Saturday Sports! K-Ranks: Top 25 and Bowl Picks
December 16, 2011
by Kevin Leyden
It was a dark and stormy night. Somewhere in the world, anyway.
I, the college football enthusiast, pored over the latest edition of the BCS standings and tried to use it to rank my Nightcap Bowl Mania Confidence Page (you can register for free for a chance to win a $50 prize or T-shirt here. Fore more info, check out the facebook page). I thought about all the ways it attempts to rank teams that might never play each other. What exactly is a strong loss? How should margin of victory be adjusted for home-field advantage? What's worse: an early loss to a team with a bad record or a late loss to equal competition?
And what in the world are the agendas of the poll voters?
The list of questions is endless, and the BCS standings aim to reveal the answers. But the real question is: Who cares?
I realized that I could break down rankings as much as I liked and never uncover much. Politics and inordinate emphasis on wins and losses, rather than actual performance, would always keep me from answering the question I really care about, the one I imagine most of you would probably like answered as well. Between two teams, who's better?
It's all fine to compare teams by resume to determine which are most worthy of playing in better bowl games. But a team's profile - headlined by its wins and losses - explains what has happened each time a team has stepped on the field. It's great at predicting the past. However, there's a lot of randomness involved in who wins and loses any given game.
I decided to create a rating system that cuts out all the garbage and is based entirely on performance. My model is built on three pillars: efficiency, tempo, and opponent strength. I crunched some numbers for every game this season between FBS teams: all 15 weeks, all 120 teams. For each game, I quantified each team's performance by fitting it to the following scenario.
Team X played at some level against Team Y. Now they're playing at the same level in a hypothetical game against a team that is the national average at everything, Team Z (roughly Washington State). Evaluated on a per-possession basis, how many points will Team X's offense score on Team Z's defense, and vice versa?
It was a fairly extensive undertaking, but I am stoked with the results. So here's my Top 25, based entirely on performance and completely adjusted for schedule strength. All of this was analyzed game by game; I took no shortcuts. I've shown records here for fun, but the other numbers you will see are offensive points per 100 possessions against Team Z and points allowed on defense per 100 of Team Z's possessions. Enjoy!
Alabama Crimson Tide (11-1, 7-1 SEC): 351, 44
LSU Tigers (13-0, 8-0 SEC): 418, 66
Wisconsin Badgers (11-2, 6-2 B10): 471, 155
Florida State Seminoles(8-4, 5-3 ACC): 292, 101
Oklahoma State Cowboys (11-1, 8-1 B12): 409, 144
Stanford Cardinal (11-1, 8-1 P12): 422, 152
Oklahoma Sooners (9-3, 6-3 B12): 303, 110
USC Trojans (10-2, 7-2 P12): 322, 129
Georgia Bulldogs (10-3, 7-1 SEC): 261, 107
Arkansas Razorbacks (10-2, 6-2 SEC): 406, 168
Texas A&M Aggies (6-6, 4-5 B12): 334, 141
Oregon Ducks (11-2, 8-1 P12): 339, 144
Michigan State Spartans (10-3, 7-1 B10): 254, 110
Michigan Wolverines (10-2, 6-2 B10): 339, 156
Florida Gators (6-6, 3-5 SEC): 283, 133
Virginia Tech Hokies (11-2, 7-1 ACC): 263, 127
South Carolina Gamecocks (10-2, 6-2 SEC): 226, 113
Boise State Broncos (11-1, 6-1 MWC): 332, 166
Clemson Tigers (10-3, 6-2 ACC): 357, 181
Vanderbilt Commodores (6-6, 2-6 SEC): 259, 132
Missouri Tigers (7-5, 5-4 B12): 261, 133
Miami Hurricanes (6-6, 3-5 ACC): 339, 179
Kansas State Wildcats (10-2, 7-2 B12): 320, 173
Nebraska Cornhuskers (9-3, 5-3 B10): 342, 186
West Virginia Mountaineers (9-3, 5-2 BE): 315, 174
But this is only the beginning; with efficiency and tempo stats in hand, this model can predict future games as well. Bowl season is right around the corner, so without further ado, I present performance-based bowl predictions for 2011-12.
Utah State 44, Ohio 35
San Diego State 35, UL-Lafayette 27
Florida International 29, Marshall 17
TCU 29, Louisiana Tech 20
Boise State 45, Arizona State 33
Southern Miss 27, Nevada22
Missouri 24, North Carolina 20
Western Michigan 37, Purdue 31
Louisville 19, NC State 17
California 31, Texas 28
Florida State 21, Notre Dame 13
Baylor 55, Washington 44
Rutgers 20, Iowa State 17
Mississippi State 24, Wake Forest 20
Oklahoma 48, Iowa 20
Texas A&M 47, Northwestern 23
Georgia Tech 23, Utah 21
Vanderbilt 30, Cincinnati 20
Penn State 24, Houston 20
Georgia 17, Michigan State 16
South Carolina 23, Nebraska 21
Florida 27, Ohio State 21
Wisconsin 41, Oregon 32
Oklahoma State 39, Stanford 38
Michigan 24, Virginia Tech 23
Clemson 42, West Virginia 39
Arkansas 44, Kansas State 34
Pittsburgh 27, SMU 12
Northern Illinois 42, Arkansas State 29
Alabama 11, LSU 9
We'll be keeping track of the model's success here at the Claremont Sports Connection and come back after bowl season with a recap. But if you can't wait until then, or you want to show me just how wrong I am, you can lay down some picks against me and the rest of the Nightcap team! Send us an email at talk@claremontsportsconnection.com and we'll get you set up in no time, or you can go to the google docs survey, which takes you to the ESPN Page, or go to our facebook page for more information. But remember, bowl season begins Saturday, December 17, so don't wait! Enjoy the bowl season!