Proof there is point-shaving in college basketball. Well, you can decide whether it’s proof or not, but it’s a pretty strong statistical case. Basically, he found that heavy favorites (12 points or more) tend to underperform by a few points, which he reads as an indication that players are purposefully playing poorly for stretches to get the final margin under the Vegas line.
There’s an alternate theory the author doesn’t deal with. Teams that are heavily favored are also more likely to pull their starters with several minutes to go and play their (not as good) bench players. Typically (although not always), the losing team will keep its starters in a longer period of time, which would have the effect of lowering the margin of victory. In a close game, the starters would never get similarly pulled.
But that’s the rub with this sort of “forensic economics” (his term); it’s great at finding broad trends in the data but not so good at determining whether or not they apply in any individual case. It’s all macro, not micro. Which makes it perfect for the sort of pop-economics analysis that has gotten so popular these days (viz Freakonomics, Gladwell, etc.) — it’s a high-brow comfort food for yuppies. (And I don’t mean that as an insult.)
The most amazing thing in the paper for me was just how accurate the Vegas line really is. It’s designed to maximize profits for the casinos, so a perfect result would be if favorites beat the line exactly 50 percent of the time. Well, over 44,120 games studied from 1989 to the present, favorites beat the line exactly…50.01 percent of the time. Those guys in Vegas know what they’re doing.