There are few things I want to burn with fire more than the RPI. The measure is an outdated performance tool that maintains numerous flaws, none greater than the fact that the RPI fails to adequately determine the performance of the teams it considers. That's the biggest issue with the RPI: It intends to determine the relative strength of teams, but doesn't actually consider actual performance on the field, and yet it used to determine the relative strength of teams. It's 2014: How is the RPI still a thing that reasonable people use in unreasonable ways?
The formula that creates the RPI is the stuff of forgotten math: 25% of a team's record plus 50% of the record of that team's opponents plus 25% of the record of that team's opponents' opponents' record. That's dumb. Margin of victory -- the best indicator of relative strength -- is totally ignored and, most importantly, none of those factors actually measure what's happening on the field. It's purely results driven ("W's" and "L's"), and weighted more toward what a team's opponents (and the opponents of those opponents) are doing. In other words, who a team plays is more important than (a) whether that team wins or loses, and (b) how well that team actually plays.
This is a real thing that exists and continues to exist despite the fact that people have developed more sophisticated methods to quantify strength and ability. This wouldn't necessarily be a problem if the NCAA didn't platform the RPI as a major tool in the decision-making process for NCAA Tournament selection, but the NCAA has done just that: The RPI drives the process for considering the pool of teams that will populate the at-large field in the bracket and the seeding of teams in that bracket. Of the primary criteria for selecting at-large teams, every single factor -- save for location of contests, head-to-head results, and results against common opponents -- is based -- in some form -- on the RPI:
- Strength of schedule is based exclusively on the RPI ranking of opponents.
- Average RPI wins and losses are based exclusively on the RPI.
- Record against cohorts of opponents -- one through five, six through 10, etc. -- is based exclusively on the RPI.
- Significant wins and losses are based on whether a team is ranked higher or lower in the RPI.
This is so incredibly, impossibly dumb. Dictating the fortunes of teams through something that is universally considered out of date and inferior to other methods utilized to determine a team's strength relative to the rest of the country deserves only the highest form of sarcastic slow clapping.