... to get all Pirsig on you and ask you to define "Quality." But I'm not an asshole. Well, I AM an asshole, I'm just not being one now.
Note to poll voters: if you did not receive an email from me, please let me know. The first poll comes out Monday and you must be apprised of URLs and such.
SB Nation's excellent Missouri blog Rock M Nation will be joining the BlogPoll this fall, and they've thrown out a question to their readers: how the hell should we put together our ballots? This shows seriousness, which is an admirable quality in a voter, but a lack of deference to the poll's President For Life, which is neither admirable nor uncommon.
I've learned over time that I can't tell people what philosophy they should follow when compiling a top 25 poll. Or, rather, I've learned I can tell people what philosophy to follow and they'll just do what they want to anyway. There's only so much control you can pretend to have when the most respected college football blogger around thumbs his nose at some of the poll's published guidelines and the funniest one slaps up haphazard ballots 30 minutes after the deadline, usually after IMing me something like "oh crap give me a few minutes."
So vote how you like, with one exception. This is the exception: ballots designed to call attention to themselves are verboten. The lone spiked ballot in poll history came from Notre Dame uber-blog Blue Gray Sky after the first week of the season. Because I am stupid I deleted it, but by BGS's own admission it was designed to highlight how silly releasing a college football poll after one week of play is. This is a perfectly fine argument to make, and one I might even agree with, but your ballot is not the place to make it. Some voters tend to call attention to their ballots by their voting patterns, whether it's Straight Bangin's sadly prescient Michigan pessimism or SMQ's resume-only first week ballot or Double Extra Point's uncanny ability to have the most boring ballot; these are okay because their notability is a side effect of the voter's habits, not the entire point.
Other than that, feel free to be stupid -- because you will be stupid, iron law of polling, that -- in whatever way you want to. But I do think a unified philosophy benefits polling. SMQ highlights how goofy this polling enterprise can be:
But no one involved with any of the mainstream polls, despite their all-too-frequent use of the term, has ever defined exactly what they mean by the concept of the best team, or how they reach that judgment in comparison with that team's peers. Most of the time, the terms are described in an abstract way, as a mental sum of perceived parts, as if there existed a secret rating system, EA Sports-style, that could settle the issue once and for all.
The BlogPoll's concept of the best team in a sentence: the BlogPoll attempts to rank teams in order of season quality. This is impossible to do before the season and silly to do in the first few weeks, and at these times the poll should be regarded as an approximate guess of which teams will end the year with the highest season quality.
Suggestions to effect this ideal follow.
Once you have enough information, vote by resume only. What qualifies as "enough information" will vary from voter to voter, but I'm sure most will agree once teams are eight or so games into their schedules there's plenty of evidence to go on. Personally, by week five I try to excise everything except results. At that point there's no reason to look at future schedules, no reason to look at preaseason expectations or shiny offensive baubles. Just the facts, m'am.
When you don't have enough information, vote by your guess at team strength, not schedule. In an ideal world everyone would play an identically difficult schedule and this wouldn't be an issue. This is far from an ideal world, and some team just have nummy soft schedules. This is often cited as a reason to rank them high -- SMQ explicitly calls it out as a factor in his preseason ballot -- and drives me crazy.
Place great importance on schedule strength. The poll's greatest development in three years of existence was its continued, extreme skepticism of a Hawaii team that barely eked out victories against poor WAC teams and found itself in the top ten of most major polls and in the BCS against Georgia. That ended with Warrior limbs flung across most of New Orleans and everyone hurredly pretending like that never happened. You should take schedules into account more than it seems the other polls do, IMO.
Style counts. This is really tricky. If a team has three fluke plays go against them and loses a game it statistically dominated, what do you do? Dan Steinberg's pet Vegas Top 25 virtually ignores fluky results and thus can claim to be a better predictive device for upcoming games. The BlogPoll aims to be descriptive, not predictive.
The sad reality of college football these days is that schedules are so watered down and multiple teams will have the same records or nearly identical records at the end of the year but they'll have taken different routes to get there. So, yeah, team A had a better season if it crushed all comers and were under serious threat only a few times while team B squeezed by by the skin of its teeth, assuming schedules are approximately constant.
Back to SMQ for a pithy summary:
That is, assumptions about "the best" are frequently proven wrong by actual events. The best system, then, is not a rigid assessment of perceived strength, but an extremely fluid, strictly achievement-based approach that systematically rejects assumptions and accounts for chaos -- the inevitable black swan -- as the natural order. If South Florida's resumé is the second-best in the country in late October, then yes, it's the second-best team at that point. But probably not for long.
Co-sign. Man the ballot stations.