The Boston Celtics are expected to win Game 1 of the NBA Finals on Thursday night, which certainly makes sense if you look at the numbers. The Celts finished the season nine games ahead of Los Angeles. Their scoring differential was by far the best in league, a full 40 percent better than the Lakers. And the majority of the championship series are scheduled to be played in Boston, where the home team has won 10 of 11 throughout the playoffs. But Game 1 notwithstanding, the bookies in Vegas are reportedly favoring the Lakers to win it all. What are they thinking?
I hope they're not relying on the superiority of the Western Conference: This season's Celtics were even stronger against the West than they were in their own conference, at one point racking up 16 wins in a row against the Pacific Coast juggernauts. It would be just as ill-advised to play up the "playoff experience" of Lakers' coach and basketball Buddhist Phil Jackson. (As I recall, Gregg Popovich has also had some modest success in the postseason.) No, I'm guessing the oddsmakers have keyed in on something else about the 2007-08 Los Angeles Lakers … something that might give them an edge in a close series against a more skilled opponent. Maybe they've noticed that the Lakers are a very, very whiteteam.
There's reason to believe that fair skin gives you an advantage in the NBA. Last spring, economists Joseph Price and Justin Wolfers published a careful analysis (PDF) of league statistics and found evidence of racial bias among the referees. According to their research, the numbers of fouls called against white and black players varied depending on the race of the referees for that game—when there were more white officials on the floor, fewer fouls were called against white players. And since the majority of the league's referees are white, this puts minority players at a disadvantage. (To be exact, the data showed only a relative effect—so it's impossible to know which direction the bias went. White refs may favor white players, or they may discriminate against blacks. Or, black refs could just as well be favoring black players or discriminating against whites.)
The commissioner's office denounced the findings, but the data it offered as a counterpoint turned out to be unconvincing. Then, last month, Price and Wolfers dropped another bombshell on the league. By comparing their previous data with Vegas betting lines, the economists tried to show that a betting strategy based on referee bias would systematically beat the spread (PDF). They found that having one additional white player on the floor could make a difference of half a point in a game's outcome, given a majority-white crew of referees. Their data also indicate that betting lines have tended to underrate teams with more white players. Have the bookmakers in Vegas finally wised up?
Take a look at the rosters for the NBA Finals. Just two players on the Celtics are white—Scot Pollard and Brian Scalabrine—and neither one has so much as set foot on the hardwood during this year's playoffs. Meanwhile, the Lakers have two white players in their starting lineup, Pau Gasol and Vladimir Radmanovic, plus two more who get significant time in the rotation, Sasha Vujacic and Luke Walton. They've got a pair of white bench-warmers in Coby Karl and Chris Mihm and a regular player in the light-skinned, Jewish point guard Jordan Farmar. In other words, the Lakers are at least three times whiter than the Celtics.
Of course, it's hard to figure out what effect this will have on the officiating in the finals. For one thing, we don't even know yet which referees will be calling the games. What's more, Price and Wolfers teased out a tiny racial-bias effect in the NBA by sorting through a mountain of data—every single box score from 12 years of regular-season games. That needle-in-a-haystack approach has its pitfalls, since a failure to correct for a single variable can make a bit of random variation seem like something important.
Last December's study (PDF) of racial bias among umpires in Major League Baseball may have suffered from exactly that problem. A team led by economist Daniel Hamermesh looked at ball-and-strike data from more than 2 million pitches across three seasons and evaluated whether white umpires were more likely to make judgments in favor of a white pitcher. The study found that there was indeed an "own-race" effect when it came to called strikes and that the bias disappeared when umpires knew they were being monitored by the "QuesTec" computer system for determining balls and strikes. In other words, the umpires were biased when they knew they could get away with it.
But subsequent investigations by other researchers—notably Phil Birnbaum of the blog Sabermetric Research—revealed some flaws in the analysis. The Hamermesh team didn't properly account for every variable, like the score in the game or the time of day. (Pitchers tend to throw more strikes when there's a lopsided score, and umpires may make their calls differently in natural versus artificial light.) If there's any bias at all, Birnbaum concluded, it's probably that minority umpires favor their own. And the effect shown in the paper might be the work of one bad apple, as opposed to rampant racism throughout the league. But even if you take the paper's central finding at face value, the effect would be so minuscule that only one or two pitches would be affected per year for each pitcher.
Stat-heads haven't made as much of an effort to debunk the original Price and Wolfers study, but Birnbaum has cast serious doubts on the betting-line follow-up. (The results, he argues, may have more to do with the fact that white players are underrated than with racial discrimination by referees.) Still, there's some reason to believe that there's more racial bias in basketball than in baseball. Wage-discrimination studies found that salaries for white NBA players in the 1980s were 20 percent higher than they were for black counterparts with similar stats. They may have been worth the extra money: Each additional white player on a team also seemed to account for a paid-attendance boost of 13,000 fans per season.