Friday, February 9, 2007

Digger Phelps Doesn't Understand Your Mathematical Formulas

It's starting to get towards that time of the year again. That's right, within only a few weeks good ole Digger Phelps will whip out yet another matching tie-lighter combo and attempt to rationalize that the RPI is flawed because it is inflating the rankings of teams from mid-major conferences yet again. Get excited for this absolute clinic in drinking the big conference kool-aid. Every Missouri Valley team in the RPI top 150?!?!? (Something no other conference outside the SEC can brag about, by the way). For shame! Those wacky computers have done it again! Hey Digger, you are slowly morphing into basketball's Joe Morgan. Let me explain...

First, let's pause for a few things. The RPI is merely one tool used by the Tournament Selection Committee used to evaluate teams. It's not the end-all-be-all factor in determining who is in and who is out for a chance at winning a national championship (Suck on it, college football!). In 1999, New Mexico went dancing with an RPI of 74, yet Missouri State's bubble burst last year with an RPI of 21. It's more of a predictor of who will not make it, if anything. There are numerous other things considered - home/road splits, record in the last 10 games, strength of schedule, win/loss record, etc. - hell, even injuries! - that get you into the field and determine your seeding. So, first and foremost, let's calm down about the RPI. It's one factor.

Second, it's a formula (using "25% your winning percentage plus 50% opponents' winning percentage plus 25% opponents' opponents' winning percentage" as its basis, adjusted for home/away, for those of you wondering), so it's objective. It can't cheat, it doesn't know who teams are, it simply plugs the math. It's only when twisted in the hands of humans and analysts that it could become subjective.

So, back to Digger's (and numerous other talking-heads) point(s) - the RPI has favored the mid-major teams last year (as I expect the argument to be this year as well) due to the fact that these teams get "overly inflated" RPIs before conference play begins, then play each other round-robin within league play, thus keeping their RPIs stable and still relatively "inflated". Last year, some analysts argued that the Missouri Valley and Colonial Athletic Conferences were rewarded with more bids, four and two (thanks, JD) respectively, from previous years that were in line with their increasing RPI ratings from previous years. Fair point. So how do these numbers get inflated to begin with? The answer to this question is two-fold. One, they all play difficult schedules in the pre-conference season (What a brilliant concept!). Two, the RPI system rewards them for playing and winning road heavy schedules and has since 2004 when home/away games became weighted and, in order to get games with power schools, middies have to play on the road.

Are these mid-major schools working the current RPI system? Absolutely. Could power conference schools do the exact same thing? Absolutely... and then some. You know how many mid (and low) majors would kill for an opportunity to get a few power conference teams on their home floors in front of their fans? The problem is power conference teams (this includes your beloved Notre Dame, Digger) are driven by money (increased revenue of home games) and the fear of losing to teams they "shouldn't lose to." Plain and simple. So, yes, the RPI is working very much so in favor of the mid-majors who smartly schedule their games, but BCS-level schools have everything on their side to do the exact same thing and many of them decline to take advantage.

My major point is this - how is the RPI incorrect when you have a strong mid-major conference schedule helping keep your personal RPI high but working perfectly when you are a major conference player who uses their tough conference schedule to boost their RPI? It's a double-edged sword and completely contradictory. Look at programs like UConn and Syracuse. Every year they feast off inferior opponents lined up on their home court through December and generally have mediocre RPIs heading into conference play, where their RPIs climb to the top due to the strength of the Big East.

For instance, let's take a look at 2006 Syracuse, which racked up a 7-9 Big East record but won the Big East Tournament to earn a 5 seed in the NCAAs, quite possibly one of the most ridiculous seeds I've ever seen the Selection Committee hand out. This is a team that was 7-9 (!) in conference play with the 249th ranked non-conference schedule, yet they had an RPI of 18. This was merely a testament of the strength of the Big East and Syracuse's conference schedule, yet I saw nobody dropping their two cents on this. Instead, Digger said the RPI was flawed because all the Valley teams racked up good, but not great, resumes in non-conference and their conference play was just them feasting off the RPIs of their own league. Isn't this the exact same thing? Aren't both teams (Syracuse & whatever contending MVC team you choose) relying on the strength of their leagues to boost their own rank?

Perhaps the mid-major bitching will cease this year due to George Mason, but somehow I doubt it. Every year, the "expert" analysis never ceases to amaze me.


-- RK

4 comments:

Anonymous said...

Nice article, kasko, but I thought the CAA only got two bids last year (UNC-Wilmington and George Mason)

Anonymous said...

Good job, Kasko. Digger Phelps is a moron - he reminds me of Joe Morgan a bit.

I think the RPI is a bit flawed, though (not particularly in favor of "big school" or "little school"). You can play a bunch of shitty teams that play in and do well in very poor conferences, which would inflate your RPI big time. I haven't kept up with the Illini's opponents' results, but I know our RPI seems higher than it should be (although our finishing schedule should make it go down big time). Just trying to point out that there are other ways it could get "inflated." You did address the fact that the RPI is just one part of the puzzle, though.

Anonymous said...

^^^that's me, edog. i comment on a few other blogs under my real name.

Kasko said...

JD - You're right, I fucked up. It was 2 for the CAA with Hofstra getting snubbed. I'll edit that out when I get a chance tonight. Thanks for correcting me.

Edog - Very good point on the opponents and scheduling. I'd like to plug some numbers and see how it works. One key seems to be refraining from scheduling teams with 200+ RPIs and keeping your non-conf slates with teams in the 100s as your easy home games.