Tuesday, December 7, 2010

You Can Fix Computers, But Not Humans

Boise State jumped LSU in the final BCS standings a day after the release. The Broncos must've had a real spirited practice. Or was it an excellent team banquet?

Thanks to an error discovered by Jerry Palm, the flip-flop in the Colley Matrix computer rankings between Boise State and LSU resulted in the Broncos moving up to No. 10 in the BCS standings, one spot ahead of LSU. The same flip-flop between Alabama and Nebraska did not change any placements, though it did bump up the Huskers' ratings score a bit (but not Alabama's).

But just what does this all mean? The truth, not much.

First, let's give Palm, the Guru's competitor, credit for this discovery. And also Wes Colley, who's the only one of the six computer operators who makes his computer formula public and therefore it's verifiable. The acknowledgment of the error speaks well for Colley's integrity.

In the big picture, though, as much as this might come as shocking to some, it matters almost not at all. The reason? The BCS is rigged anyway. (Another shocking news!)

Alarmists like to point out what if the flip-flop didn't happen at Nos. 10-11, but at the critical Nos. 4-5 or even Nos. 2-3? For one, it would never happen at Nos. 2-3 as the system is so tilted toward the voters that they have the ultimate call on who gets to play for the BCS title anyway. The computer rankings, as I wrote previously, are completely irrelevant.

The brouhaha actually masks the real problem. Computer glitches happen and they're easily correctable. Human biases, however, are not. How do you justify, for example, that the Broncos are No. 16 on New Mexico coach Mike Locksley's ballot, six spots behind Virginia Tech (a team Boise beat and has two losses) and also behind three-loss Texas A&M?

This year's coaches' votes (and Harris, too, for that matter) are relatively controversy free. But that hardly means that "the system worked." The BCS's biggest problems have always been both a lack of transparency and an unacceptable conflict of interest. Until those are fixed, the computer problems are laughably insignificant.

You can have all six computer ratings audited and verified weekly as some have suggested (and I'm all for that, too) and that will not do that much to safeguard the standings. As long as the coaches and Harris voters get to keep their ballots secret for most of the season, and not throw any outliers out in the polls (as it's done with the computers), the voters can make a whole lot more difference than any of the computers.

And let's face it, as long as they place one-third of the standings in the hands of the coaches, who have a vested interest in all of this - ranging from postseason bonuses and job security to friendship and petty jealousies - the system will remain fatally flawed.

Let's also bring up another red herring, the much hyped automatic qualifying issue. Again, the exact component of the AQ formula and the "evaluation process" are just smoke screens. If the BCS wants the Big East in and Mountain West out, the "formula" will do exactly just that. TCU finally got the memo and hopped on the Big East wagon because it realized that unless the Dallas Cowboys decided to join the MWC, it's never going to be an AQ conference.

The bottom line should be crystal clear: The BCS is rigged and therefore can never be fixed.

4 comments:

Eclectic Chad said...

So if the two teams can be swapped based on the result of a 1-AA playoff game, how would the BCS standings be affected by another 1-A game, like the Army/Navy game this weekend? Won't that have an affect on all the computer rankings? And if so, how can the current BCS results be the "Final" results?

What a stupid system.

tyamdaly said...

I agree with chademe. The Army/Navy game will do something to the computer rankings. Probably not much but it could mean a place or two. So really it is not final yet. Notre Dame is connected to both of these teams directly and this game is indirectly connected to the rest of the conferences. Army/Navy are both bowl bound this year and are not the doormats they have been in past years. So the BCS automatically throws out one of the biggest "tradition" games of the year? It won't change the top 8-9 positions. But it could have some effect on the 10-20 positions.

Anonymous said...

I assume you are stating the computers don't matter now. But was it 2003 season when USC was #1 in both human polls but somehow the computers pushed LSU and Oklahoma ahead? I think they changed the formula as a result, but I know most fans thought that was ridiculous.

reamon said...

Nice post. You're about the only person I've ever seen to point out the conflict of interest with the coaches' poll. Even figure skating doesn't let coaches vote on who gets the gold!

Google