BCS Breakdown 2007: Week 5

Horn Brain checks back in with a weekly report on the BCS Standings, which he'll provide between now and decision day. --PB--

I've gotten such a great response from everyone about my first project, I decided to turn it into a weekly analysis to help decipher the enigma that is the Bowl Championship Series.  You can peruse the sheet I based all my calculations on here.  This will likely be shorter and sweeter in the following weeks once we get some of the basics down, but bear with me for now.

So let's start this off with a quick:

Billingsley Report Card
This section is actually where I compare all of the computers, but until Billingsley is removed from the poll, he's the only thing that I can help but notice.

Check out the numbers this week:

Note:  A&H, RB, CM, KM, JS and PW are Anderson & Hester, Richard Billingsley, The Colley Matrix, Ken Massey, Jeff Sagarin, and Peter Wolfe, respectively.

Billingsley gets no gold stars this week.  He got thrown out 13 times, or from 52% of the rankings, whichever sounds nicer.  His standard deviation was once again horrendous, almost 2.5 points above the average, and about 1.8 points above Peter Wolfe, who made the mistake of ranking Hawaii at #12, and otherwise would have been unremarkable.

Here's a list of the "cutest" of Billingsley's "suggestions" that he tried to pass off as a legitimate ballot:

Ohio State gets his #2 vote after losing to Illinois, probably because of the part of his poll that doesn't let teams move around in the rankings too much from week to week, which he says he put in to prevent teams moving around without any specific reason.  Well?  I guess teams not moving around no matter what just makes more sense to him.  Wild inconsistency #1!

USC is his #7.  YAaaaaaarGghrhhghg!

Texas is #21.  By itself, this would be fine, but with USC at #7 with basically the same resume, this is chalked up as Wild inconsistency #2!

Virginia is #24, and no one else has them below 16th.  He hates teams that weren't supposed to be good.

"Wait, Horn Brain, Boise State is his #12.  He's not biased at all!"  Wrong.  Boise ended up at #10 last year, (below the other polls.  Surprise.) and he carries over rankings from last season into this season as his "preseason" poll.  His treatment of Virginia and Hawaii, coupled with this ranking, gives us Wild inconsistency #3!

Wisconsin is #14.  WTF, Billingsley?!?!?

OK, that won't be as long next time.

Overall BCS Quick Check:
Here's where we'll look at all the polls compared to teh BCS average.  This will rarely be very long.

Interestingly, Billingsley is the worst poll compared with the overall BCS, as well.  This is a departure from the norm, which is discussed in my "Playing the Numbers Game" column.  The short of it is that RB generally agrees more with the humans than the computers.  This week, though, he doesn't really agree with anyone.  Hmmm.  Perhaps with the entire college football universe turning on its head repeatedly this season, the teams we all expected to succeed have not, and thus we have adapted our ballots accordingly.  Billingsley, however, lives in the Imagine-Nation, and his prevention on big slides and his preseason poll have combined to produce rankings that reflect what we all had thought would come to pass this season, slightly skewed by the actual events.  

Winner (lowest standard deviation) is Peter Wolfe this week, with A&H a close runner up.  I'm pretty sure PW got a big boost out of ranking Hawaii, while the other polls held back and were slightly punished for it.

That was nice and short, wasn't it?

Rankology:
This section will look at the standard deviation of a ranking, and see what we information we can extract from it.  Will be shorter once we get going.

Sooooo, here:

So, let's start with the outliers.  Texas, Hawaii and Boise State.  Doesn't feel good being in that kind of company, does it?  The computers hate all three, but Texas gets to #20 after beating TTU and OSU.  Now, at first I wasn't sure what to make of these three being grouped together.  Earlier this year, I noticed this and thought that a high standard deviation across the polls was a sure sign that you were getting a lot of hype.  Now, with Texas actually beginning to perform up to expectations, and the computers starting to come around, I'm not so sure.  I now think that these huge standard deviations can mean either humans buying into the hype, or the inability of the computers to account for the way a team wins and changes in team quality over the year.

Think about this:  People say that one problem with human polls is that it matters when a team wins/loses.  I agree, of course, that timing of a win should not seriously affect the ranking of one team compared to another team with a different resume strength, but what about teams with similar resumes?  Let's look at an example:

Suppose Ohio State had taken care of Illinois, and remained #1.  Now we have to decide between LSU and Oregon, two one-loss teams.  If LSU started the year looking shaky and lost its game early, then turned a corner and blew everyone else out.  Oregon, on the other hand, looked great to start the year, then starts going to the wire with teams it would have blown out previously, and ends up dropping one to Cal this week, and we have to rank them for the final poll tomorrow.  Assuming their schedules were similar, you've got to rank LSU #2, since they're the better team right now.  A computer, on the other hand, would just rank them nearly the same, with a tiny edge going to whichever team lucked out with an infinitesimally better schedule.  It's like a computer sees you as the average of how good you've been all year, while humans have a better idea of how good you are now.  It's a total judgment call whether or not you agree with me, but I'll have to say that if two teams are pretty close resume-wise, I'm going to go with the hot team.  Call me stupid, call me unfair, call me a BCS bowl-selection committee, call me whatever you want, but that's only logical in my mind.

Conclusions:
So what have we learned?  Well, perhaps some of our arguments against human pollsters are a bit overblown.  I'd say Texas falls into the hot-team category right now (certainly in the eyes of the BCS bowl people, at least), and I'm fairly certain that Hawaii is a mediocre team playing a terrible schedule, but I don't know what to say about Boise.  Are they getting hype from last year, or are the computers not able to see that they're smashing some of their conference brethren quite convincingly?  You discuss.

--Horn Brain--

X
Log In Sign Up

forgot?
Log In Sign Up

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior users will need to choose a permanent username, along with a new password.

Your username will be used to login to SB Nation going forward.

I already have a Vox Media account!

Verify Vox Media account

Please login to your Vox Media account. This account will be linked to your previously existing Eater account.

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior MT authors will need to choose a new username and password.

Your username will be used to login to SB Nation going forward.

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Burnt Orange Nation

You must be a member of Burnt Orange Nation to participate.

We have our own Community Guidelines at Burnt Orange Nation. You should read them.

Join Burnt Orange Nation

You must be a member of Burnt Orange Nation to participate.

We have our own Community Guidelines at Burnt Orange Nation. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9347_tracker