BlogPoll Week 11 Draft: Nothing Surprising Here

But Auburn has lived a charmed existence this year and still doesn’t have a win that told people that they were for real. For that matter, they still don’t have a road win that’s worth a damn. Okay, so winning in Starkville is a decent accomplishment, but I watched that game, and it was ugly. @Kentucky? @Ole Miss? Color me unimpressed. Alabama has been dropping off a bit, but if Auburn wins convincingly in Tuscaloosa, they have a strong argument for #1. If they win at all, they have an argument. But until then, I still don’t see why we should like them more than Oregon, TCU, and Boise State.

But I can’t fathom an argument right now that says that Auburn is better that either TCU or Boise. Could they win straight-up? Sure. Also note that Cal could very well have beaten Oregon last night. Variance is ever-present. Auburn, however, is a one-trick pony on offense and has just enough defense to scrape by. Both Boise and TCU are clearly superior teams on defense, and their offenses are just as prolific.

I understand resume, and with a computer system you’d think I care more about that. But I strongly feel that a national championship should be about who’s the best, not who has the shiniest schedule. That’s the same reason that playoff systems (NFL, NBA, etc.) don’t ever care about strength of schedule except deep within the tiebreakers. If there are two teams tied for the last wild car spot in the NFC, for example, and they didn’t play head to head, we don’t start going through and figuring out who had the best pelts until after they’ve exhausted conference and common game metrics.

Now, our computers are biased towards defense, which is why Auburn gets hit. We know this and I’m really hoping to change it. (Heck, I would have tweaked it midseason.) But that’s an issue independent of Auburn, and it only affects Auburn because Auburn’s defense is simply average – you have to go to Nevada in the rankings to find a worse defense, actually, and the others are not even close.

The SEC and the PAC-10 get their reward weekly for their schedules: they’re on national TV raking in far more money and getting far more promotion than the mid-majors. Conference play shouldn’t be discounted, but it’s not the all-important monolith in the room we make it out to be either. It’s one factor among several.

Perhaps a better perspective is Auburn/Ohio State. OSU spots Auburn 25 points off W/L and SOS alone in our computer, yet ends up 6 points ahead. Auburn’s offensive metrics are consistently better thanks to Newton and Tresselball, but OSU’s defense is just that much better that the numbers tilt in their favor. (Again, the defensive bias thing which, if fixed, would at least put Auburn ahead of OSU.)

With callers-in.

I’m with Will on this. Admittedly, I haven’t been paying all that much attention to anyone’s numbers other than Tennessee’s and those of its opponent that week. And I think that there’s a reason defensive metrics ought to matter more than offensive. But wins matter most of all, and if those wins come against tougher teams, well then that seals it for me. Of course the soft spot in that argument is the word “tougher.” I suspect the dirt is in strength of schedule, although I haven’t looked at it. What can y’all tell us about how we’re getting the data and weighting SOS?

Example: take two teams ranked, oh, 5 and 6. Both have one loss. Team A lost to Team B while Team B lost to the undefeated #1 team. All else is effectively equal. Beatpaths suggests that you rank Team A at #5 and B at #6, yet Team B actually has the more respectable loss.

This logic carries forward. Instead of 5 and 6, what if it were 20 and 21? Beatpaths give you the same result, yet Team B’s lone loss is far less damaging on its face.

I always get bogged down in the logic on this. I yield to Will on beatpaths in the poll, but there’s a sense in which it doesn’t really capture quality well. (It does capture head-to-head.)