Dr. Justin Wolfers, of the Wharton School of Business at the University of Pennsylvania, is somewhat unique among academic economists. With studies such as "Prediction Markets in Theory and Practice" and "Explaining the Favorite-Longshot Bias: Is it Risk-Love, or Misperceptions?", Wolfers' has become one of the more interesting and well-published economists studying betting markets and sports wagering. Among academic researchers, Wolfers' has become somewhat of a must-read for serious handicappers. A previous study, "Point Shaving: Corruption in NCAA Basketball" sparked heated debate in both the gambling community and the mainstream media when it suggested that up to 1% of all college basketball games may involve "gambling-related corruption".
Last year, Wolfers, with Ph.D. student Joseph Price of Cornell University, moved his research into the realm of professional basketball with a paper entitled "Racial Discrimination Among NBA Referees". This study concluded that NBA players being officiated by same-race referees score up to 2-1/2% more points and earn 4% fewer fouls. Interestingly, NBA teams are 83% black, while officiating crews are 68% white, suggesting that the effects of racial bias may be widespread throughout the game. This study attracted plenty of media attention and the subsequent criticism of David Stern. Whether this has led (or will lead) to any changes in NBA officiating policies is unknown.
More recently, Wolfers has attempted to apply this discovered race bias to a more practical topic: the efficiency of betting markets. In a paper to appear in the Journal of Quantitative Analysis in Sports, "Racial Bias In the NBA: Implications in Betting Markets", Wolfers once again collaborates with Joe Price (now an assistant professor at BYU) and adds fellow BYU Cougar Tim Larsen. In this study, they use NBA game data from the 1991/92 season through the 2004/05 season (obtained from gambling website Covers.com) to evaluate whether the race bias of referees could be used to systematically beat the betting spread.
They find that, in games where the majority of officials are white, betting on the team expected to have more minutes played by white players beats the spread greater than 50% of the time, although not enough to profit at -110 odds. The same holds true for black officials and teams with more minutes played by black players. Further criteria based on differences in team racial composition can be used to bring the winning percentage to over 57% in particular situations. This suggests the NBA betting market hasn't adjusted to race bias yet - a market inefficiency.
However, just because you knew how to win yesterday doesn't mean you will know how to win tomorrow. Betting markets are dynamic and constantly evolving. The authors address this point in their conclusion, noting that their original study on referee bias attracted considerable attention, and may even have resulted in changes in NBA referee policies or attitudes. Alternatively, the underlying bias may remain but now is recognized by the betting markets, leading to an adjustment of lines so that the market inefficiency no longer exists.
When, and how quickly, the market will adjust is another question. This study, while fairly widely publicized, still is unlikely to attract the attention of the majority of gamblers. Furthermore, most handicappers have little knowledge of how to effectively utilize this race bias in adjusting their predicted margin of victory. Unless the underlying racial bias were to change (a possibility if the NBA makes adjustments based on Wolfers' research), it is likely that racial angles such as this may remain profitable for some time, though their value will certainly diminish.
A handicapper willing to dig a little deeper may find some additional ideas in Wolfers' research. One use may be in evaluating player proposition bets. Though a few percent bias alone is unlikely to overcome the heavy juice on props, it could be incorporated as one aspect of handicapping. For instance, if the Lakers are facing an all-white officiating crew, you might be more inclined to look for a sub-par performance out of Kobe Bryant, whether in points scored or rebounds or fouls committed, particularly if he is facing a team that will give above-average playing time to white players. Racial bias in other sports may also be an area that is ripe for investigation. One study has already been done on bias of umpires' strike zone based on the race of the pitcher; certainly similar situations exist in all other sports.
Regardless of whether you plan to use racial data in handicapping the NBA, these papers and others from Wolfers are worth the read for any serious handicapper. Not only do they provide in-depth snapshots of actual game data (for instance, in this case, data on home/away and underdog/favorite splits), they also present ideas for deeper analysis that can be used to discover your own unique handicapping angles.
Further information on Dr. Wolfers' work, as well as copies of his manuscripts and links to media coverage, are available at his webpage: http://bpp.wharton.upenn.edu/jwolfers/index.shtml.
If you would like to make or read comments about this article, you may do so by visiting the Mess Hall forum at MajorWager where a thread has been started. Please click HERE