March Madness Handicapping: Sweet Sixteen Seeds Trends
by Trevor Whenham - 03/25/2009
A lot has been made of seeding in the NCAA Tournament this year - the absence of high seeds and the general chalkiness. The fact that not a single No. 1, No. 2 or No. 3 seed has been eliminated yet makes this year's event somewhat unique. The seeds of the teams involved are going to be an important part of your handicapping process. To help with that, here's a collection of thoughts and facts about seeds:
1. Expected wins - By looking back at what has happened since the tournament expanded to 64 teams in 1985 we can get a general expectation for the different seeds. We can learn, for example, that No. 1 seeds have won an average of 3.36 games per year. They have won two so far. No. 2 seeds have fared significantly better, winning an average of 2.43. Based on that, we could reasonably expect two of the four No. 2 seeds to lose in the Sweet 16 if they are to hit their average. No. 3 seeds have already exceeded their historical average - they win an average of 1.79 games per year. No. 4 seeds are at 1.52 wins per year, and the No. 5 seeds typically win 1.17. Obviously, Arizona is the team that has most over-performed compared to their seed - No. 12 seeds average just 0.48 wins per tournament. These are just long term statistics, so they don't necessarily mean anything for this year. They are trends with large sample sizes, though, so they could be useful to you as a tiebreaker if you can't otherwise find a difference between two teams.
2. Diminishing importance of seeds - Some academics get to do interesting studies. For example, someone at the University of Illinois has done a comprehensive study of the tournament. Among the outcomes was that seeds are less important as the tournament progresses. Seeds are obviously key in the first round - you can be confident that a No. 1 seed is going to beat a No. 16. The further we get along, though, the less the seeds matter. The study showed that the higher seeds have just a small advantage over the lower ones in the Sweet 16. By the Elite Eight the seeds mean virtually nothing. That goes in line with the expected wins data - the No. 1 seeds are expected to get more than one more win, but by the Elite Eight they are the only ones with the expectation for more wins.
3. Closeness of ones and twos this year - This year isn't unique, but it is striking - it was very hard to differentiate between a No. 1 and a No. 2. Louisville and North Carolina were clear No. 1s, but you could argue that any of the No. 2 seeds could just as easily been a No. 1 in place of Pitt or Connecticut. The same can be said lower down as well - some of the No. 3s and No. 4s could have traded places without raising any real questions. The only way that these teams were differentiated between was by the selection committee making a relatively arbitrary decision behind closed doors. What that says to me is that it's far more important to look at what the teams bring to the table this year than it is to look at the number beside their names.
4. Passing time - During the regular season would you handicap a game by ignoring what had happened in the previous 10 days? Of course not. So why would you do it in the tournament? The seeds were assigned four days before the first round began. Since then each team has played two games, and experienced all of the highs and lows and bangs and bruises that can occur in that time. In short, these are not necessarily the same teams that the selection committee assigned seeds to. For example, the Arizona Wildcats we have seen in the last week are far better than the ones that the committee thought was barely worthy of a bid. Relying too much on seeds is just relying on stale information.
5. Overvaluing seeds - This all boils down to one big lesson - you should pay attention to seeds, but they should be well down your list of things to consider. You could think, for example, that Louisville might beat Arizona because they have more depth, or because they are faster, or because Terrence Williams is a beast that is so hard for opponents to deal with. All of those reasons would be far more valid than just blindly assuming that Louisville will roll over the Wildcats because they are seeded 11 spots higher. Paying attention to the teams and not just their seeding is so important because the oddsmakers and the public both seem to assume that the No. 1 seeds have a distinct advantage. That may be the case, but it isn't automatically so.