Implied probability & the bookmaker’s margin, explained
Last reviewed · Editorial policy
How betting prices translate into probabilities, where the bookmaker’s cut sits, & why the margin in the price differs from what a book keeps.
Implied probability in one line
Implied probability is the chance of an outcome that a price assumes. It is also the win rate a bettor needs to break even at that price.
Kalshi’s fee schedule puts it plainly for its own contracts: a contract’s implied probability is its price divided by $1. A 50-cent contract therefore implies 50%.
American sportsbook odds carry the same information. According to Covers, a bet at +100 needs to win 50% of the time to break even, a bet at -110 needs 52.4%, & a bet at +200 needs 33.3%.
Where the margin hides
The margin, also called the vig or vigorish, is the commission a bookmaker builds into its prices. In a 2004 paper in The Economic Journal, economist Steven Levitt describes the typical terms: bettors pay 110 units if a bet loses but are paid only 100 units if it wins.
Levitt quotes Lee & Smith on what this means when a book attracts equal money on both sides: “the losers pay the winners $10 & pay the bookmaker $1, no matter how the game turns out.”
The margin shows up in the breakeven rate. According to Covers, a bet at +100 needs to win 50% of the time to break even, while a bet at -110 needs 52.4%.
Margin isn’t spread evenly
A book does not have to balance the money on each side to profit. Using about 20,000 NFL wagers placed by 285 bettors in an online handicapping contest, Levitt argues that bookmakers “are more skilled at predicting the outcomes of games than bettors & systematically exploit bettor biases by choosing prices that deviate from the market clearing price.”
His illustration: if bookmakers know local bettors prefer local teams, they can skew the odds against the local team. He notes a limit on this, since bettors who know the correct price can profit if the posted price strays too far from the true odds.
Theoretical margin vs hold
The margin built into prices is a theoretical figure. What a book actually keeps is the hold, & the two are often confused.
Wizard of Odds defines the house edge as “the ratio of the average loss to the initial bet”, & the hold as the ratio of chips a casino keeps to the total chips sold. In its example, a blackjack table that sells $1,000 in chips & keeps $300 has a hold of 30%.
The site adds that a mathematician alone cannot determine the hold, because it depends on how long players stay & how the same money circulates back & forth.
In sports betting, hold is measured against handle, the total amount wagered. According to Covers, the New York State Gaming Commission reported that the state’s eight online sportsbooks held 2.9% of a $594.6 million handle in the week ending 13 September 2026, the opening week of the NFL season.
That produced $16.96 million in gross revenue & was the third-lowest hold for an NFL regular-season week since New York’s online betting launched in early 2022. Covers links it to results: eight favourites covered the spread on the Sunday, & nine of the day’s 13 games went over the points total.
Holds varied widely by operator that week. Covers reports 3.7% at FanDuel, 2.2% at DraftKings, 1.1% at Fanatics & 5.8% at BetRivers, which shows how far a single week’s results can move what a book keeps.
Fees count too
A flat fee on each bet changes the breakeven rate in a way the odds alone don’t show. Covers reported on 7 October 2026 that FanDuel had reinstated a 50-cent charge on qualifying Illinois wagers, passing along the state’s per-bet tax, which first took effect on 1 September 2025.
Covers describes the Illinois tax as 25 cents on each of an operator’s first 20 million wagers & 50 cents on all additional ones. FanDuel’s charge does not apply to straight bets of at least $25, parlays of at least $10 or wagers made with a bonus bet, & Covers says FanDuel charges no per-bet fee in 23 of its 25 markets (24 states & Washington, DC).
According to Covers, the fee raises the breakeven rate on a $1 bet at +100 from 50% to 66.7%, & at -110 from 52.4% to 71%. The effect shrinks as stakes rise: at +100 the breakeven rate is 51.3% on a $10 bet & 50.5% on a $24.99 bet.
Exchanges handle costs differently. Kalshi’s fee schedule says its trading fees are charged as a variable percentage of the expected earnings on a contract, calculated by multiplying the maximum potential earnings by the implied probability of making those earnings.
Kalshi’s fee schedule, effective 7 July 2026, sets the general fee on immediately matched orders at 0.07 × number of contracts × price × (1 − price), rounded up. The fee peaks at a 50-cent price: $1.75 for 100 contracts at $0.50, against $0.07 for 100 contracts at $0.01 or $0.99.
Orders left resting on the order book pay a smaller maker fee, & only in markets where the schedule says it applies.
Comparing books & exchanges
Implied vig is the margin read from a market’s prices. According to data from Citizens reported by Covers, Kalshi’s implied vig in Week 1 of the 2026 NFL season was 4.32%, against 4.44% at FanDuel & 4.51% at DraftKings.
Citizens tracked pregame moneyline & over/under prices on the Friday before the first Sunday slate, across 28 data points. A year earlier the order was reversed: Kalshi’s implied vig over the NFL season was 4.84%, against 4.42% at FanDuel & 4.48% at DraftKings.
Citizens credits the change to higher trading volume, more market participation & more competition among liquidity providers.
Multi-leg bets showed a different picture. Using the implied vig on overs & favourites for every NFL game, Citizens found Kalshi’s combos, its equivalent of sportsbook parlays, were priced 8% worse than at DraftKings & FanDuel, before accounting for transaction fees.
These figures measure prices, not results. In the same opening week in which Citizens put the two sportsbooks’ implied vig at about 4.5%, New York’s online books held 2.9% of their handle.
Questions
What probability does -110 imply?
According to Covers, a bet at -110 needs to win 52.4% of the time to break even. Levitt describes this as the typical structure in which bettors risk 110 units to win 100.
Is a sportsbook’s hold the same as its margin?
No. Wizard of Odds defines the house edge as the ratio of the average loss to the initial bet & hold as the share of money the house actually keeps, & says the hold term is worth defining because it comes up a lot. New York’s online books held 2.9% of handle in the week ending 13 September 2026, according to Covers.
Does Kalshi have a margin?
Kalshi charges a trading fee that is largest at a 50-cent price. Citizens measured an implied vig of 4.32% on Kalshi in NFL Week 1 of 2026, according to Covers.
How does the Illinois per-bet fee change breakeven rates?
According to Covers, FanDuel’s 50-cent fee raises the breakeven rate on a $1 bet at +100 from 50% to 66.7%. The effect shrinks with larger stakes, & straight bets of at least $25 are exempt.
Sources
Sources checked 9 October 2026. Information only. 18+.
Discussion
Ask a question, add useful context or share a source. Keep it relevant & respectful.
Add a comment