Unlocking the Playoff Committee Top 25
College Football’s #1 Projector of Weekly Playoff Rankings
By Adam McClintock and Dave Bartoo
The week prior to the Playoff Committee releasing their first top 25, Adam and I set out to create several models that could match what the committee did a year ago, in an effort to provide an advanced look at what to expect in their rankings. This exercise, if a reasonable model could be created, would also help establish a pattern of ranking behavior so we can track if they are being consistent in ranking or are an inexplicable moving target every week.
Since our first modeling we have refined our projections and have been the #1 advanced playoff committee projection model since the playoffs started the rankings in 2014.
This is our first go at it for 2016. We had to remodel the entire system after a huge bot from the Atlanta area wiped out a ton of data bases in an attack. In rebuilding the model, we were able to use the excellent work from Heather Dinich of ESPN of playoff committee members to enable us to read between the lines and find the bread crumbs of their methodology. Follow her on Twitter @CFBHeather.
We determined the best course is to model each week with the base 5.0 model, then apply the “PC” rules to re-ranking the teams. You can think of “PC” as short for playoff committee or politically correct, both may work in this situation.
Click to check out our Week 10 Ranks and Results
WEEK 11 Advanced Top 25 Rankings
UPDATE 11/8/2016: Ave top 25 miss 1.24 spots. Ave top 15 miss .80 spots. Biggest misses: A&M at #8 vs our #12 (assumed bigger drop with QB announced out for season) USC in a #20. We had them at #30 but this sets up a #4 vs #20 game with UW and USC. USC is in PAC12 N contention and this generates a bigger game frame. We should have hedged our bets and had USC at least at 25. Biggest shock is 6-2 Florida going from 11 to out. The committee has never drop a P5 team that far in a week. Violated a common “drop rule” in our formula. Funny thing, is they replaced them with 5-3 Arkansas (also SEC and bigger resume boost to #1 ranked team). SOS looks to be a big driver of ranks. Undefeated teams are in SOS order, same as 1 loss teams. Top 4 two loss teams are all #25 SOS or better. The rest all piling in the bottom half (except UNC at #42 SOS). Your 3 loss power 5s all have top 25 SOS ranks.
First group of 4:
1 Alabama SOS 5 CFB Playoff Committee Rank (1) Record 9-0
2 Clemson 10 (2) 9-0
3 Michigan 51 (3) 9-0
4 Washington 109 (4) 9-0
I think that order is pretty locked in. No changes there. Washington still has a terrible SOS. Ohio St very well could jump them; but there’s no reason to do that right now. These are the leaders of the committee’s top four conferences.
5 Ohio State 16 (5) 8-1
6 Louisville 78 (6) 8-1
7 Wisconsin 22 (7) 7-2
8 Auburn 11 (9) 7-2
Don’t see any reason to change the modeled order. Louisville won’t jump Ohio St due to their poor SOS, and Auburn did nothing to move them by Wisconsin. One loss Alabama, Clemson and Michigan will stay ahead of Louisville in the future. Badgers and Tigers still in contention for the conference title so they will stay close to the committee favorites until the lose.
9 Penn State 7 (10) 7-2
10 Colorado 53 (12) 7-2
11 Oklahoma 87 (11) 7-2
12 Texas A&M 14 (8) 7-2
PSU’s SOS is 7th and they blew out 5-3 Iowa (a QW), Colorado and OU may flip, but OU’s SOS and MOV is below Colorado’s. Plus the Big 12 is on the bottom of the power 5 for the committee. Same QWs. A&M could conceivably by ahead of Oklahoma and Colorado, but that was considered a bad loss against MSU and Trevor Knight is hurt. The committee looks at injuries so they leave A&M at the bottom of this group.
13 Virginia Tech (flip over Utah) 60 (14) 7-2
14 Utah 94 (15) 7-2
15 North Carolina 42 (17) 7-2
16 Oklahoma St 100 (13) 7-2
Va Tech’s SOS is better than Utah’s and they are closer to a conference title so we flipped what the model provided. North Carolina can’t go any higher than Virginia Tech cause of Head to head result. OSU’s SOS is trash, like much of the Big 12 so they are at the bottom of this group.
17 West Virginia 107 (16) 7-1
18 Florida 74 NR 6-2
19 Nebraska 49 (19) 7-2
20 Florida St 2 (18) 6-3
West Virginia won’t go higher than Oklahoma State cause of head to head result. Florida’s SOS is worse than Nebraska’s so those two could flip, but Nebraska’s MOD is larger than Florida’s. Florida St is the highest ranked 3 loss team due to SOS. Western Michigan could sneak into this group.
21 LSU 15 (24) 6-3
22 W. Michigan 108 (21) 9-0
23 Boise St (flipped with WSU) 63 (22) 8-1
24 Washington St 71 (23) 7-2
25 Minnesota 106 NR
LSU only took a 10 pt loss to the #1 team, they could be ahead of Florida St. Boise actually has a SOS better than Wazzu but the committee also looks at P5 teams played so they may keep them like the model predicted with Wazzu ahead of Boise. We are flipping them due to head to head.
San Diego St
USC 23 (20) 6-3
SDSU, Navy and Houston provide no support for top teams. USC could be in a 25 to make the UW/USC match up this week a top 25 tilt and a bigger TV game. Tennessee supports Alabama, but they really dont need it.