Sunday, October 16, 2011

blogpoll ballot week 7

That was fun yesterday, no?  I'll babble some thoughts tomorrow on the game itself, but for the purposes of this post, the win yesterday caused me to do something that I don't think I've ever had the chance to do: move UVA into consideration for Blogpoll ranking this week.  The result:

The astute among you may notice UVA didn't make it.  I considered 35 teams this week; here are the ones that missed the top 25 cut, in order from top to bottom:

Georgia Tech
Penn State
Southern Miss

Let's talk about UVA's chances to make it onto the ballot in the future.  UVA actually tied with Rutgers in the formula and comes out ahead for having a better best game.  The reason I decided to include UVA in the consideration this week is because I figured a win over a 6-1 team (which was previously 6th on this ballot) would at least be enough to get them above some of these other teams I'd been considering.  (GT took a double whammy this week - the loss to a team that'd been nowhere near consideration, and UNC's loss to Miami.  They're verrrrrry close to USF, but a smidge behind.

As it turns out, UVA has some incredible anchors on its resume, which will prevent our Hoos from getting onto this ballot for a while.  It looks pretty terrible when your second-best win is over William & Mary - the Indiana and Idaho games are huge anchors on this thing.  UVA will probably need to put together a winning streak in order to earn my vote - unless there's a slew of upsets, I doubt beating NC State would be enough.

As for the actual ballot, well, pretty straightforward.  It's a little weird that Illinois moves up six spots after losing, but the fact is the main factor in their ranking is that they haven't had a bye.  It may be they need to drop a little.  The 10th through 17th spots are all tremendously close, and if you've got a convincing argument that a few things should be shuffled in that range, I'm open to hearing it.


Anonymous said...

In your formula, are wins automatically "better" than losses? In the case of UVA, for instance, I would consider our second-best outcome to be a close loss to Southern Miss (which is not impressive, but still a couple notches above the W&M win).

Also, you gotta figure out a way to work out that bye glitch. Get your geek on! Consider it a challenge. :)

Brendan said...

When I rank one team's games, yes, wins are automatically better than losses. This lets me compare wins to wins and losses to losses for the most part. It really minimizes me having to figure out whether a "competitive loss" is better than a medium-good win. However, when I compare Team A's Xth best game to Team B's, and one is a win and one a loss, then it's not automatic and a good loss is often better than a bad win.

I could easily fix the bye week thing, the question is whether or not I want to. Believe me, I've gone back and forth on that. The question is this: Isn't 6-1 better than 5-1? Since I think the answer is yes, the bye week "glitch" remains. As the season progresses, it becomes less of a thing, which is kind of the desired effect - there's a more pronounced difference between 4-0 and 3-0 than there is between 7-0 and 6-0.

The way it works out is that 6-1 (or 5-2 or whatever) teams with solid resumes end up benefitting from not having skipped a game, and 6-1 (or 5-2 or whatever) teams with weak resumes end up not being able to take advantage. Penn State is a perfect example of the latter.

Brandon said...

Had a busy weekend and didn't get around to looking at rankings 'til today, so this might be too late to effect change, but... did something go horribly wrong with Arkansas? They edged out your #11 team and slaughtered your #21, with a loss good ol' #1 which is a large reason #1 should _be_ #1, so I'm assuming them missing the cut is just a mistake.

Anonymous said...

Hmm, should've looked closer the first time through. What's going on with Illinois and UNC? Illinois gets spanked at home by an unranked (if respectable) opponent to fall from the ranks of the unbeaten, none of their other wins improve (you moved asu down a ton this week, for example), and yet they go up 6 slots? Meanwhile I can't imagine having UNC ranked at this point, given their best wins are us and rutgers, with 2 losses, but they move up off a home loss to an unranked opponent as well? Did something just break this week? I typically disagree quite a bit with your system, but there's usually logic behind things.

Brandon said...

Derp, last comment was mine as well.

Brendan said...

One thing to keep in mind is that I practically never take into account where a team was last week. I think that's a mistake - laddering - that the regular pollsters sometimes make: "This team was here, and they lost, so they must go down." Each week, each team's resume is looked at with a fresh eye.

That said, I agree about Arkansas. On another look, their top two wins should really be enough to get them in ahead of USF and WVU at a minimum. The system penalizes them for a huge dropoff in quality from the top two wins to the rest of their resume, but that's where common sense should've stepped in, as it sometimes does. It is too late to change, though - the poll comes out Tuesday mornings. I should make that clearer.

As for UNC - true, they lost, but their wins started looking better this week, which is why they got a little boost. Rutgers is 5-1 (a weak 5-1, which is why they ended up 32nd, but 5-1 regardless) and UVA, of course, moved up the ladder too. I do think they're about where they belong.