- Open one licensed sportsbook account and one CFTC-regulated exchange account, after confirming each is legally available in your state.
- Price this weekend's side on both: add up the book's implied percentages, then add the exchange's fee to its contract price.
- Bet the cheaper one, as a single, from a bankroll you set before kickoff.
FanDuel — No Sweat First Bet up to $1,000 (code PBFD) FanDuel is the book in our Bills–Broncos comparison, so price your first single there against the exchange, and judge the offer by its playthrough terms, not the headline. 21+ only, T&Cs apply. Bet Now
You must be 21+ to bet with regulated sportsbooks, and gambling should stay entertainment. If it stops being fun, stop, and call 1-800-GAMBLER for free, confidential help.
Bottom line
Prediction markets didn't kill the house cut. They moved it off the price and onto a fee line, which on liquid NFL sides can mean a cheaper bet and on thin markets can mean a worse one. FanDuel's 104.1% against Kalshi's 100%-plus-fee is the comparison to run every week, on every bet. Use the exchange for liquid sides, totals and futures, the book for props and live betting, bet singles, and keep your bankroll fixed. In 2026, that habit beats any pick.