Wednesday, March 10, 2010

Does NCAA Tournament Bracket Placement Matter?

Let me share a few quotes:

Quote 1: It doesn’t matter what seed you get as long as you are in. (Various)

Quote 2: It doesn’t matter if you are a 1 seed. If you are a 1, 2, or a 3 and you have a good team, you should be able to get to the Final Four. (Bob Knight)

Quote 3: I’d much rather be a 6 seed than a 5 seed anyway. (Fan)

The first reaction to Quote 1 should be a laugh. After all, 1 seeds are much more likely to reach the Final Four. But do they reach the Final Four because they have an easier path or because they are better teams?

One way to answer this question is to assume the beautiful log5 predictions of Ken Pomeroy are correct, and show how team's odds change in different bracket positions. And that’s exactly what I’m going to do today.

Consider a good, but not great NCAA team in last year’s NCAA tournament. (I’m looking at someone with a .9550 Pythag Winning Pct.) How likely would this team have been to advance to the Final Four in various bracket slots?

Best Slots
Final Four Odds - Slot
28.9% - 1 Midwest
25.0% - 1 East
24.5% - 1 South
21.5% - 2 East
20.2% - 4 East
19.5% - 2 Midwest
17.9% - 3 East
17.5% - 1 West
16.8% - 2 West
16.4% - 2 South
15.9% - 4 Midwest
15.2% - 6 East
15.0% - 6 Midwest

Worst Slots
Final Four Odds - Slot
6.3% - 16 West
6.2% - 15 West
6.1% - 9 West
5.9% - 14 West
5.9% - 11 West

Note that I’m projecting a good, but not elite team in each slot. Louisville’s Final Four odds last year were closer to 40%.

But I actually think this provides some decent support for Quote 1. The odds are not impossible in even the worst slot and the odds are not fantastic in even the best bracket position.

As for Bob Knight’s point, certainly 1, 2, and 3 seeds generally have better odds of making the Final Four. But what these numbers should emphasize is that the bigger question is who you draw in your region.

Last year the West was brutally stacked. To get to the Final Four, Connecticut, the 1 seed in the West was projected to face the tournament’s toughest 8 seed in BYU, one of the tournaments toughest 4/5 match-ups in Purdue or Washington, and either a dominant Memphis or Missouri team in the Elite Eight. On paper, Duke had an easier path with a 2 seed in the East.

Certainly, 1, 2, and 3 seeds have some of the easier Final Four paths, but region plays a huge role. And I find even better support for a long-held belief by fans. Sometimes being a 6 seed and facing a number of good but not great teams is an easier path to a deep tournament run.

Finally, in terms of getting to the Final Four, the 16 seeds are not the worst Final Four slot. Theoretically if you knock off the 1 seed, you take over their favorable draw. An 11 seed that has to face a strong 6, 3, 2, and 1, could have worse Final Four odds. The reason 16 seeds never advance deep in the tournament is not their slot. It is because 16 seeds aren't very good.

After the brackets are announced, check out Basketball Prospectus for the official log5 breakdown, a complete break down of tournament odds for every team. It is the combination of team quality and likely opponents.

But as a plug for myself, stop by late Sunday night and I’ll give you a statistical answer to the long-held question. What is the toughest region and toughest slot in the tournament? The TV announcers always like to say what they feel is the toughest. I’ll give you the statistical answer.