NFL bettor biases and price setting: further tests of the Levitt hypothesis of sportsbook behaviour

Abstract

Empirical support for the Levitt hypothesis of sportsbook behaviour, where sportsbooks set prices to maximize profits, not to clear the market, is shown using data from actual sportsbooks. Betting percentages on favourites and underdogs (pointspread market) and overs and unders (totals market) were obtained using actual dollars bet (www.sportsbook.com) and percentage of bets made (www.sportsinsights.com). Both data sets reinforce the idea that sportsbooks are not setting prices to attract even betting dollars on both sides of the proposition. Big favourites, road favourites and overs on high totals are all shown to attract a significantly higher percentage of bets in both samples. Betting against public sentiment is shown to be statistically profitable for the National Football League pointspread market, but not in the market for totals.

Notes

¹Sportsbook.com is included as one of the four sportsbooks that share their percentage of bets data with Sports Insights. Although we use the data from sportsbook.com individually, the data are captured directly from their webpage and present the percentage of dollars bet, not the percentage of bets made on each side of the proposition.