I got behind schedule
last week and didn’t really catch up. This week I changed up my process a bit
so hopefully this will help going forward. The experimental numbers take a more
hands on approach than the current set do. (Full ranks for current and experimental
sets)That said…
We finally see some
change in the top 25 and the top 10 in the current set of numbers. Georgia
moves up, probably to where it should have been a couple of weeks ago. Florida State
takes a dip after a couple poor performances. Wisconsin continues to win and
win solidly, but I’m not sure the upside is much more than fringe top 10 with a
weak schedule. Miami (FL) is the biggest mover this week coming up from 29 to rest
much closer to its current perceived value, but still as the fourth best ACC
team. If it helps, my numbers bother me sometimes as well. LSU moves back to 20
after a good win on the road at Florida, and Kentucky sneaks back in at 25.
|
Rank
|
Team
|
Rating
|
Last
Week
Rank
|
|
1
|
Alabama
|
25.879
|
1
|
|
2
|
Washington
|
21.098
|
2
|
|
3
|
Clemson
|
19.714
|
3
|
|
4
|
Ohio State
|
17.021
|
4
|
|
5
|
Penn State
|
17.013
|
5
|
|
6
|
Georgia
|
15.240
|
10
|
|
7
|
Auburn
|
15.146
|
6
|
|
8
|
Washington State
|
15.102
|
9
|
|
9
|
USC
|
14.183
|
8
|
|
10
|
Wisconsin
|
13.704
|
12
|
|
11
|
Florida State
|
13.385
|
7
|
|
12
|
Oklahoma
|
11.272
|
11
|
|
13
|
Oklahoma State
|
10.963
|
14
|
|
14
|
Colorado
|
10.917
|
13
|
|
15
|
TCU
|
9.451
|
16
|
|
16
|
Virginia Tech
|
9.289
|
18
|
|
17
|
Georgia Tech
|
8.523
|
17
|
|
18
|
Miami (FL)
|
8.131
|
29
|
|
19
|
USF
|
8.086
|
22
|
|
20
|
LSU
|
8.085
|
28
|
|
21
|
Oregon
|
8.075
|
19
|
|
22
|
Michigan
|
8.010
|
15
|
|
23
|
Texas A&M
|
7.969
|
21
|
|
24
|
Louisville
|
7.892
|
20
|
|
25
|
Kentucky
|
7.851
|
26
|
|
Dropped
Out
|
|||
|
26
|
Tennessee
|
7.803
|
23
|
|
27
|
Florida
|
7.666
|
24
|
|
28
|
Vanderbilt
|
7.544
|
25
|
As we move to looking
at the new set of numbers it’s becoming immensely clear to me that an
adjustment to my system will be necessary this off-season. I need to strike a
balance between a system that doesn’t make swings just for the sake of
adjustment (current system) and a system that moves a bit quicker to make
adjustments based on results and performance. The experimental system has
caught and surpassed my current system on a basis of accuracy, or more
accurately, based on difference from actual results. A combining of the two
will allow me to more quickly eliminate the preseason ranks from the
predictions, but I want/need to keep some semblance of the stability in ranks
that we see in the current system. There is lots of tumult from week to week in
the experimental rankings, and that makes them more adaptable, but I we also
see teams drop or rise and return to similar rankings from week to week and that
is frivolous change that needn’t be made. Sorry for the ramblings, I needed to
verbalize this to start to figure out what changes need to be made. I think
limiting the swings in the new system, and a quicker deletion of the impact of
the preseason rankings are necessary, but, a games played component for the phase
out may be necessary to ease the smoothing out process.
Again, sorry about
the ramble, it’s more therapeutic writing, but it’s the point of following both
systems. In the experimental top 25, the first thing I notice is that the top
four swaps in Georgia for Ohio State and that seems fine. Washington State and
Wisconsin move up after good in conference performances, Oklahom takes a
sizable seven spot drop for the Iowa State fiasco, TCU falls a touch off hard
fought win against WVU, and Miami continues to correct up after a good win over
formerly 18th ranked FSU. 20-25 shows the most turnover outside of
23rd ranked Georgia Tech, with UCF making the jump to 20 after back
to back thrashings of AAC foes, and if I’m honest I think UCF/SDSU/Stanford all
self-corrected here because of some transcription errors that I only just
caught and rectified this week. Vandy falls back to, if we’re honest, they
should have been all along, FSU/UF took rivalry losses, and Colorado takes a long
fall to a much more understandable rank. You see here why I like this set of
numbers, and if I’ve fixed the issues, hopefully will see less huge swings at
the top.
|
Rank
|
Team
|
Rating
|
Last
Week
Rank |
|
1
|
Alabama
|
51.149
|
1
|
|
2
|
Washington
|
41.341
|
2
|
|
3
|
Clemson
|
40.670
|
3
|
|
4
|
Georgia
|
36.546
|
6
|
|
5
|
Penn State
|
35.604
|
4
|
|
6
|
Ohio State
|
32.991
|
5
|
|
7
|
Wisconsin
|
28.928
|
11
|
|
8
|
Washington State
|
28.317
|
12
|
|
9
|
Auburn
|
27.872
|
8
|
|
10
|
USC
|
25.273
|
10
|
|
11
|
Oklahoma State
|
24.456
|
9
|
|
12
|
TCU
|
23.987
|
13
|
|
13
|
Miami (FL)
|
22.085
|
17
|
|
14
|
Oklahoma
|
21.841
|
7
|
|
15
|
Notre Dame
|
18.531
|
27
|
|
16
|
USF
|
17.733
|
15
|
|
17
|
Virginia Tech
|
16.882
|
20
|
|
18
|
Michigan
|
16.228
|
14
|
|
19
|
Oregon
|
14.347
|
16
|
|
20
|
UCF
|
14.159
|
34
|
|
21
|
NC State
|
14.014
|
29
|
|
22
|
San Diego State
|
13.434
|
32
|
|
23
|
Georgia Tech
|
13.170
|
23
|
|
24
|
Stanford
|
12.895
|
43
|
|
25
|
Kentucky
|
12.507
|
26
|
|
Dropped
Out
|
|||
|
27
|
Louisville
|
11.540
|
21
|
|
30
|
Utah
|
11.027
|
19
|
|
31
|
Florida State
|
10.879
|
18
|
|
34
|
Florida
|
9.934
|
22
|
|
37
|
Vanderbilt
|
8.943
|
25
|
|
39
|
Colorado
|
7.592
|
24
|
There we have it, we’re
quickly encroaching on the half way point in the season, and we’re learning a
lot. I’m looking forward to continuing to watch how both sets of numbers
perform and improving my system as time moves along. Thanks to all who read
what I write, and sorry for missing last week.
No comments:
Post a Comment