My analysis is based on taking the computer's rankings that are available (Sagarin), those that publish their algorithm that I've reproduced (Colley Matrix), and guessing at the changes for the three others that have been releasing their rankings. I've also included the latest Coaches poll and guessed at updates to the Harris poll.
Here is the rough result:
| Rank | Team | Points |
|---|---|---|
| 1 | Oklahoma | 0.9244 |
| 2 | Oregon | 0.9215 |
| 3 | Boise State | 0.9189 |
| 4 | TCU | 0.8606 |
| 5 | Auburn | 0.8310 |
| 6 | LSU | 0.8113 |
| 7 | Alabama | 0.7327 |
| 8 | Michigan State | 0.7188 |
| 9 | Utah | 0.6516 |
| 10 | Ohio State | 0.5847 |
Oklahoma makes its move based primarily on the computer where they should be a near unanimous #1. Oregon's #1 in both polls is offset by the computers having them around #7 and with Boise State #4 in the computers, the gap is there, narrow as it is, for OU to slip in. It is very close though so any slight error in my extrapolation/guessing could change it so any of the 3 are #1.
There is a bit of a gap to #4 TCU with Auburn and LSU fairly close for #5 and #6 (they should be #2 and #3 in the computers). It is interesting that it appears Alabama will be #7 ahead of undefeated Michigan State and Utah, although Michigan State is close so could eke by.
We'll see how close this is!
Where does Colley have OU, Oregon and Boise?
ReplyDelete@Scott By my calculations Colley has OU #1, Boise St #7, and Oregon #9. It is Oregon at #9 and at least 2 others having them about the same that kills their lead in the polls.
ReplyDeleteMy guesses for the others are as follows (is this close to what you have?):
ReplyDeleteTeam Mass Bill A-H
OU 1 2 3
Boise 4 4 7
TCU 6 5 8
Oregon 8 6 10
@Scott
ReplyDeleteI'm estimating
Team Mass Bill A-H
OU 1 1 2
Boise 2 4 5
TCU 4 6 6
Oregon 9 5 9
Just proves it's difficult to predict the computers! My thesis was on NCAA tourneys, so it's a passion of mine as well. Keep up the good work!
ReplyDeleteYeah, each one seems to have some challenges and without being able to use MOV, some teams are a bit low in many of the computers. But I'm pleased with how accurate my prediction of the BCS top-10 was. I had all 10 and just had a couple swapped, otherwise dead on!
ReplyDeleteI was happy that I had OU at #1 by 8 pm cst Saturday and predicted Auburn over TCU by 1 pm cst Sunday. Didn't attempt 6 through 10. OU, ironically, has been helped by MOV before it was removed and helped again after it was removed. Humans certainly have more influence now than before, but not as much as one would think. As you know, poll rankings mean a LOT less than poll points now ...
ReplyDeleteI believe it is better now using poll points instead of poll rankings. In situations where there is a large gap or tightly packed group of teams, it gives a much more accurate result. Even with my computers rankings I tend to focus more on the rating. The ranking is just the result of sorting based on the rating.
ReplyDeleteYes, it would be nice if they just took all of the computers' ratings and transformed the final rating to a normal(0,1), then used some type of BCS computer score based on the ratings. Maybe average each team's middle 4 and then rescale the averages so that the maximum value is 1.
ReplyDelete