The Cleveland Plain-Dealer’s Doug Lesmerises* has caused quite a bit of commotion early in the college football season with his unorthodox manner of voting. This week, Lesmerises’ top five teams in the country are Alabama, Boise State, and Houston, Florida, and Iowa. No other AP voter ranked the Gators lower. No other AP voter ranked any of the other four teams higher.
*Lesmerises isn’t alone here. Jon Wilner of the San Jose Mercury News pretty much does the same thing. It’s just that I read about Lesmerises first.
Lesmerises has laid out the rationale for his voting in painstaking detail each week. Instead of voting based on talent or coaching or preseason predictions, Lesmerises bases his ranking purely on results. Alabama’s win over Virginia Tech is the best in the country, and thus the Crimson Tide are No. 1, etc. His rankings alter dramatically from week to week because, well, the results alter dramatically. In his preseason rankings, Lesmerises had Oklahoma State ranked 14—one of the lowest rankings for the Cowboys. After their opening win over Georgia, he moved them all the way up to third, behind only Alabama and BYU. When OSU lost at home to Houston the next week, Lesmerises bumped the Cowboys down again to 22 and the Cougars from unranked to fourth. Needless to say, his ballot embraces flux.
It also has drawn a lot of attention, both positive and negative. Let’s focus on the negatives: There’s the basic argument that Lesmerises’ ballot is a poor reflection of how good teams really are. Most people make this by saying, “If Boise State were to play Florida on a neutral field on Saturday, who would you expect to win?”* Most people think the answer is Florida. So does Doug Lesmerises, who points out that his rankings are reactive rather than predictive.
*I cleaned it up a little. Most of the people making this argument in comments misspell Boise, don’t use the subjunctive mood, and generally include some vulgar ad hominem attack at Lesmerises at the end.
That leads to the philosophical question of the day: What should the role of rankings be? The institution of college football itself is confused over the issue, considering the AP tells its voters to be reactive while simultaneously putting out a preseason poll.
If we want our rankings to be predictive, they should rarely change, even after losses. Let’s throw out a hypothetical: I am an AP voter who thinks Florida is the best team in the country. But one week from Saturday, even with Tim Tebow returning from his concussion, the Gators lay an egg in Baton Rouge against LSU. Tebow throws an uncharacteristic two interceptions, Florida misses a key field goal, and the Tigers win by three. Does this necessarily make LSU a better team? No, I can still tenably expect Florida to beat the Tigers on a neutral field. And so I still rank the Gators ahead of LSU because I have that much faith in them.
The problem here is that the college football season is short, games rarely take place on a neutral field, and it’s hard to differentiate between aberrational bad games and ones that point to intrinsic flaws (i.e. if Tim Tebow’s throwing two interceptions and Florida’s missing a key field goal happen regularly). And if we all did our rankings predictively, the actual games lose most, if not all, of their significance.
However, reactive rankings cause problems, as well. The first has already been illustrated in Lesmerises’ ballot: continual flux. Dramatic movement by several teams in an individual voter’s ballot isn’t that big a deal; it becomes one if all the voters rank this way. Without any real consistency, rankings begin to lose their meaning. You only have to look back two college football seasons, when the top teams lost with such regularity that being No. 1 or No. 2 didn’t mean much, even at the end of the season (quick: Who was No. 1 and No. 2 at the end of that season?). If the top teams are always changing, then the impact of beating them is diminished. This is particularly troubling in a system, like that of Lesmerises, that acknowledges its temporality. For instance, let’s say his No. 3 Houston team loses tomorrow at UTEP—the same team Texas beat by 57 last week. Does Lesmerises give UTEP the same amount of credit for beating No. 3 Houston that he gave BYU for beating No. 3 Oklahoma earlier in the year? Or does he weigh them differently because he thinks, predictively, that OU would beat UH on a neutral field? And if it’s the latter, why rank at all, if the No. 3 designation means something different for different teams?*
*A more practical, less teleological argument on behalf of the fans: A large part of the fun (and perhaps the purpose) of rankings is to pander to fans and infuse games with some quantifiable meaning—a meaning that is lost when there’s the insidious notion that the rankings, if accurate at all, are only accurate for seven days. Instead of beating No. 3, underdogs will have to settle for beating The First Week of October’s No. 3.
Furthermore, reactive rankings rely exclusively on scheduling. Right now, the teams at the top of Lesmerises’ rankings are not only the teams that themselves have played tough games, but have played them against teams that have proven themselves in other tough games. To clarify: Alabama’s win over Virginia Tech is legitimated by Tech’s win over Miami which is legitimated by Miami’s win over Florida State who beat BYU who beat Oklahoma. This all just begs the question: Is Alabama’s win any less impressive if Sam Bradford doesn’t get hurt?
Scheduling becomes a bigger problem, though, in a sport with as little balance as college football. Critics of Lesmerises’ system point out that it’s wrong to put Houston in front of Florida just because the Cougars have played tougher teams to this point. These critics, however, will likely shy from making that same argument on behalf of Houston at the end of the season: Why is Florida ahead of Houston? Just because the Gators played an SEC schedule and not a C-USA one?
The biggest problem with college football is that its schedule offers minimal opportunities for perspective. The only major non-conference battles occur early in the season, when we’re still basing our thoughts on preseason expectations.* In fact, the only time we get a good sense of how conferences match up with one another is the bowl season, when it no longer really matters. The result is that we can’t objectively judge whether a 12-0 mark in the SEC is better than a 12-0 mark in the Big East.** The conferences have played their two games against each other already this season: Auburn and Kentucky won home games against West Virginia and Louisville, respectively.
*Wake Forest’s needing a last-second field goal to beat Ole Miss early last season looked like a bad win for the ACC. Turns out it may have been its best non-conference victory of the year.
**Most everyone would quickly answer with the SEC here. What’s more debatable is if an 11-1 SEC or Big XII record is more impressive than a 12-0 Big East mark—a debate that may determine who plays in the title game this year. And yeah, I’m that confident in Cincinnati.
Now I’m not going to argue that the Big East is better than the SEC; it’s probably not. It’s just that the argument for why the SEC is better than the Big East relies far too much on things like, “You serious?” and “Come on,” and “Open your eyes.” The only on-field proof for the SEC’s superiority in 2009 is a pair of contests between non-contenders in each conference. And what happens if West Virginia holds on in the fourth quarter at Auburn and Louisville scores a mild upset at Kentucky? Is that enough ground to say the Big East is better? There will likely come a time when the Big East is better than the SEC, and we probably wouldn’t realize it.
This is all a roundabout way of saying that, Lesmerises’ method included, there is no ideal way of ranking college football teams. There’s no ideal way of ranking teams in any sport really. College basketball has the same issues with rankings; it’s just that those rankings don’t mean quite as much. And if we can agree that there is no ideal method of ranking, well, can’t you see where I’m going here?*
*For those who can’t: WHY DO WE HAVE A SYSTEM THAT DECIDES A CHAMPION BASED ON FLAWED RANKINGS?
In the end, though, I do know one thing: Among AP voters, Doug Lesmerises is my No. 1.