Weekly Playoff Probabilities - Week 9

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

Rank	Team	Agg Rank	POFF Prob	Change
1	Clemson	5	96.15%	12.3%
2	Alabama	1	94.52%	6.0%
3	Michigan	2	78.98%	0.1%
4	Washington	6	44.28%	6.1%
5	Ohio State	3	34.39%	4.4%
6	Texas A&M	9	17.36%	2.7%
7	Louisville	4	13.16%	-1.9%
8	Western Michigan	23	11.34%	0.1%
9	LSU	8	2.84%	-0.1%
10	Auburn	7	1.57%	0.7%
11	West Virginia	19	1.48%	-10.3%
12	Wisconsin	12	0.74%	0.1%
13	Penn State	20	0.69%	0.3%
14	Colorado	15	0.60%	0.1%
15	Oklahoma	10	0.49%	0.2%
16	Baylor	17	0.48%	-5.4%
17	Florida	13	0.40%	0.0%
18	Nebraska	25	0.31%	-0.7%
19	North Carolina	24	0.08%	0.0%
20	Oklahoma State	22	0.06%	0.0%
21	Washington State	28	0.05%	0.0%
22	Virginia Tech	18	0.03%	0.0%
23	Utah	36	0.01%	-0.1%
24	Minnesota	46	0.00%	0.0%

1.  Since the first official playoff rankings will be released on Tuesday, a disclaimer:  My rankings are not intended to mirror the week-to-week rankings of the committee.  Rather, the goal of this is to show who has the best odds of making the playoff in the end.  This means that a team that has a difficult remaining schedule (and thus a better chance to lose) will rank lower here than in the committee's eyes.  As this has been a rather straightforward year thus far, there probably won't be any stunning differences from the CFP ranks, but it's still an important distinction to remember.

2.  You may notice that the teams at the top have a really high chance of making the playoff.  For comparison's sake, the top team at this point last year (also Clemson) had just a 56% chance of making the playoff.  Part of this difference is the change in my methodology: My model now recognizes that Clemson and Alabama would have far better one-loss resumes than other teams, and rewards them accordingly.  But the main reason for the dominance of the top three is that they have three incredibly strong cases for playoff inclusion.  Here's a chart that helps explain this:

Team	0-Loss Odds	1-Loss Odds	P(POFF|1LOSS)	2-Loss Odds	P(POFF|2LOSS)	Adj SOS
Clemson	58.65%	34.45%	93.60%	6.68%	29.52%	0.600
Alabama	47.25%	42.84%	94.39%	9.75%	32.49%	0.603
Michigan	42.87%	48.35%	67.66%	8.49%	5.57%	0.555

Obviously, all three teams have about a 50% chance of going undefeated at this point, and if that happens any and all three of these teams is guaranteed a playoff berth.   Where the real difference comes into play is in the scenarios where these teams lose one game.  In this case, Clemson and Alabama's insane strength-of-schedules (yes, a .600 SOS is really good) put them in the company of teams that almost always make the top four.*  Even if these teams drop a couple of games, there's still some chance of making the playoff, although you would have to think the teams that take them down (LSU?  Auburn?  Virginia Tech?) might leap them in the committee's estimation.  Yes, my model is more aggressive with these teams than others are, but all available evidence points to a runaway for the teams at the top.  We shall see.

*For reference, the eight teams that have finished the season with one loss and an SOS of .600 or greater have all made the top four.  This group includes two eventual champions: 2006 Florida and 2015 Alabama.

3.  We say goodbye to six more teams this week: Boise State, Florida State, Tennessee, Pittsburgh, Wake Forest, and Maryland are all officially eliminated.  The first three teams resigned in the middle of last week's rankings (10th, 12th, and 14th), which is part of what helped to collapse the rankings even more than they currently were.

4.  The Big 12 is basically eliminated now, so the picture is even clearer:  If Washington slips up, we'll see a second team from a conference in all likelihood.  Should be fun.

Conference	Exp Playoff Teams
SEC	1.167
B10	1.151
ACC	1.094
P12	0.449
MAC	0.113
B12	0.025

5.  Conference races continue to take shape.  Clemson is a mortal lock to win the Atlantic after their win, and several other divisions how have a dominant leader as well.  Wisconsin is not in first in the Big Ten West, but they are a better team than Nebraska and have an easier remaining schedule, plus that tie-breaker due to their win last night (grrr).  Boise is still the favorite after their loss (they're still two touchdowns better than Wyoming in the composite ratings), but it should be a much more interesting race now.  For my money, the best race is in the Sun Belt, where Troy and App State should battle it out until the end.

