Power Ratings Forecast 2015

Each year I post my pre-season rankings based on 2014 finish, returning starters, and regression to the team’s weighted average historical mean. Here are my forecasted ratings for this year. Note that this is not a forecast for what the Final Top 25 will be; those rankings are partially associated with W/L merit and human perception rather than raw scoring margin in context of strength of schedule (ex. Arkansas wasn’t ranked because they were 7-6, but they were very unlucky. It was clear to computers that their ability to play with Alabama, and shut out Ole Miss and LSU, warranted a top 25 ranking). Instead this is a forecast for what their final power ratings will be.

TEAM Average PR 2015
Ohio State 95.51
TCU 93.85
LSU 88.33
Alabama 88.02
Baylor 87.35
Arkansas 87.20
Georgia 87.06
Oregon 86.63
Michigan State 86.10
USC 85.81
Notre Dame 84.81
Tennessee 84.61
Oklahoma 84.07
Clemson 83.08
Georgia Tech 82.95
Ole Miss 82.71
Texas A&M 82.26
Stanford 82.09
UCLA 82.08
Auburn 82.01
Wisconsin 81.70
Missouri 81.13
Virginia Tech 80.90
Utah 80.38
Florida 80.35
Nebraska 78.78
Arizona State 78.38
Mississippi State 78.24
Kansas State 77.74
Florida State 77.14
Boise State 77.02
Texas 76.79
Penn State 76.58
Oklahoma State 76.52
West Virginia 76.49
California 75.79
Pittsburgh 75.75
Miami (FL) 75.11
Louisville 74.94
Iowa 74.69
Texas Tech 74.61
Arizona 74.26
BYU 74.25
Minnesota 73.78
South Carolina 73.34
North Carolina State 72.97
North Carolina 72.77
Michigan 72.44
Kentucky 71.86
Washington 70.96
Cincinnati 70.70
Marshall 69.86
Utah State 69.63
Virginia 69.58
Appalachian State 69.22
Temple 69.08
Colorado 68.11
Northern Illinois 68.08
Louisiana Tech 67.61
Memphis 67.29
Houston 67.29
Boston College 67.29
Purdue 67.02
Duke 66.52
Toledo 66.34
Western Michigan 66.26
Navy 66.12
Illinois 66.07
Northwestern 66.00
Iowa State 65.61
Western Kentucky 65.54
Vanderbilt 65.25
UCF 65.16
Maryland 64.85
Oregon State 64.76
Washington State 64.62
Arkansas State 64.42
Georgia Southern 64.32
Wake Forest 64.12
Ball State 63.83
Indiana 63.25
East Carolina 63.20
Syracuse 63.18
Colorado State 62.43
Rutgers 61.97
San Diego State 61.79
Rice 61.77
Fresno State 61.52
Bowling Green 60.69
Central Michigan 60.21
Air Force 60.17
Hawaii 60.08
Middle Tennessee 60.04
Massachusetts 59.84
Nevada 59.25
Tulsa 58.52
Ohio 58.39
Florida International 58.25
New Mexico 57.65
Louisiana-Lafayette 57.31
South Florida 56.48
San Jose St 56.33
Southern Miss 56.28
Tulane 56.22
Kansas 55.84
Akron 55.24
Kent State 54.53
Texas State 54.25
Louisiana-Monroe 54.10
UTEP 54.04
North Texas 53.66
Florida Atlantic 53.14
South Alabama 52.97
Troy 52.39
Buffalo 52.22
Old Dominion 51.66
New Mexico State 51.45
Wyoming 50.66
Connecticut 50.41
Miami (OH) 49.57
Charlotte 49.17
UNLV 48.71
Idaho 48.60
SMU 47.48
Georgia State 45.04
Army 44.52
UTSA 42.18
Eastern Michigan 41.50

Comments:

  • Wow, LSU is high. But maybe this isn’t as unreasonable as it seems. Despite an 8-5 season the Tigers finished with a pretty high power rating. Maybe I should account for Chavis leaving (if I deduct 2-3 points to adjust for Chavis leaving that would put LSU in the range of Oklahoma and Michigan State). But even so, with Leonard Fournette returning and Travin Dural and Malachi Dupre back at WR, if they can get a quarterback (a rather big “if”) they’ll be a major contendor in the SEC West.
  • If LSU is overvalued due to Chavis leaving, then Texas A&M is undervalued due to Chavis arriving. It’s not quite the same impact necessarily – a coach leaving has an immediate impact, while one arriving may take a while for him to start seeing results. Still, a Texas A&M with a defense would be pretty dangerous.
  • Stanford’s using some interesting technology. Stanford has been one of the first teams to adopt virtual reality training for their team, and it sounds like it’s helping out a lot. I don’t know how to quantify that, and right now my rankings match the preaseason polls. But if there’s something to this…watch out for the Cardinal.
  • Texas shouldn’t be that high. I’d prefer to have them in the 77 power rating range, which would have them ranking in the low 40’s. One of the inputs of this system looks at a regression to a historical mean, and Texas has a very high historical mean. But coaching changes often change the culture of a team. Then again, the foundation of the program is still high 4-star and 5-star recruits, so the idea that they should eventually perform well makes sense. I have made a manual adjustment to Baylor because I don’t think it’s fair to regress them to a mean based on games before Briles was the coach. The difference is pretty stark — it would put the Bears at an 83 rather than an 89 (the latter seeming much more sensible).
  • UTSA almost at the bottom? They return 5 starters. That’s really bad, and our sample size is pretty limited on teams that started a new team from scatch with all freshmen, graduated them with seniors, and had to start over. Larry Coker is a good enough coach and has been planning to fill some of these holes though. He signed a very small 2014 recruiting class so that he could have room for 2015 early enrollees and JUCO transfers. He also did his best to work in freshmen last year even if they weren’t starting. I see them being better than teams like Georgia State, Idaho, etc.