Sunday, November 27, 2016

Weekly Playoff Probabilities - Week 13

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

Rank Team Agg Rank POFF Prob Change
1 Alabama 1 99.74% 0.5%
2 Ohio State 3 96.80% 15.3%
3 Clemson 4 88.95% -3.7%
4 Washington 5 48.07% 21.0%
5 Michigan 2 22.60% -29.1%
6 Western Michigan 22 16.51% 4.1%
7 Wisconsin 12 8.24% 1.8%
8 Penn State 14 6.49% -0.6%
9 Colorado 13 6.45% -2.3%
10 Oklahoma 8 4.85% -3.0%
11 West Virginia 19 0.76% 0.0%
12 Oklahoma State 17 0.54% -0.5%

1.  So, The Game came and went and as expected, the results shook things up.  You may notice that a few of the two-loss teams dropped a bit at the same time, but the cause was something else: Washington clearing a major hurdle.  The Huskies' blowout win in the Apple Cup served to keep them in the one-loss bucket and boosted their schedule a bit by adding the one good Pac-12 team they hadn't played (Colorado).  We now have roughly even odds of having a straight-forward playoff with three conference champs with one or fewer losses, and one really strong 11-1 non-champ. 

2.  That said, if we do get a little chaos, it would appear from the chart above that Michigan has the best odds of making the playoffs.  While that may be true (their SOS is slightly ahead of everyone else in the two-loss bucket), the main reason for the gulf is that they have finished their season, and everyone else has to play a tough game.  Once the results shake out, the victors of championship week should pull much closer to the Wolverines.

3.  Fun week for eliminations with one obvious (Florida), one not-too-surprising (Nebraska), and one WTF (Louisville).  In case you were wondering, Louisville is still 7th in the aggregate ratings because of how awesome they were the first two months of the season.  But they reeeeeally aren't making the playoff now, nor do they deserve to (in any definition of the word).


Week 14 Preview

Home Away Home Win Prob Playoff Teams Lost
Colorado Washington 30.9% 0.119
Clemson Virginia Tech 81.7% 0.081
Penn State Wisconsin 46.1% 0.035
Alabama Florida 91.0% 0.030
Oklahoma Oklahoma State 73.8% 0.017
Ohio Western Michigan 9.1% 0.005
West Virginia Baylor 77.6% 0.002

As one would expect, we have a decent slate for championship week.  That said, it's not as exciting as last year's edition (which had the two biggest games of the year), nor is it as big as last week (which had OSU-Michigan).  The "playoff importance" of the season peaked last week for a couple of reasons.  One, Ohio State and Michigan were in much stronger positions than the two-loss glut, so they each had further to fall.  Two, while it's decently likely a two-loss team makes the playoff, there's so many of them that it divides up the odds such that none of them can be involved in that "important" of a game.  And finally, there isn't a true win-and-in game out there.  The Pac-12 title game comes the closest, but I still think Colorado is more likely than not to miss the playoffs if they win.  Penn State-Wisconsin also exists in this vein, but I wouldn't be shocked if Michigan remains above them due to the head-to-head wins and a non-conference victory over a potential conference champ (Colorado).  We shall see.

Sunday, November 20, 2016

Weekly Playoff Probabilities - Week 12

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

Rank Team Agg Rank POFF Prob Change
1 Alabama 1 99.25% -0.3%
2 Clemson 4 92.66% 3.0%
3 Ohio State 3 81.49% 13.3%
4 Michigan 2 51.71% 9.0%
5 Washington 6 27.06% 3.5%
6 Western Michigan 23 12.39% -0.2%
7 Colorado 13 8.75% 6.5%
8 Oklahoma 8 7.82% 4.6%
9 Penn State 14 7.12% 3.9%
10 Wisconsin 12 6.41% 2.6%
11 Louisville 5 1.75% -38.4%
12 Nebraska 32 1.54% 1.0%
13 Oklahoma State 19 1.07% 0.8%
14 West Virginia 28 0.74% -8.9%
15 Florida 18 0.26% 0.2%

1.  A relatively quiet Week 12 has just one major victim: Louisville.  What's interesting about their fall is that the top teams benefit more than the cream of the two-loss crop.  The reason for this is the same as it was earlier in the season:  My model likes the current one-loss teams better as two-loss teams than the current two-loss teams.  Clemson, Ohio State, and Michigan would have far stronger SOS metrics than anyone else in that group, so a second loss may not hurt any of them as much as you might think.  Still, some of the two-loss teams did improve their standings, especially those (Colorado and Oklahoma) who passed strong tests.  Yes, I'll fix the WMU problem next year, but I'm leaving it for now, because it's funny.

2.  Only two eliminations this week:  Utah and Washington State.  Clearly I jinxed the Cougs when I wrote a section on them last week.  Sorry.

3.  All season it's looked like there would be room for a conference to send two teams to the playoff; we just weren't sure if it would be the SEC, ACC, or Big Ten.  After 12 weeks of play, the first non-conference champ playoff berth is the Big Ten's to lose.

Conference Exp Playoff Teams
B10 1.483
SEC 0.995
ACC 0.944
P12 0.358
MAC 0.124
B12 0.096

4.  Most conferences and divisions are down to one or two contenders.  The main exception is of course, Ohio State, which is looking like a strong playoff contender in spite of the long odds to win the division.  In three instances (Bedlam, Apple Cup, and WMU-Toledo), we have a game that will decide a division/conference for the victor, so that's fun.

Conference Favorite Perc Runner-Up Perc
ACCA Clemson 100.0%
ACCC Virginia Tech 90.3% North Carolina 8.6%
AMEE Temple 93.3% South Florida 6.7%
AMEW Navy 100.0%
B10E Penn State 48.5% Michigan 42.6%
B10W Wisconsin 88.5% Nebraska 7.3%
B12 Oklahoma 72.4% Oklahoma State 27.6%
CUSAE Western Kentucky 92.1% Old Dominion 8.0%
CUSAW Louisiana Tech 100.0%
MACE Ohio 85.0% Miami (OH) 15.0%
MACW Western Michigan 75.3% Toledo 24.7%
MWCW San Diego State 100.0%
MWCM Wyoming 61.4% Boise State 35.4%
P12N Washington 65.9% Washington State 34.1%
P12S Colorado 80.5% USC 19.5%
SECE Florida 100.0%
SECW Alabama 100.0%
SUN Arkansas State 79.6% Appalachian State 12.4%


Week 13 Preview

Home Away Home Win Prob Playoff Teams Lost
Ohio State Michigan 57.3% 0.317
Alabama Auburn 86.0% 0.046
Washington State Washington 34.9% 0.046
Colorado Utah 79.6% 0.017
Clemson South Carolina 96.8% 0.015
Penn State Michigan State 84.7% 0.010
Western Michigan Toledo 75.3% 0.010
Iowa Nebraska 59.9% 0.009
Wisconsin Minnesota 86.1% 0.008
Iowa State West Virginia 29.2% 0.002

The Ohio State-Michigan game is officially the most important game of the past two years, surpassing last year's Big Ten title game.  The next three games on this list are all good and in different time slots than all the other games.  So not only is this a great weekend, but it is also set up optimally for your viewing pleasure.  Enjoy.