Sunday, February 15, 2009

Basketblogpoll ballot #2

This week's ballot, with comments after the jump that isn't really a jump but more of a "ballot."

RankTeamDelta
1 Connecticut
2 North Carolina
3 Pittsburgh 1
4 Oklahoma 2
5 Duke 2
6 Memphis 1
7 Michigan St. 3
8 Louisville 10
9 Clemson
10 Missouri 6
11 Kansas 8
12 Xavier
13 Illinois 10
14 Wake Forest 7
15 Villanova 2
16 Marquette 3
17 Gonzaga 8
18 Washington
19 Butler 8
20 Arizona St. 4
21 Louisiana St. 1
22 Utah St. 7
23 Davidson 2
24 Dayton 10
25 Utah
Last week's ballot

Dropped Out: UCLA (#8), Florida (#22).


Unlike last week, there's a methodology and a philosophy. Last week I rolled in from my ski trip at 8 at night, sat down, and threw the thing together, so I was afraid I'd see all these new faces on the ballot and win the Mr. Schizo award or whatever gets handed out in the Blogpoll writeup as a reward/punishment for uber-rejiggering of your ballot. Only two teams means I must have sorta done something right last week.

First, I looked at the top 40 teams in the RPI, Sagarin, and Pomeroy ratings; if a team doesn't show up in at least one of those top 40s, it's not worth a look. That leaves about 55 teams.

Then I got rid of all the teams that had pretty shaky records. For example: all three rankings systems have West Virginia jacked up pretty high - Pomeroy has them 6th - but their resume consists mainly of beating Villanova at home, two other ranked teams (OSU and Georgetown) that got bumped from here for the same reason (shaky records) and a lot of losses to good teams. That's way more bubblicious than their computer rankings say.

Whatever teams are left that all three ranking systems think are top-25 get "autobids" and are then ranked amongst themselves. This is where I start looking at who you beat and who you lost to, etc. This is also where the conference rankings come in. I used Sagarin's. Rather than spend all afternoon figuring out whether beating TFSU is better than beating South Carolina, I take the semi-lazy way out and call a 6-4 record in the ACC better than the same in the SEC. This is, for example, why you don't see Florida any more.

After that there's a pool left of about 25 teams to fill out the last 10 or so spots. Philosophy takes over here, and I'll freely admit that these aren't necessarily the "best" 25 teams in the country at this point. If I wanted that, I'd just get superlazy and stop at the ranking averages. Top 25 by average win! So, sure, could California, who I omitted, beat LSU? Could Purdue beat Utah State? Very possibly. But LSU, Utah State, Utah, etc. are in much better shape for getting to the Big Dance than the mid-good teams in the power conferences. They've also done a lot more winning. Therefore I'm more inclined to rank 'em. The line gets drawn at the extremes, like Siena, but in general, a team that's 24-2 ought to be rewarded and ranked, even if they didn't quite make the KenPom top 50.

UPDATE: per the below comment, which makes sense, I've rejiggered the Pac-10 teams a bit. I didn't really take last week's results into account when moving teams up and down because I started this week's rankings with a clean slate and the assumption that last week's might have been a bit hasty. But now that you mention it, Anonymous, there's really no reason UCLA should be ranked and ASU not. In fact there's really no compelling case UCLA should be ranked at all as they seem to be no better than third best (at best) in the Pac-10.

Also punished Clemson and Duke a spot each and Butler two spots for losing Sunday.

1 comment:

Anonymous said...

Dropping ASU out after they beat UCLA? Interesting.