What would you say about a Coaches' Poll voter who looked over the college football landscape after Saturday's action and ranked Auburn , undefeated and owner of more top 25 wins than any team in the country, all the way down at No. 11? What would you say if he voted four different teams with two losses ahead of the Tigers, including a Missouri team coming off of a loss to Texas Tech and an Arizona team that lost at home to 4-4 Oregon State ? What would you say if he kept Alabama all the way up at No. 5, ahead of not only the Auburn team two games ahead of it in the SEC West standings but 12 spots ahead of the 8-1 LSU team that just beat it head-to-head?
What you would say is that this voter had lost his damn mind and deserved to have his voting privileges revoked. You would say he deserved no part in a BCS process where so, so much is riding on every ballot. And you would be right.
Then why do so many college football analysts, fans, and statisticians insist that the six computer rankings that enter into the BCS formula -- each of which carries far, far more weight than any single Coaches' Poll or Harris Poll ballot -- be allowed to use margin-of-victory as part of their calculations? Because the hypothetical ballot above is exactly what the computers would spit out; it's the current "Predictor" rankings as produced by ratings guru Jeff Sagarin , where margin-of-victory is all-important and straight wins and losses irrelevant. Sagarin has stated unequivocally that he would prefer submitting the "Predictor" rating to the BCS-mandated margin-of-victory-ignoring "ELO_Chess," for the reasons laid out here by fellow BCS computer rater Kenneth Massey and baseball statistical godfather Bill James :
“You’re asked to rank teams that don’t play each other, that don’t play long seasons, and you can’t include margin of victory?” said Massey, who provides a “better version” on his Web site, masseyratings.com . “It’s a very challenging problem from a data-analysis standpoint. It does require sacrificing a bit of accuracy. It’s not the best way to do it" ...Maybe the math is nonsense. But shouldn't that be weighed against the fact that to virtually everyone else who follows college football, ranking Alabama ahead of LSU is an act of even greater nonsense?
“This isn’t a sincere effort to use math to find the answer at all" [according to James.] It’s clearly an effort to use math as a cover for whatever you want to do. I don’t even know if the people who set up the system are aware of that.
“It’s just nonsense math.”
The problem is that ratings system like the Sagarin "Predictor" and Massey's preferred system (which also ranks the Tide over the Bayou Bengals) aren't even trying to do the same things the BCS rankings are attempting to do. Their goal is to identify which teams are the "best," the most powerful, the most likely to win a given matchup; as its name implies, what "Predictor" wants to do is forecast the future, and there's no doubt it would do a better job of this than "ELO_Chess."
But certain unfortunate tiebreaks (like TCU 's and Boise State 's current predicament) excepted, BCS berths aren't awarded on the basis of hypothetical future results, or guesses at perceived strengths. They're awarded on the basis of achievement, on wins and losses and conference championships. Including margin-of-victory may make the BCS computer rankings "more accurate" when it comes to selecting which teams are playing the best football, but it would make them less accurate when it comes to answering the question the BCS rankings are trying to answer: which teams are most deserving .
That ought to be cause enough to keep the rankings margin-of-victory-free, even before we start wondering whether we really want the BCS nodding in approval as Boise desperately tries to hit the century mark week-in and week-out on the San Jose State s and Wyoming s of the world. (Not to mention it's already a shame when a player injures himself in a game that's well in hand; what happens when LaMichael James or Justin Blackmon tears an ACL trying to tack on a computer-mandated score at the end of a 60-7 blowout?) No, it's not particularly fair for TCU's annihilation of Utah to go in the BCS computers' books as nothing more than a W. But as the Horned Frogs' jump up the human polls shows, it's simply not true to say the BCS doesn't take the impressiveness of their victory into account at all.
The bottom line is that by including scoring margin (even one capped at, say, three touchdowns) in their computer rankings, the BCS would officially declare every win numerically judged like a figure skating routine, would give the thumbs-up to coaches like Bob Stoops who'd prefer to quit on a potential win over risking an embarrassing loss, would agree with "Predictor" that Alabama beating Duke by 49 points is more important than LSU beating Alabama by 3. The computer rankings could be better, but the way forward isn't to open a Pandora's box that college football would be much the worse for having opened.