What causes the favourite-longshot bias? Further evidence from tennis

Abstract

In sports betting markets, bets on favourites tend to have a higher expected value than bets on longshots. This article uses a data set of almost 45 000 professional single tennis matches to show that the favourite-longshot bias is much stronger in matches between lower-ranked players, in later-round matches and in high-profile tournaments. These results cannot be solely explained by bettors being locally risk-loving or overestimating chances of longshots, but are consistent with bookmakers protecting themselves against both better informed insiders and the general public exploiting new information.

Keywords:

• favourite-longshot bias
• tennis
• sports betting
• market efficiency

JEL Classification:

• L83
• G14

Notes

¹ The data set was downloaded from the website tennis-data.co.uk on 22 June 2013. The men’s tennis matches start in 2002; the women’s matches start in 2007. The betting odds are the latest available odds by the bookmaker Bet365. Originally, there were 48 042 matches, but 3171 matches (6.6%) were discarded due to missing odds or a withdrawal of one player before the match started.

² The estimation method in the whole article is OLS with heteroscedasticity-robust SEs.