First, an admission: This doesn't really matter. In most seasons, teams will have more than enough opportunities to differentiate themselves, and first-, second-, and third-tier teams will end up approximately where they belong. Oklahoma State didn't necessarily finish behind Alabama in the polls last year because they were ranked lower in the preseason; they finished lower because more voters thought Alabama was a better team. Most of the time, this probably doesn't matter as much as we think it does.
But it could. And if there is a better way to go about a season's worth of polls and poll movement, why wouldn't we pursue it?
In 1936, the AP began releasing a Top 20 poll in college football. For grins, here was the first Top 10: 1. Minnesota. 2. Duke. 3. Army. 4. Northwestern. 5. Purdue. 6. USC. 7. Notre Dame. 8. Washington. 9. Pittsburgh. 10. Yale. 11. Duquesne. 12. St. Mary's (CA). 13. LSU. 14. Texas A&M. 15. Nebraska. 16. Fordham. 17. Holy Cross. 18. Tulane. 19. SMU. 20. Marquette.
(St. Mary's was totally the Boise State of its day.)
Bernie Bierman's Minnesota Golden Gophers would go on to finish first and claim the inaugural AP crown. Pittsburgh would win the second in 1937, then Davey O'Brien and TCU would win the third in 1938. Those first few years, the first poll was not released until three weeks into the season. A decade later, we had backtracked to two weeks. But in 1950, preseason polls began. In the "Some things never change" department, Notre Dame was your preseason No. 1; after blowout losses to Purdue (which finished 2-7) and Indiana (3-5-1), the Irish were out of the polls entirely four weeks into the season.
Presumably, the preseason polls were put in place for one primary reason: argument. It is a lovely debate topic to fill the summer months. I don't necessarily have a problem with this. I start to have problems, however, with just about everything else.
1. How are we determining our preseason rankings?
There are basically two approaches to filling out a preseason poll. A voter is either making his or her picks based on who he or she thinks is the best, or how he or she thinks teams will finish. The former is preferable, and for one simple reason: with the latter, you double dip.
If you pick a team 16th because you think they will lose two or three games to teams ranked higher than them, that's fine; but what inevitably happens is, when your No. 16 team loses to a team above them as you expected, you drop them in your rankings. You punish them because you think they will lose a certain game, and then you punish them again when they actually lose it.
If the goal is to get an accurate read for the best 25 teams in the country, then your arbitrary preseason predictions are muddying the waters quite a bit. If we must have a preseason poll, and if we must continue to track it week to week, then it should really be based on perceived strength, with no schedule issues taken into account.
2. Why do we automatically punish teams for losing, even to higher-ranked teams?
Boise St falls out of AP Poll for 1st time since Sept. 2008. It had been ranked for 62 straight wks, 2nd longest active streak behind Bama.— ESPN Stats & Info (@ESPNStatsInfo) September 4, 2012
Seemingly out of respect as much as anything else, AP voters selected Boise State 24th to start the season. The perennial top-10 team is replacing a nearly unprecedented number of starters, and not a single one of us had a clue what to expect of them heading into the season. In their opening game, at No. 13 Michigan State, they showed potential offensive issues but held steady enough defensively to hold a lead deep into the second half. They eventually fell to the Spartans, 17-13, but if nothing else a four-point road loss to the No. 13 team probably proved that the Broncos were of Top 25 caliber. But they were bumped from the polls because they lost. You almost always drop when you lose, no matter what.
My go-to example also applies here: in 2006, No. 2 Michigan lost to No. 1 Ohio State on the road by three points, proving with almost complete certainty that if you thought they were the second-best team in the land before the game, they should definitely be considered No. 2 after the game. But because they lost, they dropped behind other one-loss teams. Polls slowly shift from perceptions of strength early in the season to comparisons of resumes late in the year. But we basically treat them like ladder matches at a Thursday night tennis scramble. Win, and you move up. Lose, and you move down.
Oh yeah, and there's another issue at play in this week's polls.
Michigan dropped from 8th to 19th, yet another reminder that there's more disincentive than incentive to schedule tough.— Stewart Mandel (@slmandel) September 4, 2012
In the preseason AP poll, Michigan ranked eighth with 1,000 points, Oklahoma State was 19th with 430, and Boise State was 24th with 212. Michigan got whipped by Alabama, lost 554 points and fell to 19th. Boise State barely lost to a higher-ranked team on the road, lost 133 points, and fell out of the Top 25. Oklahoma State pantsed what was basically an NAIA-caliber Savannah State team ... and gained 128 points.
We ache longingly for better non-conference matchups. We complain about teams scheduling cupcakes. We demand better early-season viewing. But we give teams absolutely no motivation whatsoever for scheduling these games. In fact, we provide exactly the opposite motivation. If Boise State had beaten Central Michigan in Boise on Friday night, they would still be ranked. If Michigan had whipped UAB instead of playing Alabama on a neutral field, they probably would have moved up in the polls.
Now, in the end we might find out that Michigan indeed isn't a top 10-caliber team and that Boise State isn't Top 25-caliber. We are, after all, still mostly guessing right now. But Boise State did nothing that reflected poorly on them in barely losing to a top-15 team, and in getting destroyed by Alabama, Michigan simply proved that they are nowhere near the nation's top two to four teams. We had an inkling that the top four were leaps and bounds ahead of everybody else to start the season, and Michigan and Alabama proved that. But just about anybody would have probably been destroyed by Alabama with the way the Tide played in Arlington; we punished Michigan simply because they were in the building at the time.
There's an obvious solution to this, of course, which makes the entire process more maddening. And the AP got it right 75 years ago, for that matter. We could solve so much of this randomness issue by simply waiting to take a poll. Let's say that, in the coming weeks, Michigan destroys Air Force and UMass, then also destroys Notre Dame and Purdue on the road. Chances are, they will have proven themselves as a potential top-10 team, or something close to it. And if they're lucky, they'll make it back into the top 10, depending on who above them loses.
But if we were to take our first poll on or around Oct. 1, when we actually know a decent amount about these teams, then the Wolverines' fate won't be tied simply to whether No. 12 Clemson, or No. 14 Ohio State, or No. 17 Texas loses; it will be tied to who has looked better over the course of the first month of the season. The same goes for Boise State: beat a good-looking BYU team in two weeks, then win at Southern Miss on Oct. 6, and you'll have probably proven that you are not only a Top 25 team, but perhaps a top 15-20 one.
Again, preseason polls are great for conversation, for argument, for website hits, and for keeping college football in the opening segment of SportsCenter. That's fine. But even if we retain preseason polls, we don't need to have another poll until three or four weeks have passed. We'll have begun to learn what we need to learn about the sport's best teams, and we won't have simply moved teams up and down based on wins and losses in the process.
While we’re here, let’s watch some of the many fine college football videos from SB Nation’s YouTube channel: