RPI...myth v fact

[courtside]

Andy Katz also has apiece up on ESPN.com as well as Vahe's.

http://www.stltoday.com/stltoday/sports/st...A7?OpenDocument

[kwyjibo]

For the record I always try to temper my talk of the RPI with the fact that it is only part of the selection process. The most recent formal statements by the committee seem to indicate even more strongly that the RPI takes a reduced role. The selection committee process is such that in order to place a team on the table for selection you must have an objective argument in place though and in that context RPI is used as a starting point for many discussions. It is not the end of discussion.

There are wider issues that almost always get missed in Katz/Gregorian/Bilas articles. What SHOULD the role of some type of objective index be? I think it should be more of a role but I have never personally thought the RPI was a very good index. People here completely forget the history of the RPI when they argue about it (i. e., it was invented in a pre-internet age, 1981, to be calculated ONCE at the end of the season to rank teams and make sure big schools with a lot of losses get in and teams near deserts with great records could be left out). It is with irony that any complaints from BCS apologists have about the MVC/Colonial/Horizon looking artificially good should be heard ("hoisted on their own petards" as it were).

So now that the RPI occasionally overrates MVC and Colonial teams as well as doing its intended job of making BCS schools look good (I actually think it does that job very well as the RPI tends to overrate BCS schools far more than it overrates the MVC, Colonial, and A10), the BCS media-which is most of ESPN and BCS beat reporters--complains that something needs to be done. They want to make sure that the selection process lets in big conference teams that never left home in 2006 and never played anybody tough solely because they were able to beat a few good teams in conference play (an option that does not apply even when smaller conferences have a bunch of good teams). So, the committee, chaired for the first time in a while by a non-BCS school but always at least half BCS, makes it absolutely clear that non-objective factors will be used as much as possible to give the BCS schools as many bites at the apple as they need.

I think they do an OK (in comparison to TV analysts for example they at least only allow 65 teams in) job but when they select teams like Seton Hall and not Missouri State or Hofstra last year something is seriously wrong. If nothing else, the last few games of season seem to be of far higher import than they should be. Anyway, there are better measures to use than the RPI (the best measure both predictively and for fairness is simply the measure that minimizes out of all possible rankings of teams the number rankings a team that beat another is placed below the team it beat--otherwise known as the ELO method which Sagarin publishes).

I hope you read the first myth of the article you link to--the one about the myth of conference quotas.

I think this is very interesting reading about the actual process in detail:

http://www.ncaasports.com/basketball/mens/story/6985142