March Madness: Why RPI Is Not a Great Measurement of a Team's Ability

Rob Goldberg@TheRobGoldbergFeatured ColumnistMarch 11, 2013

NEW ORLEANS, LA - APRIL 02:  Confetti is seen on the NCAA logo after the Kentucky Wildcats defeat the Kansas Jayhawks 67-59 in the National Championship Game of the 2012 NCAA Division I Men's Basketball Tournament at the Mercedes-Benz Superdome on April 2, 2012 in New Orleans, Louisiana.  (Photo by Ronald Martinez/Getty Images)
Ronald Martinez/Getty Images

In the weeks leading up to the NCAA tournament, college basketball fans hear all about RPI but know very little about it. The truth is that it is an overused and outdated ranking system that needs to be abolished.

The Ratings Percentage Index (RPI) is one factor that the selection committee uses to decide not only who should be included in the Field of 68, but where each team should be seeded as well.

While each member can choose to use whatever information he or she feels works, the NCAA admits, "The RPI will be more helpful to a committee member when evaluating a team out of his or her own region, especially if they never have seen them play."

Obviously, there is no excuse for someone making a decision on a team they never saw play, but the question remains whether RPI should be used at all.

According to the NCAA website, the rankings are based on three factors: winning percentage (25 percent), opponent winning percentage (50 percent) and opponents' opponent winning percentage (25 percent).

The only wrinkle in the system is that road wins are weighted more than home wins, as it is much more difficult. However, this is an overly simplistic way of evaluating and just plain wrong in many cases. 

In the rankings, 75 percent of the score is based simply on strength of schedule. In some ways, this is valuable in making which teams are beaten an important factor. In the current rankings (via ESPN), Indiana (26-5) is ranked at No. 5, while Stephen F. Austin (22-3) is only No. 66. The Hoosiers certainly played a much tougher schedule to achieve their rankings.

Still, winning has to account for something. Nebraska ends the regular season at the No. 100 team in the country despite a 14-17 record. Sure, the Cornhuskers played one of the toughest schedules in the nation, but they finished with a 1-12 record against the Top 50 teams, and most games were not even close. 

If my school schedules all of its games against the best in the country and loses by 40, should it be highly rated? Not at all.

The fact that winning accounts for so little in the rankings is almost as bad as a lack of scoring margin in each game.

Middle Tennessee had a good season, posting a 19-1 record in the Sun Belt conference with a 28-5 overall record. Unfortunately, a closer look shows that an RPI of 28 might be a little too generous. In the games against the three toughest opponents on the schedule, the Red Raiders lost by five to Akron, 15 to Belmont and 21 to Florida.

Beating up on weaker opponents only tells us so much at this time of year, but a squad that cannot win against good teams should be much lower in any rankings system.

Last season, this issue played out in the tournament, with overseeded teams like Georgetown and Florida State being ranked higher due to a great strength of schedule. Both were given a three-seed and each lost in the second round.

Specifically, Florida State lost to a No. 6-seeded Cincinnati team that had a similar overall record and an identical 7-5 mark against the Top 50. However, the Bearcats were 30 places lower in the RPI because it played teams like Maryland-Eastern Shore and Mississippi Valley State.

Even though they won by more than 30 each time, it did not matter in this rankings. The on-court results show that this is flawed.

The good news is that there are other systems to use when evaluating programs. The Sagarin Rankings take into account a more advanced measurement for strength of schedule, and KenPom adjusts the scoring margin for the pace of a game.

Each of these systems has a more technical approach than RPI and in turn is a more accurate measurement of a team's season.

Additionally, ESPN recently created a system known as the College Basketball Power Index (BPI) that takes into account playing games without a key player. While this also has its faults (a win without a top player is worth less than with him), it gives a different dimension toward evaluating teams.

Still, the best approach for seeing if a team is good is to simply watch it play multiple games. Anyone who watched Indiana play Michigan on Saturday could tell that those were two very good programs.

Those Big Ten teams combine to only have one loss to a team outside the Top 50. Are they both better than No. 2 New Mexico, whose best non-conference win is Cincinnati and lost to Air Force and South Dakota State? Probably. 

The selection committees have a lot of information at their disposal over the next week. Between the computer systems, human polls and the "eye-test," there are plenty of ways to evaluate which teams are better.

Hopefully, they do not use the flawed RPI system that should be retired by the NCAA.

Follow me on Twitter for the latest college basketball information throughout the NCAA tournament.