Large spreads in college basketball
T.O. Whenham of Doc's Sports Predictions
This is the time in the college basketball season when top teams are looking for warm-ups against easy competition. Tough competition is on the horizon, either with non-conference games or the start of conference play, so athletic directors try to schedule games that they know their teams can win. Some teams are more obvious about this than others. My beloved Michigan Wolverines, for example, aren't among the elite in the country, but they have a very reasonable chance of being 12-0 before they play their first ranked opponent, UCLA, in December because of the weakness of their schedule until then. Of course, the simple scheduling can backfire in very embarrassing fashion, as Kansas learned this week against Oral Roberts.
The obvious result of these lopsided games is some gigantic spreads. When a good team meets up with a weak team we will often see double digits lines. That's especially true when one team is a heavily-hyped ranked team and the other is a team that few people know anything about. Since we see a lot of giant spreads in a season, and especially early in the season, it only makes sense that we have to figure out how to deal with them. One team is clearly much better than the other, as the spread reflects, but we have to decide whether to jump on them, whether to bet that the weak opponent will stay close or whether we should skip the game entirely.
The first kind of spreads we can consider are the truly gigantic ones - those of 20 points or more. In those cases the books are obviously saying that one team is dramatically better than the other. This kind of game doesn't happen a whole lot. Last season, major conference teams were involved in just 68 games with spreads of 20 or more. They covered 36 times and failed to 32 times. That's essentially meaningless. With a sample size that small, and results that evenly split, there is no clear conclusion to draw about how to play those spreads.
Within that group of games with huge spreads there are some games that are particularly interesting. By virtue of their name recognition and public appeal, some teams are going to face these spreads much more than the average. Duke, for example, played in 11 games in the 2004-05 season in which they were favored by at least 20. A team like Duke is going to get a lot of attention against a team perceived to be weak because of their popularity, and that means that they are going to have a lot of money bet on them. That, in turn, means that spreads could be higher than they should be. Duke's performance against the spread in those games, 5-6, shows that their spreads may not be entirely warranted all the time.
To me, this all means that I need a good reason to go one way or another before I touch a game with a big spread.
How about spreads that are still big, but not so huge? If we say that the line between a favorite and a heavy favorite is at 12 points, then there is some interesting data to look at. Justin Wolfers, a University of Pennsylvania economist, looked at results from 16 seasons of college basketball. Teams favored by less than 12 points covered about as much as they didn't cover. The heavy favorites, with spreads more than 12, only covered 47 percent of the time, despite the fact that they were clearly the better team.
Wolfers' research concluded that point shaving occurred in 5 percent of those games with heavy favorites. I'm not sure if that's true or not (though it seems a bit extreme), and it's not really relevant. Regardless of whether there are some dastardly deeds going on, there are lots of reasons why heavy favorites don't cover. The motivation to hold on to win by 20 when the game is well in hand might not be high - you aren't going to be as tight on D when you're up by 17 in the last minute. Coaches can use a big lead as a chance to give bench players playing time. Teams built around one or two stars can struggle when those players have a bad day. An injury or early foul trouble can change the course of a game. The fact is, obviously, that it is much easier to win a game than it is to win it by 13.
Like pretty much everything involving handicapping, this isn't a case where you can apply the data to make quick money. It also doesn't point towards whether you should play these games or avoid them. The simple but unhelpful answer is that it depends. There is certainly no reason to avoid playing an underdog facing a large spread, and you also don't need to avoid playing a heavy favorite if you have a good reason to do it. The only thing that we have learned is, if all other things are equal in your analysis, you'll do better in the long run by backing the underdog that doesn't stand much of a chance.
