Teams with Surprising RPI’s

by David Mihm | January 29th, 2007

It’s about this time every year that the RPI formula, and its impact on selecting and seeding teams for the NCAA Tournament, comes under heavy fire from media, fans, and even coaches.  On the media side this week, both Jay Bilas and Seth Davis voiced strong opinions about the RPI, with Bilas re-iterating his desire to see the RPI dis-used by the Selection Committee, and Davis commenting that a team’s RPI rating is “absolutely meaningless in determining its NCAA Tournament standing.”

The RPI formula has its flaws, and in some cases it has been over-used by the Selection Committee, but to say that it is meaningless is just flat wrong.  Appendix E of the Men’s Division I Basketball Handbook reveals that the RPI influences seven of the 13 criteria used by the Committee to select and seed Tournament teams.  And until that changes, the debate will continue over the extent to which the Committee should rely on RPI data.

Thankfully, RPI is only one component of a team’s resume.  Plenty of other criteria, such as overall Division I record, last 10 games record, and road record, are used by the Committee in evaluating a team’s Tournament worthiness.

In most cases, a team’s overall profile parallels its power numbers.  But the exceptions are what makes seeding the bracket so interesting, as we try to gauge just how much the Committee will rely on computer data versus human judgment.  Let’s take a look at ten of the most surprising disconnects likely to affect the 2007 bracket.

RPI data from Ken Pomeroy (kenpom.com) through games of Sunday, January 28.

RPI = “Remarkably Positive Impact”

RPI 7, SOS 3

The Wildcats are clearly a Tournament-caliber team, but not even the most partisan fan in Big Blue Nation would argue there are only six teams in the country better than UK.  Tubby Smith’s team did pick up a couple of very nice non-conference victories, including a home W against Indiana and a win at Freedom Hall.  But they dropped all three of their truly marquee non-conference games (N-UCLA, N-Memphis, @UNC), and until Sunday’s home win over Tennessee hadn’t beaten anyone of note in SEC play.  Why is UK ranked so highly by the computers? A distinct lack of cupcakes: only three out of 19 opponents have come from outside the RPI’s top 150.

RPI 16, SOS 8

Another Wildcats team is also getting plenty of love from the computers, despite a 3-4 Big East record (with a visit from Pitt looming this evening) and a 6-6 Top 100 RPI record.  Villanova just seems to be hitting its stride, and may very well be a top-20 team by the end of the season.  But until their Big East standing catches up to their overall profile, these ‘Cats are taking advantage of the bonus points awarded by the RPI formula for road/neutral games, where they sport a gaudy 8-4 record, including W’s at Georgetown and Providence.
RPI 27, SOS 28

I praised Gary Williams’ team in this column a few weeks back for its strong chemistry and solid non-conference resume.  Prior to ACC play, these numbers would have sounded just about right for the Terps.  But Maryland is now just 2-4 in conference, including a home loss to Miami.  I sense the Terrapins will turn things around, but even if they only go .500 the rest of the way, numbers like these might be just enough to get them in the Tournament.
Florida State
RPI 29, SOS 23

Florida State has played one of the tougher ACC schedules to date, and is likely in better shape than Maryland, game-in-hand notwithstanding, heading into their big tilt on Wednesday night.  Though FSU is just 3-4 in conference, its power numbers are dramatically better than they were last year, and they’re probably “in” the field of 65 at this point.  FSU would do itself wonders, however, by winning even one game next week (@Duke, @Clemson).

RPI 36, SOS 42BYU will be an interesting case for the Committee and for bracketographers everywhere on Selection Sunday.  Given last year’s acceptance of borderline Western teams to balance the field geographically (see: Air Force, Utah State, and Cal’s #7 seed), Brigham Young would seem to be in good shape with these power numbers.  But on closer inspection, their profile reveals no good wins until Air Force this past weekend (which, unlike last year, is a true quality win).  Bad losses at Boise State and Lamar mar the Cougars’ non-conference resume, and a Mountain West regular-season finish ahead of UNLV would be advised.

RPI = “Really Puzzling Integer”

51 RPI, 32 SOS

Virginia’s numbers are finally starting to catch up with its performance. Prior to Sunday’s comeback win at Clemson, UVA stood at 73 / 49 – these numbers for a team which had defeated Maryland, Arizona, and Gonzaga and had compiled a 4-2 ACC record.  Even after one of the best road wins the country has seen this year, 51 still seems surprisingly low.  Cavalier fans can only hope the Committee takes a look at the actual teams Virginia has beaten, rather than its bloated RPI, should the Wahoos stumble down the stretch.
RPI 57, SOS 59

The team that notched arguably the best road win of the year (@Oregon), and one which stands just a game back of the lead in the country’s best conference, is placed at a clear disadvantage by its power numbers.  The Trojans beat Long Beach State, George Washington, and Wichita State in the non-conference season, and could really use some help from those three teams down the stretch.  I’d guess the Trojans will end up in the 8-9 game at this point, but a dip of ten RPI points or so might be enough to knock them out of Tournament consideration, should they falter at all in Pac-10 play.

RPI 58, SOS 60

The hottest team in the country (or co-hottest team with Stanford) isn’t exactly blazing its way up the computer rankings.  Despite a number of quality wins in the past two weeks, including at Kentucky and at LSU, Vandy still finds itself squarely in the middle of bubble territory.  Even a non-conference W over Georgia Tech isn’t really helping the Commodores, as the Yellow Jackets are having troubles of their own.  The remaining SEC schedule is pretty favorable, except for two tilts against the defending national champions, the first of which occurs on Wednesday in Gainesville.  A win in either one would assuage any remaining fear of missing the Tournament.

RPI 50, SOS 99

Here’s a team I just do not know what to make of.  Kevin Durant is only some of the talent leading the way for Rick Barnes, but the Longhorns are just 4-5 against the RPI Top 100, and struggled with both Nebraska and Baylor this week.  They’re 5-1 in the Big XII, and not in any serious danger of missing the Tournament, but one has to wonder how high this team’s seed will rise, given that its toughest stretch of conference games is just beginning.  #1 seeds around the country will be rooting hard not to see Texas’ name drawn as an 8 or a 9.

RPI 155, SOS 21

The 7-13 Utes are well under .500 and have no chance to make the NCAA Tournament as an at-large team.  But Ray Giacoletti’s squad has some pretty amazing wins on its resume: Air Force and Washington State at home and Virginia in San Juan.  BYU had better watch out on Wednesday in the first of two Beehive-state showdowns.  The Committee might be forced to ignore a sub-150 loss if Utah’s RPI holds steady the rest of the way.

Leave a Reply

Latest Headlines

Browse By Category

Browse Archives By Author