NBA odds comparisonNBA odds comparison — books vs prediction markets

Right now BeatVig is comparing 1,615 bets across NBA where one venue is paying measurably more than the rest of the market for the same outcome — game moneylines, spreads and totals alongside every player prop we price.

What "pp" means. Every price implies a win rate you would have to hit for the bet to break even: +108 implies 48.1%, -140 implies 58.3%. The headline on each row is how many PERCENTAGE POINTS the best price knocks off that number against the baseline — so "+11.4pp" means a bet that needed a 59.5% win rate at the rest of the market needs 48.1% here. It measures the same thing at every price level, which is exactly why it is the number the board is ranked by.

The "pays X% more" line underneath is the same comparison in payout terms: both prices converted to decimal odds — the total return per unit staked — and the best one measured against the baseline. It is true, and it is the more flattering number, but it may not rank the board. On a deep longshot a ratio measures how close to zero its own denominator is rather than an edge: +35000 against a market median of +17500 does pay double, and is 0.29 points of probability. Ranking by it put the board's smallest real advantages on top.

American prices are not on a linear scale for either number — -108 and -110 are not "2 better" of anything — so every price is converted before anything is compared.

The baseline is the median of the other venues on the same outcome, not the worst of them. Comparing against the worst price on the board would inflate every number on this page. The median is the price you would plausibly have taken had you not seen this comparison. A bet quoted by fewer than three venues in total is left off entirely — with one other price there is no market to be better than — and so is anything under half a percentage point, where the precision would be false.

Spreads and totals are only ever compared at the same line, and on the same side of it. A price on a 9.5-point favourite and a price on a 9.5-point underdog are opposite bets, not two quotes on one bet, so the board picks the line the most venues quote and compares within it. Player props are grouped by their exact line for the same reason.

Daily fantasy pick'em apps are excluded everywhere on BeatVig. They list nearly every market at the same price because their margin lives in the parlay multiplier, so including one would manufacture a gap on every row here.