RebKell's Junkie Boards

[beknighted]

For the last (okay, of two) Fun with RPIs for this season, I’m looking at how the tournament committee’s seeding compared with RPI in predicting results. It’s a pretty simple exercise.

Upsets versus seeding (14)

1st round (7):

12 BYU over 5 NC State
11 JMU over 6 Gonzaga
11 Florida over 6 Dayton
10 Florida State over 7 Iowa State
9 St. Joseph’s over 8 Georgia
9 Oregon State over 8 Middle Tennessee
9 Arizona State over 8 Vanderbilt

2nd round (4):

12 BYU over 4 Nebraska
7 DePaul over 2 Duke
7 LSU over 2 WVU
5 Oklahoma State over 4 Purdue

Sweet 16 (2):

4 North Carolina over 1 South Carolina
4 Maryland over 1 Tennessee

Elite 8 (1):

4 Maryland over 3 Louisville


Upsets versus RPI (11)

1st round (3):

38 BYU over 18 NC State
44 JMU over 11 Gonzaga
35 Oregon State over 17 Middle Tennessee

The old Jerry Palm RPI-based game predictor would have called 39 Florida State over 42 Iowa State an upset because it was on the Cyclones’ home court, but this is a pure RPI analysis.

2nd round (4):

38 BYU over 19 Nebraska
25 DePaul over 4 Duke
29 Oklahoma State over 9 Purdue
15 LSU over 12 West Virginia

The Jerry Palm game predictor would have said that LSU would beat West Virginia.

Sweet 16 (2):

21 North Carolina over 6 South Carolina
20 Maryland over 5 Tennessee

Elite 8 (1):

20 Maryland over 8 Louisville

Championship game (1):

2 Connecticut over 1 Notre Dame

I will note that the final was not an upset from a seeding perspective because (not that this is a surprise to anyone) UConn had the overall #1 seed.

* * *

10 games were upsets both against the seeding and against the RPI. 4 games were upsets only against the seeding. 1 game was an upset only against the RPI.

Last year, there were 15 seeding upsets and 17 RPI upsets, so RPI did a bit better this year than last year in predicting results. In 2012, it was 7 upsets both from an RPI and seeding perspective.

Fun with RPI will return after the World Series. I’ll try to make it November instead of March this time.