Weekly Playoff Probabilities - Week 8

As always, explanation here, ranking below, thoughts after.

Rank	Team	Agg Rank	POFF Prob	Change
1	Alabama	1	85.76%	-3.2%
2	Penn State	3	60.35%	8.7%
3	Georgia	6	49.97%	-1.4%
4	Notre Dame	4	30.36%	12.7%
5	Ohio State	2	29.38%	-9.7%
6	Miami (FL)	15	28.82%	-10.1%
7	Central Florida	16	26.67%	12.1%
8	Clemson	5	25.43%	1.9%
9	Wisconsin	9	24.44%	-2.9%
10	TCU	12	17.59%	-0.3%
11	Washington	7	11.01%	1.2%
12	Oklahoma	10	2.74%	-1.4%
13	North Carolina State	18	1.75%	-0.2%
14	Oklahoma State	11	1.74%	-0.1%
15	Virginia Tech	14	1.15%	0.7%
16	Washington State	23	0.72%	0.0%
17	Michigan State	30	0.60%	-0.1%
18	Auburn	8	0.41%	-0.1%
19	USC	20	0.35%	-5.9%
20	Stanford	13	0.34%	-0.6%
21	Mississippi State	17	0.19%	0.0%
22	Michigan	22	0.09%	-0.7%
23	Texas A&M	37	0.04%	-0.1%
24	Georgia Tech	21	0.04%	0.0%
25	LSU	25	0.02%	0.0%
26	South Carolina	39	0.01%	0.0%
27	South Florida	28	0.01%	0.0%
28	Iowa State	26	0.01%	0.0%
29	Kentucky	73	0.00%	0.0%
30	Arizona	35	0.00%	0.0%
31	West Virginia	33	0.00%	0.0%
32	Virginia	71	0.00%	0.0%

1. The "pretender" one-loss teams (Kentucky and Virginia) lost this past weekend, which leaves us with all of the legitimate contenders occupying the top 17 slots.  The most notable jumps belong to Penn State and Notre Dame, which makes sense as 1) they won the two biggest games of the day in dominant fashion, and 2) they're both really good.

2.  Since it was a pretty quiet week, a quick note about the differences between my model and 538's.  For the most part our results line up pretty well, but there are a few exceptions.  Most notably, the SEC teams, the Big Ten teams, and Notre Dame are all a bit lower in their implementation.  The reason for this is that their model weights conference titles fairly heavily whereas mine does not even consider them as a separate variable.  This means my model sees a number of possibilities where Georgia and Alabama (or Penn State and Ohio State if you prefer) both make the top four, while 538's model thinks there needs to be a fair amount of chaos for that to happen.  Furthermore, if you click the "win out" box next to Notre Dame, you'll see that their model only gives an 11-1 Irish team a coin flip's chance of making the playoff.  Mine thinks an 11-1 Notre Dame is basically a lock with a 92% chance in that scenario.  I generally respect the work that 538 does, but I think a 40 percentage point difference based on a single variable* is a bit strange.

*And remember, this is a variable with zero evidence of its own importance.  Just last year, we saw non-conference champ Ohio State beat out Penn State, and every other playoff participant has effectively qualified on their own merits.

3.  A brisk week for eliminations leaves us with Wake Forest, Iowa, Utah, and Texas Tech on the outside looking in.  As a reminder that my model is actually good, the Utes were at just 0.05% to make the playoff when they were 4-0 three weeks ago.

4.  Same story, different week for the conference-level playoff odds.  Ohio State's loss to the Sooners and Wisconsin's bad SOS makes it more difficult to envision a second Big Ten team in the playoff when compared to the Bama-Georgia pairing.  But if Penn State finishes a strong 11-1, I can see them pulling off a 2016 Ohio State and squeezing themselves in.

Conference	Exp Playoff Teams
SEC	1.364
B10	1.149
ACC	0.572
IND	0.304
AMER	0.267
B12	0.221
P12	0.124

5.  Two weeks ago, San Diego State had an 89% chance to win their division.  Now, Fresno State has the second best odds of any team in any division.  Part of this wild swing is due to the rest of the division being terrible, but part of it is due to the Bulldogs legitimate rise in quality in Jeff Tedford's first season.  In other news, every division now has a clear favorite thank in large part to Memphis and FAU turning into near-dominant squads.

Conference	Favorite	Perc		Runner-Up	Perc
ACCA	Clemson	57.3%		North Carolina State	42.4%
ACCC	Miami (FL)	77.3%		Virginia Tech	20.7%
AMEE	Central Florida	75.2%		South Florida	24.9%
AMEW	Memphis	73.6%		SMU	14.3%
B10E	Ohio State	64.9%		Penn State	33.3%
B10W	Wisconsin	99.7%		Northwestern	0.2%
B12	TCU	77.7%		Oklahoma State	52.2%
CUSAE	Florida Atlantic	56.2%		Marshall	38.4%
CUSAW	North Texas	57.1%		Southern Miss	36.2%
MACE	Ohio	58.6%		Akron	24.4%
MACW	Toledo	60.0%		Northern Illinois	27.1%
MWCW	Fresno State	98.6%		San Diego State	1.4%
MWCM	Colorado State	60.2%		Boise State	36.3%
P12N	Washington	52.8%		Stanford	38.5%
P12S	USC	58.0%		Arizona State	23.6%
SECE	Georgia	93.7%		South Carolina	3.1%
SECW	Alabama	84.6%		Auburn	11.5%
SUN	Appalachian State	60.9%		Arkansas State	28.3%

Week 9 Preview

Home	Away	Home Win Prob	Playoff Teams Lost
Ohio State	Penn State	67.3%	0.183
Notre Dame	North Carolina State	79.9%	0.038
Clemson	Georgia Tech	82.2%	0.023
Georgia	Florida	86.8%	0.022
Iowa State	TCU	31.9%	0.019
North Carolina	Miami (FL)	10.6%	0.010
Washington	UCLA	89.1%	0.006
Illinois	Wisconsin	3.1%	0.002
Oklahoma	Texas Tech	85.2%	0.002
West Virginia	Oklahoma State	23.3%	0.002

Now it starts.  Four of the next five weeks are the four biggest weeks of the year, per the "Playoff Teams Lost" metric.  Week 9 is a more top-heavy than some (which happens when the #2 team has a 2 in 3 chance of losing), but there's enough depth to keep me interested all day.  My favorite clash is ND's amazing O-line versus Bradley Chubb and the Wolfpack's ferocious D-line.  Just off the top ten list is a lovely late-night tussle between Arizona and Washington State.  There's also the game that could decide the Pac 12 South, with USC traveling to Arizona State.