clock menu more-arrow no yes

Filed under:

BCS Breakdown 2008 V1.4

New, comments

Not a whole lot is shaking out in the polls this week, at least as far as my pet project is concerned.  Even our old friend Billingsley is merely tied for most throwouts this week (with Anderson and Hester, the poll that Huckleberry from BC says was created to prove Washington was better than Miami in 1991.).  I'd like to take this opportunity, then, to go into a little more detail than I normally do with RB's ballot, just to show you how systematically unfair the program is, even when it doesn't stick out like the sore thumb it is.  In light of the added emphasis on the psychopath who thinks that the Longhorn fans who hated him last year on the "message boards" love him now, we'll skip the poll analysis and dive straight into Billingsley Report Card.

(Don't worry, we'll get back to the poll analysis later if you're on the edge of your seat ready to adjust your picks for this week's pool.)


Billingsley (Extended) Report Card:

First, the usual numbers:


At first glance, Billingsley looks to have had a tame week.  He's tied for most throwouts with A&H, and instead of blowing away the field in terms of standard deviation, he actually appears to have chopped his crazy-score in better than half (Remember, usually he's the guy with such a high standard deviation that it's the only one above average in the group.).  But all this means is that without the crazy pseudo-random rankings thrown in as cover, we can see exactly why his poll is the antithesis of what the computer polls were envisioned to be when the BCS was introduced.

What do I mean?  Well look at his poll compared to the other polls and the computer average, highlighted below:


Look at the teams that Billingsley is high on (anomalously high based on being an obvious wack-job compared to computer average and just plain old looking at the numbers), in green.  USC, Ohio State, BYU and LSU.  Meanwhile, he has Utah, Ball State, and UNC mysteriously below the other computers.  Notice anything?  Like how the poll appears biased towards traditional powerhouses and against up-and-comers?

That would be a product of his inane idea that all teams should start the season ranked where they finished the previous year, coupled with a number of his other ridiculous rules like adjusting strength of schedule ratings based on venues played in (he actually factors in average attendance over the past few years in determining how tough it is to play at, say, Michigan's Big House vs. Oregon's Autzen Stadium.).  This sounds like it may be a valid metric, but then you realize that you're effectively going to make every team in a BCS conference seem like they're playing a lot tougher schedule than the non-BCS teams (moreso that actual), just because there are more fans of the bigger programs.  BIAS!  If you're expecting teams like Ball State and Utah to consistently perform at an undefeated/one-loss level for several years in order to garner the respect of your preseason poll, then you're effectively killing their chances to be ranked very highly.  BIAS!  If you give teams special bonus points for holding teams to low scores, then you're going to bias your rankings towards teams that play in conferences with crappy offenses (Ahem, Schlabach).  BIAS!  Billingsley's ranking system is an absolute joke.

Note -  I know that BYU isn't a traditional powerhouse, but they did finish ranked in the top 20 of Billingsley's rankings last year, and so that's where they started this year.

Outliers watch:

Boom, Mother Numbers!  Yeah!


Again, the strange group consensus continues, especially with such teams as Missouri (Does something about "Missouri" and "10-12" elicit a Pavlovian response from voters or something?  This is weird.  So is this.).  Again, I'm calling on all 5-10 of my readers to go ahead and speak up if you have any idea what conclusions I could draw from a really low standard deviation.  Until that glorious day, however, we'll have to just look a the the high weirdos.

Alabama -  Watch out for the teeth on those Gators.  I think it's pretty clear that the Tide will have to bring a much tougher game to the SEC championship bout with Florida than the one they brought to Death Valley.  Can they?  I doubt it.  Tebow isn't going to throw four INTs even if he plays 1 on 11 with the LSU defense.  Will Florida lose before they get there, though?  We'll have to hope.

Ball State - Here's a weird one.  The humans have them 15th and 16th, and the computers have them at 16th on average, yet their standard deviation is way high.  That's because the computers are totally divded on them.  Two have them at 12th, two have them at 16th, Billingsley hates their non-traditionally powerful guts at #23, and Massey thinks they belong at #19.  I'm pretty certain this is because they haven't played anyone but Pitt.  I wouldn't bet on them to cover a spread in a BCS game if they got to one.

BYU - Pretty clear case of humans likes, computers hates.  UCLA was (is) a horrible, horrible team.  TCU has now lost twice, and they whipped BYU's ass.  Get over it, people.

Pittsburgh - Computers haven't the faintest idea how to agree on them.  Some have them as high as 16th or so.  They were supposed to be half decent this year, and I think that's exactly what they are.  They got unlucky and thought that Ball State would let them have a win, but beyond that, not too bad.  Beat Iowa, beat Notre Dame, Navy, USF, but got a surprise F***stomping from Rutgers.  Wannstachio may yet live to fight another season.

Wake Forest - Who the hell cares about the ACC?

Good luck with those picks, gentlemen.  Oh, and don't forget to spread the word about Billingsley.  He remains the very worst part of the BCS, and that's saying a lot.

College Football BCS Rankings, Scores,
Schedule and Blog Posts - SB Nation - SB Nation