Bracketology Grades

Considering how closely people follow the projected brackets, I thought it'd be fun to see who ended up being the closest. I took the final predictions from the three big sites- ESPN (Lunardi), Sportsline (Parrish), SI (Mandel). Here's how they stacked up.

Lunardi was the only one of the three to guess the field perfectly, going 65/65. The other two each missed one. Parrish had Illinois St. in, and Oregon out; Mandel had Virginia Tech in, and Villanova out.

For me, the longer you looked at Illinois St.'s resume, the more obvious it became that they were not deserving. Their RPI is 33, which is quite good for a bubble team. They went 13-5 in MVC, a conference that has made a name for itself over the last few years, and reached the conference tournament final. They were also 2-5 against the RPI Top 50.

But a closer look reveals that there's really not much there. There was one very good team in the Valley this year- Drake. Illinois St. played them three times, and lost all three. Their two Top 50 wins were both against Creighton, 46th in the RPI and not exactly an elite team this season. They also have one terrible loss (Eastern Michigan, RPI #237), and a couple other bad ones (Northern Iowa, RPI #131, and Indiana St., RPI #128). At the end of the day, they didn't have any impressive wins, and had some bad losses, which I think canceled out their strong RPI.

Virginia Tech was pretty clearly undeserving, in my opinion. People were arguing for them because they almost beat UNC. This is not an acceptable argument. Oregon only lost to UCLA by 5; is that at the top of their resume? Of course not, because they have real wins. This is not the case for Virginia Tech, whose best win was against a pretty mediocre Miami squad. They were 1-7 against the Top 50, and their second best win was probably @Maryland. Not very impressive.

These guys don't only predict what teams will be in, of course- they also guess what seed each team will be. So, who was the best at that?

Somewhat surprisingly, it was SI's Mandel, whose average miss was only 0.63. Parrish was second, at 0.72, and Lunardi came in third at 0.77. If we use RMSE, to penalize more for big misses, Lunardi jumps ahead of Parrish, but Mandel still leads.

The most surprising seed was Oregon being a 9. Lunardi and Mandel both had them as a 12, and Parrish didn't even have them in the field. They'll face Mississippi St. in the first round.

Other surprising seeds: Kansas St. (11, expected to be an 8/9), Indiana (8, expected 5/6), Miami FL (7, expected 9/10), and Butler (7, expected 4/5). I was shocked by Indiana, and also Vanderbilt, who somehow got a 4 seed.

Related: Selection Sunday 2008 Part II [The Money Line Journal]