Conference	Favorite	Perc		Runner-Up	Perc
ACCA	Clemson	98.6%		Louisville	1.4%
ACCC	Virginia Tech	78.4%		North Carolina	21.1%
AMEE	Temple	76.4%		South Florida	22.3%
AMEW	Navy	46.3%		Tulsa	27.5%
B10E	Michigan	70.5%		Ohio State	23.5%
B10W	Wisconsin	58.0%		Nebraska	25.2%
B12	Oklahoma	65.0%		West Virginia	14.8%
CUSAE	Western Kentucky	78.6%		Middle Tennessee	15.4%
CUSAW	Louisiana Tech	65.9%		Southern Miss	32.6%
MACE	Ohio	81.1%		Akron	13.9%
MACW	Western Michigan	87.7%		Toledo	12.2%
MWCW	San Diego State	99.2%		Hawaii	0.7%
MWCM	Boise State	55.4%		Wyoming	39.3%
P12N	Washington	79.4%		Washington State	20.6%
P12S	Colorado	74.3%		Utah	14.3%
SECE	Florida	86.2%		Tennessee	9.5%
SECW	Alabama	75.1%		Auburn	13.6%
SUN	Appalachian State	47.2%		Troy	46.6%

Week 10 Preview

Home	Away	Home Win Prob	Playoff Teams Lost
LSU	Alabama	30.7%	0.120
Ohio State	Nebraska	87.0%	0.026
California	Washington	13.8%	0.020
Mississippi State	Texas A&M	17.8%	0.016
Clemson	Syracuse	96.3%	0.012
Michigan	Maryland	97.8%	0.006
Ball State	Western Michigan	10.1%	0.004
Boston College	Louisville	5.3%	0.003
Northwestern	Wisconsin	30.6%	0.003
Penn State	Iowa	69.5%	0.002

Much like previous weeks, there is one game that far and away has the most playoff implications.  Verne's final night game should be a doozy.  Other than that, there are still a couple of other big games, including some decent road tests.  If Virginia and Michigan State could compel one-score games last week, it's reasonable to expect that one or two of these home underdogs will make some noise.

Weekly Playoff Probabilities - Week 8

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

Rank	Team	Agg Rank	POFF Prob	Change
1	Alabama	2	88.48%	10.5%
2	Clemson	5	83.86%	6.0%
3	Michigan	1	78.87%	14.0%
4	Washington	6	38.20%	0.3%
5	Ohio State	4	29.95%	-38.6%
6	Louisville	3	15.11%	3.2%
7	Texas A&M	9	14.65%	-12.9%
8	West Virginia	18	11.78%	7.6%
9	Western Michigan	22	11.22%	1.5%
10	Boise State	25	7.40%	2.5%
11	Baylor	15	5.89%	2.5%
12	Florida State	10	5.01%	-0.2%
13	LSU	8	2.93%	1.8%
14	Tennessee	20	2.25%	0.8%
15	Nebraska	23	0.98%	-0.2%
16	Auburn	7	0.88%	0.4%
17	Wisconsin	12	0.66%	0.2%
18	Colorado	16	0.46%	0.3%
19	Florida	13	0.40%	0.1%
20	Penn State	28	0.35%	0.3%
21	Oklahoma	11	0.30%	-0.1%
22	North Carolina	24	0.12%	0.1%
23	Utah	37	0.10%	0.1%
24	Washington State	26	0.07%	0.0%
25	Oklahoma State	27	0.04%	0.0%
26	Virginia Tech	19	0.02%	0.0%
27	Pitt	33	0.01%	0.0%
28	Wake Forest	60	0.00%	0.0%
29	Maryland	62	0.00%	0.0%
30	Minnesota	54	0.00%	0.0%

1.  Week 8 saw Ohio State and Texas A&M basically lose half a playoff team from their probabilities.  The majority of the spoils from this went to the top three teams, which all saw slight upward movement in their favor.  As a result, the odds for Alabama, Clemson, and Michigan may seen quite high, but I think I can account for this fairly logically.  First of all, there are just nine other power conference teams with fewer than two losses.  This creates a relatively small pool of true competitors, and many of those teams (Utah) are probably more pretender than contender.  Second, all three of the top teams are legitimately good which means they're not likely to collapse (all three teams have better than a 80% chance of finishing with one loss or fewer).  Finally, Alabama and Clemson boast two of the top strength of schedules in the nation.  If either were to trip up and finish with one loss, my algorithm (based on the NCAA's definition of SOS) says they would both have better than a 90% chance of being selected for the playoff, based on what happened to similar teams in the past (for reference 2008 Florida - one of the best teams of the millennium - is the nearest comparison in terms of SOS for one-loss teams).  There is certainly still a chance for chaos, but we still need an awful lot of it to avoid the seemingly inevitable.

