How de-vigging a sportsbook line works

The event contract (a digital)

A binary claim on a prediction-market venue that pays a fixed amount (commonly $1) if the stated side wins the game and nothing otherwise. It is cash-settled to the game's official result, fully collateralized: the stake is posted up front. Its price is, directly, the market-implied probability that the side wins.

The sportsbook moneyline (two-sided)

A pair of posted prices, one on each side of the same game, quoted in American odds. Each price pays a fixed amount on top of the stake if that side wins and loses the stake otherwise, settled to the same official result. Read as probabilities the two prices sum to more than one; the excess is the book's margin. A sharp book is one whose prices move quickly on informed money and carry a thin margin — the reference used here.

The Scopes Fair Value for "this side wins" is the de-vigged probability read from a sharp book's two-sided price. Each American price converts to a raw implied probability: a negative price of magnitude m implies m ÷ (m + 100); a positive price n implies 100 ÷ (n + 100). For a two-sided market those two raw figures sum to more than one, by the book's margin.

De-vigging removes the margin by rescaling. The simplest method — proportional, or multiplicative, normalization — divides each raw probability by their sum, so the pair sums to one while their ratio is unchanged. The result is the number a bettor would read off the book for the same payoff the event contract offers, net of the margin the book adds, computed independently of the prediction-market venue.

The normalization is not the only method, and the choice is not neutral. Proportional de-vigging assumes the margin sits on each side in proportion to its probability; alternatives — additive normalization, which subtracts an equal share from each side, or the Shin method, which models the margin as protection against better-informed bettors — distribute it differently and move the favorite's de-vigged figure by up to a point or more on lopsided lines. The de-vigged fair value is therefore a method-dependent estimate, not a single identity.

Method risk

The de-vigged number depends on the method chosen and on the assumption that the book's margin is distributed a particular way across the two sides. Two reasonable methods return two different fair values, most visibly on heavy favorites and longshots where the sides are far apart. A gap to a prediction-market price that is smaller than the spread between de-vig methods sits inside the method's own uncertainty, not a discrepancy to be reconciled.

The margin and transaction costs

Reconciling the two instruments crosses the prediction-market venue's fees and the sportsbook's margin on the other side. The vig is not a quote artifact to be wished away: a bettor who takes the sharp side pays it, and the prediction-market side carries its own fee and spread. A several-point gap can be inside the round-trip cost of establishing and holding both sides to settlement.

Line movement and timing

A sportsbook line and a prediction-market price are captured at moments that need not coincide, and sharp lines move on news, injuries, and money right up to kickoff. A gap measured against a line taken minutes earlier can be the two venues pricing different information, not a standing difference — the two prices describe the "same" game only when they are read at the same instant.

Limits and access

Sharp books cap stake size and restrict or close the accounts that consistently take the sharp side, and prediction-market venues impose their own position limits; the size at which the two instruments can be held against each other is small. Even a gap that survives the other frictions applies to a quantity too small to matter for a participant with the operational capacity to trade it.

The margin is not the probability. A raw implied probability read straight off a single posted price includes the book's margin, so it overstates the chance of that side. De-vigging is what separates the market's probability from the price of the book's service. A number quoted without de-vigging is a price, not a probability, and the two differ by the hold.

De-vigged is still market-implied, not frequency. Even after the margin is removed, the de-vigged figure is the market's clearing probability — it embeds the balance of money across the two sides and the book's risk management, not a forecast of how often the outcome occurs over many identical games. A gap between a de-vigged sharp line and a prediction-market price is a difference between two market-implied reads of the same outcome, informative about relative pricing across two venues, not a measure of true frequency.

The method assumes a sharp book. De-vigging recovers a clean probability only to the extent the input is clean: a sharp book with a thin margin and fast-moving lines de-vigs to a number close to the market's consensus, while a soft book with a fat margin and slow lines de-vigs to a noisier one. The operation is only as good as the book it is applied to, which is why the reference is drawn from the sharpest line rather than any line.

The participants who could reconcile a de-vigged sharp line with a prediction-market price are professional bettors and trading syndicates: they already model games, hold accounts at both the sharp books and the prediction-market venues, and can act on both sides. For them a gap of a few points sits against the book's margin, the prediction-market fee, the movement between the two capture moments, and the small size the limits allow — and the sharp books that price these games most tightly are also the quickest to restrict the accounts that lean on them.

Casual bettors face the frictions more sharply, not less: they typically hold accounts at softer books with fatter margins and slower lines, so the number they can act on is noisier and the margin they pay is larger, with none of the infrastructure that would make the reconciliation cheap. So a gap is not left open because it is unnoticed; it is left open because, for each participant able to act on it, the cost of doing so — margin, fees, timing, and restricted size — exceeds the gap at the size available.

That is the structural reason a de-vigged sharp line and a prediction-market price can sit apart. The de-vig recovers the market's probability net of the book's margin; it says nothing about which venue is better calibrated, or which side will move if the two converge — that is what the calibration record measures, over many resolved cases rather than one.