RPI and SOS are subjective and shouldn't even be considered in selection.

[jamie72921]

This is really getting out of hand here.  Is the selection flawed?  Of course, but so will any selection system of 100+ teams.  I can tell you without a doubt, that RPI and SOS is OBJECTIVE.  Any numerical data that is calculated using a formula is objective.  Even if it is weighted, as long as it is weighted using a formula then it is objective. Any test, data, or information that is definitive in nature/criteria is objective.  Now an example of something subjective would be an opinion or something not confirmed by data.  An example of subjective data is a coaches poll, an example of objective data is the final electorial college voting count.

Is this data discrete or empirical.

My stat professor doesn't agree with your point of view here, because you cannot measure HOW good a team is in comparison to another. You can only say that one is better and one is worse. It is impossible to prove by how much, a team is better than another. This is what many on this board believe is going on with the RPI.

Formulas can be subjective, and are when data is unevenly weighted based on criteria other than the number of observations within a sample or population and the RPI, definitely fits that category.

[jamie72921]

You might consider doing a LITTLE research & learning how the RPI is calculated.

The RPI adjusts for home vs road games. For eaxample, IF your team was somehow able to play nothing but home games & had a 20-10 W-L record---THEN your W-L record for calculating your RPI would be 12-14.

Playing a ton of home games in the non-conference schedule is not the best way to boost your RPI.

If it doesn't work, as you say, why does North Carolina do it then? Seems like it would cost them seeding at least, and yet, it doesn't.

Your example of playing nothing but home games is ridiculous to the point of absurdity.

When in a BCS level conference, you can load up on home games in the non conference because you will play a minimum of 9 road/neutral games against teams that will likely raise your RPI even if you don't leave your home court. More than making up for a lack of non conference away games.

Name a program that plays away games even at a 50/50 rate in the pre season.

[Biggus Piggus]

1. Relating to, proceeding from, or taking place within an individual's mind, emotions, etc. 2. Originating from or influenced by one's personal interests, prejudices, emotions, etc. 3. Of the mind or emotions only. (As opposed to objective).

Dead horse, die already.

I think what "noledge" he's trying to show off is confused in semantics.  If one were to use RPI and SOS to calculate odds of the outcome of a particular game (proxy for which team is better), then the probability would be estimated as subjective or Bayesian probability.  Limited number of observations.  But he leaps from this to slur RPI/SOS as fatally flawed because they are subjective, which is at best misleading.  They are a limited data set, but they are not probabilistic math.  RPI/SOS are a method of ranking teams in sequential order according to wins/losses, opponents' wins/losses, opponents' opponents' wins/losses, with the subject team's wins/losses (not the others) weighted according to home/road/neutral.

The committee uses RPI/SOS as one input, one view on how to adjust W/L record for schedule strength.  It is obviously a rough guide.  Georgia Tech (52), Texas Tech (53) and Stanford (63) were not in the top 50 of the final RPI.  Air Force (30), Missouri State (36), Bradley (38), Drexel (39), Florida State (41), Utah State (43), Clemson (45), Alabama (49) and Syracuse (50) did not get in.