2.  While a few teams avoided elimination in notable ways, we still wave goodbye to seven teams this week.  Arkansas, Miami (FL), North Carolina State, Stanford, TCU, Arizona State, and Iowa all took their third loss this week.  None of these teams were in the top 20 of the aggregate rankings (Miami is 21st...so close), so none were realistic contenders at this point anyway.

3.  Ohio State's loss (in concert with real improvement from West Virginia) means we no longer have a conference with strong odds of placing two teams in the playoff.  This doesn't mean it still can't happen, but the Pac 12 and Big 12 are looking stronger than they were, which makes it more likely one of them sneaks a team in.

Conference	Exp Playoff Teams
B10	1.108
SEC	1.096
ACC	1.041
P12	0.388
B12	0.180
MAC	0.112
MWC	0.074

4.  I made two changes to the conference title odds this week.  One, upon popular request, I added the Big 12 and Sun Belt to the list.  I had previously only been calculating the division winners for conferences with title games (so I could in turn simulate those).  But, it still matters who wins the other conferences, so I wanted to include them here.

The second change concerns tie-breakers.  You may have noticed that (for example) Louisville had maintained a greater than 20% chance of winning the ACC Atlantic, in spite of having an effective 1.5 game deficit to Clemson.  This was because I had simply randomized the divisional winner in the case of ties.  Since there are now several examples where head-to-head tie-breakers are playing important roles in determining division champs, I did the work to make my code able to pick the winner.  This new functionality also shows up a couple of other places, most notably in swinging the ACC Coastal odds towards Virginia Tech and the West division of the American towards Navy.

Conference	Favorite	Perc		Runner-Up	Perc
ACCA	Clemson	91.3%		Louisville	8.5%
ACCC	Virginia Tech	57.9%		North Carolina	38.0%
AMEE	Temple	71.9%		South Florida	24.6%
AMEW	Navy	66.7%		Houston	13.7%
B10E	Michigan	70.0%		Ohio State	25.1%
B10W	Wisconsin	45.5%		Nebraska	43.0%
B12	Oklahoma	47.5%		West Virginia	28.6%
CUSAE	Western Kentucky	71.1%		Middle Tennessee	22.7%
CUSAW	Louisiana Tech	60.4%		Southern Miss	33.0%
MACE	Akron	50.3%		Ohio	46.6%
MACW	Western Michigan	79.3%		Toledo	20.7%
MWCW	San Diego State	96.0%		Hawaii	3.7%
MWCM	Boise State	93.8%		Wyoming	5.5%
P12N	Washington	77.8%		Washington State	22.0%
P12S	Colorado	73.3%		Utah	17.2%
SECE	Tennessee	51.7%		Florida	46.1%
SECW	Alabama	79.0%		LSU	9.0%
SUN	Troy	51.3%		Appalachian State	38.8%

Week 9 Preview

Home	Away	Home Win Prob	Playoff Teams Lost
Florida State	Clemson	38.3%	0.142
Utah	Washington	18.5%	0.026
Oklahoma State	West Virginia	48.6%	0.022
Wisconsin	Nebraska	70.8%	0.010
Texas	Baylor	36.3%	0.009
Ohio State	Northwestern	94.4%	0.009
Michigan State	Michigan	3.2%	0.008
Ole Miss	Auburn	40.5%	0.004
South Carolina	Tennessee	16.5%	0.004
Wyoming	Boise State	14.9%	0.004

Week 9 is similar to the past couple of weeks in that there is one big game with a couple of nice runners-up.  That said, the big difference is that this top ten has eight home underdogs, many of which have a real chance of pulling the upset.  In addition to this top ten, there are a number of other matchups featuring home underdogs (Oregon State-Wazzu, Virginia-Louisville, Pitt-VT), so get ready for ESPN to call it "Home Underdog Week Presented by Home Depot" or something.