College Basketball Bubble Watch

Updated: February 17, 2012, 8:38 AM ET
By Eamonn Brennan | ESPN.com


For better or worse, the RPI still plays a major role in the bubble world

Editor's note: This file has been updated to include all games through Thursday, Feb. 16.

On Thursday, the humble author of this here Watch let loose with a bit of a rant: The RPI is dumb! The NCAA keeps using it despite its dumbness! Loud noises!! Arghhh!!!!

These things are all true, but they're slightly more nuanced than that. When David Worlock, NCAA associate director of men's basketball and one of the staffers responsible for organizing and guiding the selection committee through its annual process, responded via Twitter -- he wrote: "And yet you reference the RPI in your own bubble watch column. We do use Sagarin, LRMC, KenPom, etc." -- I decided to strike up an email conversation. (It's always better than a public Twitter feud. The only thing more annoying than a Twitter feud is when people live-tweet the mock selection committee. But I digress.)

David shared a few points worth noting. To wit:

"I appreciate your points, and you are largely spot on," Worlock wrote. "The one thing the committee recognizes is that using the RPI solely as a basis for discussion for selecting and seeding teams is a flawed process. And that's why we don't. Your points about top 50 wins, SOS, etc. being based on the RPI is valid. But the committee does reference other computer rankings and it is noted when there are significant discrepancies between the RPI and other rankings.

"Bottom line is this: the committee has to pick the best 37 teams, regardless of which rankings are used. We could throw the RPI away today and the field wouldn't change dramatically, if at all, if we were doing the bracket tonight. Will there be arguments about VCU, UAB, Virginia Tech, Colorado, Alabama and the like? You bet. And that's OK."

That made me feel like a big meanie, because I did sort of call the NCAA (and by proxy David) luddites, and it's never nice to call names. My mom would be very disappointed.

Point is, the NCAA does at least have these things in mind. David even sent along the committee's sample composite rankings sheet, which includes the RPI ranks alongside KenPom, LRMC, Sagarin, and so forth. Encouraging email attachment, that. And at the end of the day, David's right: The NCAA does a good job selecting and seeding the field each season, and when you're dealing with large sets of numerical data spread across 70 or 80 teams, the immediate impact of each RPI number is certainly lessened. This isn't life-or-death stuff. I get that.

Unfortunately, as long as the RPI underpins all the nitty-gritty sheets the committee uses to compare teams at a glance -- as long as the main factors the NCAA considers (losses vs. sub-151 RPI, noncon SOS, top-50 wins, top-100 wins, and, you guessed it, RPI) are either the metric itself or based solely on it -- it's not exactly fair to say the RPI is just one more tool the committee considers. When the metric underpins everything the committee does, it isn't just one tool. It's the toolbox. As a fan of not merely information but good, accurate information, I think we can do better. Most fans seem to agree.

Anyway, rant over, at least for now. But in case you needed a reminder of exactly why the Watch criticizes the RPI in one breath and uses it repeatedly in the other, it's because the Watch, to be even remotely informative or predictive, must play by the same rules the selection committee uses. They are bad, outdated rules, but such is life, kids.

Without further ado: