Weekly Playoff Probabilities - Week 11

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

Rank	Team	Agg Rank	POFF Prob	Change
1	Alabama	1	88.25%	11.7%
2	Miami (FL)	11	75.87%	36.6%
3	Clemson	6	59.04%	15.2%
4	Wisconsin	8	48.38%	8.8%
5	Georgia	7	43.33%	-23.7%
6	Notre Dame	10	29.10%	-22.8%
7	Central Florida	18	19.97%	-21.9%
8	Oklahoma	9	13.71%	7.1%
9	Ohio State	2	7.45%	2.0%
10	Penn State	3	6.32%	-0.1%
11	Auburn	5	4.57%	3.6%
12	USC	17	2.59%	1.4%
13	Washington	4	0.55%	-11.7%
14	TCU	13	0.37%	-5.6%
15	Washington State	24	0.22%	0.2%
16	Oklahoma State	12	0.16%	0.1%
17	Michigan	19	0.15%	0.0%

1.  The biggest weekend of the year predictably leads to a bunch of shuffling up top.  Most of this probably makes sense....the top five seems like it's in the right order, and it makes sense that Auburn is behind the Big Ten teams because the prospect of beating Alabama and Georgia in back-to-back weeks is still pretty daunting.  But the teams in spots 6-7-8 seem pretty weird.  So let's talk about them.

Notre Dame - The Irish took their second loss in grand fashion, which seemingly knocks them out of the running.  But my model gives them a near 1 in 3 chance of making it.  The main reason for this is their preposterous .622 predicted SOS, which is easily the best in the field.  Only two two-loss teams have finished with an SOS in that neighborhood and both of those teams (1999 Bama and 2002 USC) made the BCS top four.  Now, obviously the committee has slightly different logic than the BCS did, but my underlying assumption that that logic has basically the same effect suggests that the Irish aren't dead yet.  It will be interesting to see where ND falls in the next set of rankings, especially in relation to fellow two-loss teams Auburn and Ohio State.

Central Florida - I mentioned previously that, while I love the Knights, my model was probably over-estimating them.  Much like the Irish, they are exhibiting a very strong SOS relative to their bucket.  Their .524 projected SOS is well above the gold standard of mid-majordom, 2009-10 TCU. But from the committee's initial rankings of UCF, it's pretty clear that nothing short of utter chaos is getting them into the field.  So I did a half measure, and re-ran the Group of Five model with 2017 UCF as a (negative) data point.  This has the effect of flattening out the curve for such teams, while still rewarding SOS to some degree.  The end result probably still overrates the Knights, but for a mid-season change I made in 30 minutes, I'll take it.

Oklahoma - The Sooners collected their third big win of the season, and will probably sit third or fourth in Tuesday's rankings.  So why is the team so low?  Well first of all, they only have a ~50% chance of winning out, so there's still some work to do.  But mainly, the really weak SOS of the Big 12 is hurting them a bunch.  A .513 projected SOS puts them in mixed company...for example, 2014 TCU and 2015 Iowa had similar ratings with one loss, and neither made the Playoff.  While it's probably true that the Sooners have little margin for error, I think my model is underrating them.  Sure, a good deal of their schedule consists of terrible teams, but if they finish 12-1, they'll have 4 or 5 wins against top-25-level teams, which the committee appears to eat up.  Should the Sooners make the Playoff, I may have to examine if other SOS measures are more predictive of the committee's behavior.

2.  Just three eliminations this week, with Mississippi State, Michigan State, and Virginia Tech all taking a third loss.  We are left with all of the top 13 teams in the aggregate rankings and only one team outside of the top 20.  This (in theory) means the championship games should give us the most even set of games in some time.

3.  There's been talk of two bids for the SEC for some time, but now we also have to entertain talk of two bids for the ACC as well.  Miami's position is suddenly quite strong, and it's entirely possible that they could survive a loss to Clemson and still make the Playoff.  In other news, the Pac 12 now has to rely on a USC team that will probably be 10th in the standings on Tuesday.

Conference	Exp Playoff Teams
SEC	1.361
ACC	1.349
B10	0.623
IND	0.291
AMER	0.200
B12	0.142
P12	0.034

4.  Five divisions have been decided, all in power conferences (this makes sense, as those teams are more likely to take the penultimate week off and are more likely to have non-conference rivalries to end the year).   In the Mountain West, we will likely get a title game preview the week before (great scheduling, guys).  And the Sun Belt is nuts after another wild upset...this time USA over Arkansas State.

Conference	Favorite	Perc		Runner-Up	Perc
ACCA	Clemson	100.0%
ACCC	Miami (FL)	100.0%
AMEE	Central Florida	78.6%		South Florida	21.4%
AMEW	Memphis	98.8%		Houston	0.8%
B10E	Ohio State	90.0%		Penn State	4.5%
B10W	Wisconsin	100.0%
B12	Oklahoma	97.1%		TCU	65.1%
CUSAE	Florida Atlantic	99.7%		Florida International	0.3%
CUSAW	North Texas	97.4%		Southern Miss	1.4%
MACE	Ohio	81.3%		Akron	18.7%
MACW	Toledo	72.0%		Northern Illinois	19.1%
MWCW	Fresno State	76.4%		San Diego State	23.1%
MWCM	Boise State	97.5%		Wyoming	2.5%
P12N	Stanford	73.3%		Washington State	16.2%
P12S	USC	100.0%
SECE	Georgia	100.0%
SECW	Alabama	61.1%		Auburn	39.0%
SUN	Appalachian State	39.4%		Arkansas State	26.7%

Week 12 Preview

Home	Away	Home Win Prob	Playoff Teams Lost
Wisconsin	Michigan	76.9%	0.038
Temple	Central Florida	12.5%	0.025
Notre Dame	Navy	93.0%	0.020
Miami (FL)	Virginia	92.8%	0.018
Georgia	Kentucky	93.8%	0.013
USC	UCLA	80.8%	0.005
Penn State	Nebraska	95.9%	0.003
Kansas	Oklahoma	1.8%	0.001
Texas Tech	TCU	28.2%	0.001
Ohio State	Illinois	99.0%	0.001

The best week of the year is followed by the worst.  Six of the top ten games have >90% odds for the favorite, and the rest aren't all that great.  Watch Wisconsin-Michigan in the early slot and then go find something else to do.  Rest up for the last two weeks.