Dewey Rankings Week 4
If you care to look, please see my updated rankings below. Technically, week four is not yet complete, though I doubt that Hawaii - Louisiana Tech will change things. Here we go:
Rank Team Rank Points
1 Iowa 1.726
2 Alabama 1.713
3 Texas 1.660
4 Florida 1.631
5 Virginia Tech 1.565
6 Boise St 1.533
7 Nebraska 1.504
8 Stanford 1.457
9 UCLA 1.453
10 Kansas 1.453
11 LSU 1.442
12 Oregon 1.418
13 Auburn 1.412
14 South Florida 1.395
15 Ohio State 1.391
16 Oklahoma 1.371
17 Texas A&M 1.367
18 TCU 1.363
19 Michigan 1.352
20 South Carolina 1.336
21 Cincinnati 1.324
22 Southern Cal 1.317
23 Penn State 1.315
24 Clemson 1.285
25 Missouri 1.269
If you read my post last week, you know that I am trying to improve on Colley. In that vein I have compared my results to that of Colley and the ESPN/USA Today Poll. The first part of success is total convergence among the teams - the teams in my top 25 are the teams in theirs - without taking the differences in rank of each team into account.
Last week it took 37 teams to represent the total of our top 25 lists. This week it is also 37 teams.
Last week I had 19 teams in common with Colley, this week it is 16. Last week I had 17 teams in common with ESPN, this week it is 18. Last week Colley had 15 teasm in common with ESPN, this week it is 17. Both last week and this week 13 teams were common to all three of our lists.
Of note: Colley and I both have Iowa #1, Florida #4, and Va Tech #5.
To the comment that I look at the retrodictive accuracy of my rankings, I am going to check my system against the MinV system at the end of the season and will post those results. I haven't looked too deeply, but it seems MinV breaks down where you have multiple undefeated teams or no undefeated teams and multiple teams with one loss - if the only criterion is to minimize where victors are ranked below a team they beat, how do you distinguish amongst one loss teams for example?
Again, I welcome all feedback.
909
September 27th, 2009 at 1:26 PM ^
It's interesting, but no sane person thinks Iowa is the best team in the country. It's something different though.
September 27th, 2009 at 1:58 PM ^
I can see the Iowa win on the road being huge and vaulting ahead of the other undefeated teams. However, then the system needs to rank PSU higher, to justify its Iowa ranking.
On a side note, Florida's schedule is a joke this year. I'm sick of SEC hype. Florida plays one tough game in the SEC before the title game, and they have a bye before it, two weeks from now at LSU.
I'm going to be upset when Florida loses and is put in Title game anyway. Shenanigans I say!!
September 27th, 2009 at 2:09 PM ^
Bama's schedule IMO is easier than Florida's..
Regardless.. there's a good chance that 1-4 will all drop a game along the way.. I wonder what pollsters will do about Boise State.. this might be the year they send a non BCS team to the natl championship...
September 27th, 2009 at 2:15 PM ^
There are no words to describe the craziness of the rankings I see above.
**edit**
It should be called the knee-jerk rankings.
Iowa
South Florida
September 27th, 2009 at 2:12 PM ^
Iowa at one after beating an extremely overrated PSU team. Hmm...don't see much sense at all in that
September 27th, 2009 at 2:57 PM ^
Per Sagarin
#21 Arizona 3-1
#23 PSU 3-1
#49 Northern Iowa (highest FCS team) 3-1
#70 Iowa State 3-1
Sagarin ranks them #5 this week, in the same ballpark as #1 for the numerical systems.
http://www.usatoday.com/sports/sagarin/fbt09.htm
Michigan is #30
#90 WMU 2-2
#32 ND 3-1
#159 EMU 0-3
#68 IU 3-1
FWIW, Delaware State 1-2 is ranked #184
September 27th, 2009 at 3:05 PM ^
Thanks ITG - I was going to make the same point. Iowa has beaten three solid FBS teams, two on the road. My system doesn't take location into account, but I think Sagarin does.
Anyway another point is there is no knee-jerk - my system is a margin of victory model where only the results count.
909
September 27th, 2009 at 3:37 PM ^
I respect you putting forth the effort. I am a subscriber to using models in conjunction with perception. I think Iowa should be ranked, but not #1. USF beat FSU, but also beat a 1-3 Charleston Souther team, a 0-4 Western Kentucky team, and a 1-3 Wofford team. I would give them votes, but not thrust them into the top 15.
(sorry if I was being a little harsh earlier)
September 28th, 2009 at 1:40 AM ^
Your system only makes sense, if at all, after ALL the games have been played. It makes no sense in the middle of the season because it only measures games that have been played and is biased toward teams who have their tougher games earlier. If that was your goal, then there is no sense checking it on a weekly basis. It does not make any sense on a weekly basis. And, obviously, it has no predictive value. I am not sure if that is your goal or not, but if it isn't it seems like it would be more useful to analyze past years than the current year because you will not have the data you need to analyze the current year until the season is over.
September 27th, 2009 at 3:05 PM ^
I dont want to sound like a homer but A&M shouldnt be in the top 25 in my opinion, let alone be ahead of UM. They are 3-0 and have played New Mexico, Utah State, and UAB. All powers right? They play Arkansas this week and if they win, sure they should be in the Top 25 i guess, but I still wouldnt move them ahead of Michigan assuming Michigan wins at EL.