Sunday, November 13, 2016

Weekly Playoff Probabilities - Week 11

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

Rank Team Agg Rank POFF Prob Change
1 Alabama 1 99.54% 1.9%
2 Clemson 5 89.64% -5.6%
3 Ohio State 3 68.22% 25.4%
4 Michigan 2 42.73% -37.0%
5 Louisville 4 40.11% 26.1%
6 Washington 7 23.54% -24.6%
7 Western Michigan 23 12.57% 0.2%
8 West Virginia 19 9.68% 8.0%
9 Wisconsin 9 3.79% 3.0%
10 Penn State 17 3.24% 2.2%
11 Oklahoma 11 3.20% 2.6%
12 Colorado 13 2.27% 1.3%
13 Washington State 15 0.63% 0.5%
14 Nebraska 39 0.54% 0.4%
15 Oklahoma State 25 0.22% 0.2%
16 Utah 29 0.09% 0.1%
17 Florida 20 0.01% 0.0%

1.  When everybody loses, not everyone has to fall.  This week, that maxim applies to Clemson.  Even though a home loss to a decent Pitt isn't the best resume-builder, Clemson still has overwhelming odds to make the playoff.  The main reason for this is that they have nearly a 75% chance of winning out, and finishing as a 12-1 conference champ.  The secondary reason is what I mentioned a few weeks ago: Clemson has a very strong strength of schedule that ranks above everyone in the above list, save Alabama.  An 11-2 or 10-2 Clemson still has a reasonably good case.  If Clemson remains #2 in the committee's poll on Tuesday (which they very well could), then that will show that there's still some margin of error for the Tigers.

2.  You may notice that even though a lot of top teams lost, there still isn't a two-loss team with more than a 4% chance of making the playoff.  There are a couple of reasons for this.  First, only one of those nine teams has a greater than 50% chance of winning out (Penn State at 53%), while many are long shots (that many of these teams play each other and the teams above them has something to do with that).  My simulation had a median of just two of these teams winning out.  If you can be one of those two teams, then sure, you might have decent odds of making this year's playoff  But good luck figuring out which two teams those are.

Second, none of these teams have particularly good projected SOS numbers, which my model practically requires for teams in the two-loss bucket.  The best as of now is Colorado with a decent but not spectacular .537.  Only one of the nine two-loss teams to finish in the top four of the BCS standings had a lower SOS, and that was from the ridiculous 2007 season (Oklahoma with a .532 SOS).  It's not exciting, but if chaos ensues, you're still better off betting on one of the higher ranked teams.

3.  This was a strangely bloody week, as we say goodbye to more than a quarter of the remaining teams.  Auburn, Texas A&M, North Carolina, Virginia Tech, Baylor, and Minnesota were all eliminated.  Auburn and A&M represented the pair of two-loss teams with the best odds of making the playoff, which further reinforces that the current group of two-loss teams isn't as "playoff-worthy" as you might think.

4.  The Big 12 has life after all, as it's the only conference that avoided major upsets this past weekend.  The SEC is down to essentially just Alabama, so if a conference gets two teams, it'll be the Big Ten or ACC.

Conference Exp Playoff Teams
ACC 1.297
B10 1.185
SEC 0.995
P12 0.265
B12 0.131
MAC 0.126

5.  Congrats to Alabama and Louisiana Tech for their equally impressive accomplishments of winning their respective divisions.  In other news, Penn State is now a divisional favorite, and neither division of the Pac 12 has a team with better than 60% odds.  These last two weeks should be fun.

Conference Favorite Perc Runner-Up Perc
ACCA Clemson 92.9% Louisville 7.1%
ACCC Virginia Tech 90.5% North Carolina 8.4%
AMEE Temple 81.7% South Florida 17.0%
AMEW Navy 82.8% Houston 10.6%
B10E Penn State 47.8% Michigan 42.0%
B10W Wisconsin 88.9% Nebraska 7.1%
B12 Oklahoma 65.0% West Virginia 23.0%
CUSAE Western Kentucky 91.7% Old Dominion 8.3%
CUSAW Louisiana Tech 100.0%
MACE Ohio 91.4% Miami (OH) 8.7%
MACW Western Michigan 79.4% Toledo 20.6%
MWCW San Diego State 100.0%
MWCM Boise State 72.3% Wyoming 21.6%
P12N Washington 59.8% Washington State 40.2%
P12S Colorado 59.7% USC 26.5%
SECE Tennessee 54.1% Florida 45.9%
SECW Alabama 100.0%
SUN Troy 77.6% Arkansas State 15.9%

6.  One last random note:  Washington State has been on an absolute tear.  After starting the season 0-2 with an aggregate computer rating of 56, they have won eight straight games with an increasing measure of dominance.  After their second loss, my system had the Cougars winning out zero times, and averaging 7.7 losses per simulated season.  The chart below shows how Mike Leach's squad has improved with each week.  If they can finish this out, it will be the best recovery from two early losses since 2002 USC won the Orange Bowl and the Heisman.

After Week # Win Out (Per 10000) Avg Losses
2 0 7.7
3 0 7.0
4 1 6.9
5 11 5.9
6 75 4.9
7 170 4.3
8 213 4.1
9 185 4.0
10 387 3.7
11 735 3.4


Week 12 Preview

Home Away Home Win Prob Playoff Teams Lost
West Virginia Oklahoma 43.0% 0.041
Wake Forest Clemson 7.4% 0.033
Houston Louisville 14.4% 0.029
Michigan State Ohio State 5.3% 0.018
Colorado Washington State 62.2% 0.012
Washington Arizona State 94.2% 0.007
Michigan Indiana 97.8% 0.005
Rutgers Penn State 4.4% 0.001
Purdue Wisconsin 3.4% 0.001
TCU Oklahoma State 56.5% 0.001

Week 12 once again lacks a marquee game, but that didn't stop last week from being crazy.  The first four games all feature home dogs with a chance to wreck a team's playoff chances.  Odds are at least one of those games comes down to the wire.  Outside the top ten, LSU and Florida play the Purdue transfer QB bowl (winner gets David Blough), and San Diego State plays Wyoming in a critical game for Group of Five positioning.

No comments:

Post a Comment