clock menu more-arrow no yes mobile

Filed under:

Faulty Math

It's getting down to that time of year where people are going to start paying a lot of attention to the Pairwise Rankings, which are used by the NCAA tournament selection committee.So I figured I'd take this opportunity to point out what a stupid, arbitrary system it is.

In looking over the big grid, you'll notice that Massachusetts is currently winning their comparison over Alaska. That seems a bit strange given that Alaska is 9th in the RPI with an adjusted RPI of .5392 while UMass is 24th in the RPI with an adjusted RPI of .5185.

So how exactly is UMass winning that comparison? For starters, UMass beat RPI earlier this year. Alaska played RPI the following weekend and tied them. That is the only common opponent the two teams shared--UMass didn't play anyone from west of Niagara this year. Anyways, there you go. UMass is 1-0-0 against common opponents while Alaska is 0-0-1. Point UMass. That one game completely cancels out Alaska's significant advantage in the measurement of all 30-some games the two teams played.

Since the two never played head-to-head, that leaves record against Teams Under Consideration as the deciding factor. UMass' record against TUCs is 11-9-0 while Alaska's record is 6-5-6. Slight edge to UMass, which means they win the comparison. Of course, if the tournament committee decided to draw the completely arbitrary TUC line at the top 26 teams, that would introduce UMass' 1-2-0 record against Northeastern into that category, and Alaska would have a slight edge in the category and win the comparison. If they drew the line at 23 or fewer teams, getting closer to what the NCAA used to use when they really only looked at teams actually under consideration for the tournament, and this isn't an issue because UMass isn't even in the picture to begin with.

Obviously this is a fairly extreme example, but it highlights what is wrong with this system. Some people will say the Pairwise shouldn't be looked at until every conference tournament has finished, but it's at least plausible for a situation like this to occur in a couple weeks. UMass and Alaska won't play any more common opponents so that category is impossible to change. Alaska's fate--with respect to that individual comparison right now, but potentially also their NCAA tournament chances should they drop a couple spots--will likely be determined by which order teams 23-27 in the RPI end up in. That doesn't seem like the most logical way to determine who belongs in the NCAA tournament field.