The tool favored by the NCAA tournament selection committee is the RPI, which rates teams' accomplishments to date based on winning percentage and strength of schedule, giving extra weight to road or neutral-court wins but ignoring margin of victory. The tool favored by proponents of advanced stats is Ken Pomeroy's ratings, a predictive system which utilizes efficiency stats to provide a snapshot of every Division I team's current level of play.
With the NCAA releasing the first edition of its official RPI ratings on Monday, it seemed like a good time to examine the new ratings systems and see where they differ thus far. Here's a look at some of the teams whose RPI and KenPom rating vary the most and some thoughts on which is more accurate:
1. Colorado (10-4, RPI: 6, KenPom: 41): The RPI is overrating Colorado as a result of a tough early schedule that has already included five true road games and three games on a neutral floor. The Buffs deserve credit for beating Baylor, Colorado State and Murray State and nearly upsetting Arizona in Tucson were it not for a questionable game-ending call, but No. 6 seems way too generous for a team that got destroyed at Kansas and suffered losses to unproven Wyoming and Arizona State.
Which is right? KenPom is closer. I'd rate Colorado somewhere in the 25-40 range with the potential to crack the Top 25 later this winter if it can bounce back from an 0-2 start to league play. The Buffs lack depth and their shooting can be erratic, but Spencer Dinwiddie and Askia Booker form a potent backcourt, Josh Scott is already an effective low-post scorer as a freshman and Andre Roberson may be the Pac-12's best rebounder and most versatile defender.
2. Indiana (13-1, RPI: 16, KenPom 1): The KenPom ratings love the way Indiana has mercilessly destroyed the many bad teams on its schedule. The RPI punishes the Hoosiers' cupcake-heavy non-conference slate. Aside from quality wins against Georgetown and struggling North Carolina and a loss to Butler in Indianapolis, Indiana played nobody of note in non-league play and beat them all convincingly. What's more, the Hoosiers hadn't played a true road game before Big Ten play began.
Which is right: KenPom again is much closer. The Hoosiers haven't accomplished as much as Duke so far this season, but this is still a title threat. Assuming they're able to win on the road in the ultra-strong Big Ten, their RPI will climb into the top 10 as the season goes along and they'll have a chance to contend for a No. 1 seed. Yogi Ferrell has been poised and effective as a freshman point guard, Jordan Hulls remains a premier shooter and Victor Oladipo is an impact player on both ends. Plus, frontcourt play remains a strength even if Christian Watford remains erratic and Cody Zeller hasn't been quite as dominant as expected.
3. Kentucky (9-4, RPI: 62, KenPom: 9): Even though Kentucky will very likely be one of the top 25 teams in the nation by the end of the season, it's reasonable not to rank the Wildcats that high now because they haven't accomplished much. Their only notable win is against a Maryland team that hasn't lost since but also hasn't played anyone relevant. Nonetheless, the RPI has the Wildcats way too low because it doesn't account for margin of victory and it penalizes Kentucky for having yet to beat an RPI top 50 team.
Which is right: The RPI is way off and KenPom may eventually be on target. Kentucky's resume isn't the ninth-best in the nation at this time but the Wildcats could evolve into a top 10 caliber team by the end of this season. Kyle Wiltjer's outside shooting has been much better of late and Ryan Harrow's increased confidence and aggressiveness has solidified the point guard spot. Kentucky still needs to improve its free throw shooting and get more consistent production from Alex Poythress though.
4. Pittsburgh (12-3, RPI: 71, KenPom 11): Advanced stats love Pittsburgh, which began Big East play averaging the most points per possession in the nation and in the top 20 in points per possession surrendered. Granted the Panthers' best win came against Lehigh, but they destroyed the subpar competition on their schedule, which explains the massive gap between their RPI and KenPom rating. RPI penalizes Pittsburgh for its soft non-league schedule and doesn't take into account margin of victory.
Which is right: At this point, I'd have Pitt somewhere in the middle, maybe in the 35 to 50 range. I was of the belief Pittsburgh was underrated prior to their losses to Cincinnati and Rutgers this past week. I'm not ready to bury the Panthers after a pair of early league losses, but their stagnant offense, rebounding issues and and inability to keep speedy guards out of the lane last week was a concern.
5. Wisconsin (11-4, RPI: 90, KenPom: 17): Last year, Pomeroy had to pen an FAQ addressing why he rated Wisconsin in the top 10 much of the season when other rankings had the Badgers far lower. This year, his formula again has Wisconsin higher than anyone else because once again the Badgers' efficiency stats merit it. The Badgers have sustained losses to Marquette, Virginia, Florida and Creighton, but they've been so ruthlessly efficient against bad teams that they're top 30 nationally in points per possession and points per possession surrendered.
Which is right: Split the difference. A four-loss team whose best wins are against Cal and Arkansas probably doesn't belong in the top 20, but there also aren't 89 teams better than the Badgers either. Strong frontcourt play and interior defense will keep Wisconsin in contention for an NCAA bid, but they miss Jordan Taylor's ability to create for others and Josh Gasser's toughness and perimeter defense. Point guard play and the erratic shooting of Wisconsin's guards will continue to be a season-long issue.
Others of note: New Mexico (RPI: 11, KenPom: 51); Illinois (RPI: 12, KenPom 34); Butler (RPI: 18, KenPom 53); Oklahoma (RPI: 19, KenPom 56); VCU (RPI: 32, KenPom: 10); Notre Dame (RPI: 37, KenPom: 20); Ohio State (RPI: 41, KenPom: 12) Virginia (RPI: 133, KenPom: 27)
- Sports & Recreation