KING JAMES BIBLE: Human Football Polls Need Rebooting

As the returns from the latest college football Saturday come pouring in, it seems that Harvard has polled amazingly among machines aged one to five, but terribly among people aged 30-60.

The disparity between the opinions of the animate and inanimate is quite distinct and one worth illustrating, if for no other reason than to further discredit the merit of human polls.

Here is a quick sampling of the rankings given to the Harvard football team by the eight most widely cited computer polls:

Massey—6th

Matthews—3rd

Sagarin—9th

Laz Index—7th

Dwiggins—6th

Self—4th

Sauceda—3rd

Ashburn—1st

The average rank bestowed upon the Crimson by those eight polls is essentially fifth in all of I-AA (4.88). Let that sink in. Fifth.

Now call up your friends at Florida or USC and tell them you go to a football school. When they stop laughing, pass along the information presented above. When they continue laughing, hang up on them.

Now, there are a few of you out there who would make the argument (which usually includes the crackpot New York Times computer ranking as exhibit A) that computer polls are out of touch with reality and lack the subjectivity necessary to create an applicable ranking system. That reasoning can explain why one poll might place a team way too high or low, but in the face of eight different polls—all of which essentially agree—that contention seems to fall apart.

Then, there are the three human polls—ESPN/USA Today, Sports Network and AnyGivenSaturday.com. Harvard barely cracked the top 25 in all three, finishing 19th, 22nd and 19th in each of the aforementioned polls, respectively.

The Sports Network poll is by far the most ridiculous, as the Crimson (5-0) finds itself a spot behind Penn (4-1). The Quakers’ only loss came to Villanova, a team that Northeastern beat by four two weekends ago. And we all know what Harvard just did to Northeastern.

Here are a few more ridiculous tidbits that I’d be remiss if I didn’t mention. In each poll, there are at least four teams with two losses ahead of the Crimson. In The Sports Network rankings, Harvard finds itself behind one 3-3 team (Maine) and directly in front of two 4-3 teams (Villanova and Appalachian State). Strength of schedule is important, but it can rarely explain away the gap between a program which is undefeated and one which has three losses.

I’ve got some time, so let me point out a structural flaw which points to the meaningless nature of these polls.

The ESPN/USA Today poll is voted on by the coaches. Why has no one ever questioned this system? Coaches only play eleven opponents a year. On Saturdays they spend their time coaching, not watching three different games at the same time on three different televisions positioned strategically around a common room.

People knock the computers for merely taking in scores and spitting out a numerical ranking.

But what do you think the coaches do? Can you see a coach heading into his office after a tough loss merely to watch film of Southern Illinois and Georgia Southern in order to decide who should get his number one vote? Of course not. They’re looking at the same scores as the computers and trying to rationalize who should be ranked where from that information. Actually, that’s even a lie. They’re handing off their ballot to sports information directors who take the time to fill it out for them and send it in. Since the computers have a defined system for processing statistics, I’ll put my money on them, thank you very much.

If you need any more proof that it’s time to deemphasize the human polls, just look at the body that’s ramping up its support of them—the BCS. When you see that mess headed your direction, you know it’s probably best to abandon ship.

—Staff writer Michael R. James can be reached at mrjames@fas.harvard.edu. His column appears every Tuesday.