Search This Blog

Monday, March 27, 2006

Overperforming and Underperforming

Here is a great post by Rocky (a Pitt fan) on

Overperforming and Underperforming

There's been a fair amount of discussion about which teams are over or under seeded in the tournaments, and which teams' reputations get them favorable or unfavorable seeds. So I figured I'd do some numbers.

I took all of the tournaments in the 2000's (2000-2005) prior to the current one. I looked at each team's seed, and compared that with how many games that seed would say they should win. That is, seeds 9-16 should win 0 games, seeds 5-8 should win 1 game, 3-4 should win 2 games, #2 seed should win 3 games and the #1's should win 4 games. (I've ignored the play-in game).

For each team I calculated the difference between how many games they "should have won" and how many they actually did win.

Basics: there have been 163 different schools making the round of 64 in that period.
65 teams were in just once
43 were in 2 tournys
20 were in 3
15 were in 4
7 were in 5
13 were in all 6

Of those teams, the ones that performed most over their seeds:
Mich St. +6 (in 6 tournys)
Wisconsin +5 (6)
Kent St +4 (2)
Tulsa +4 (3)
Missouri +4 (4)

Those that performed worst vs their seeds:
Stanford -9 (6)
Cincinnati -6 (6)
Wake Forest -5 (5)
Kentucky -5 (6)
Oklahoma -4 (5)
Duke -4 (6)

Seeing schools like Kentucky and Duke on the most over seeded list would seem to confirm a bit of reputation bias. Though it also looks like there is some bias against the Big 10.

Stanford though is a case unto itself, having underperformed in every year except 2002 (when they did as expected):

Stanfords seeds, wins, expected wins and difference:
2005 8/0/1/-1
2004 1/1/4/-3
2003 4/1/2/-1
2002 8/1/1/0
2001 1/3/4/-1
2000 1/1/4/-3

No comments: