Month: September 2009

It’s still way too early to know what sort of college football season we’re going to have, although it is starting to take shape. The general consensus so far seems to be that, while not quite as wild as 2007, the 2009 college football season may well be very upset-ridden… but it’s hard to know whether any of these games should really be considered upsets.

People have complained about the preseason polls for a long time, and this year seems to have underscored it. Ole Miss was ranked very highly in the polls, rising to #4 after the first three weeks, only to fall to South Carolina. That sparked a discussion as to whether Ole Miss should have had that much hype thrust upon them without this year’s version of the team having done anything to deserve it. After all, we’re hit with surprise teams in both directions all the time in pro sports; why should college be any less of a crap shoot to predict?

But the rest of the weekend seems to have suggested that’s not the problem. Miami (FL) wasn’t ranked at all in the preseason poll and rose to the top ten in only three weeks, only to be demolished by Virginia Tech. Meanwhile, the two teams considered the undisputed top two in the country on the basis of the preseason polls, Florida and Texas, thanks to the volatility of the early rankings, have zoomed up to take two of the top three spots in the C Ratings – they are who we thought they were, despite the Gators losing Tim Tebow during the Kentucky game. That the C Ratings already reflect the national consensus in this area suggests people should pay close attention to the fact Iowa stands between the two in the 2 hole – they are the real deal, and should be considered a budding national championship contender.

I heard someone on ESPN (Kirk Herbstreit?) suggest that the preseason polls are too useful for focusing our attention on certain teams early in the season, so we should keep them but make them meaningless later. Individual conferences have media and coaches’ preseason polls and don’t even have polls during and after the season; to say the conferences should have preseason polls and the entire country shouldn’t seems absurd. Here’s my solution: Have a preseason poll, but hold off on putting out any other polls until week 4 or so, maybe even as late as the release of the first BCS standings. That way, instead of having pollsters make requisite “tweaks” from week to week, you force them to look at the entire body of work of the first four weeks and free them to make a fresh, unencumbered ranking.

How the C Ratings are tabulated: First, A Ratings are tabulated by multiplying the total score ratio, which is expressed by (points-opponents’ points)/points, by the winning percentage. Score ratio minimizes the effect of running up the score. Next, B Points for each game are tabulated by (margin of victory)/(opponent’s A rating)+/-1 for wins, and -(margin of loss)/(1-opponent’s A Rating)+/-1 for losses. The “+/-” is + for road games and – for home ones. The total number of B Points is multiplied by the A Rating to get the B Rating. Conference Ratings are tabulated by averaging the B Ratings of all teams in the conference. (Independents are counted separately, and Army and Navy are counted as one conference.) Finally, the C Rating is tabulated by taking the difference between the team’s B Rating and his conference’s rating, taking a fraction of that equal to the fraction of Division I-A the conference makes up, and taking the result off the B Rating. The three ratings go A, B, C across. Click here to see the complete ratings.

38 teams total with positive C Rating (none with negative B Rating)

Off Top 25: #29 Florida State (was #16), #30 Arizona State (was #23), #32 Indiana (was #24), #33 Mississippi (was #14), #51 Miami (FL) (was #15), #74 California (was #7)

Unbeaten teams not on top 25: #26 Texas A&M*, #28 TCU, #36 Wisconsin

Rest of Watch List: #27 Stanford*, #29 Florida State, #30 Arizona State, #31 South Carolina*, #32 Indiana, #33 Mississippi, #34 Iowa State*, #35 Utah*

Other Positive B Ratings: #37 West Virginia*, #38 Wake Forest*, #39 Hawaii, #40 Duke, #48 SMU* (*=Newly Positive)

No Longer Positive: #43 Minnesota, #51 Miami (FL), #52 Southern Miss, #56 North Carolina, #57 Washington, #58 Kentucky, #63 Northern Illinois, #71 Colorado State, #74 California, #76 Louisville

Bottom 10: #111 Akron, #112 Memphis, #113 Washington State, #114 Buffalo, #115 Tulane, #116 San Jose State, #117 New Mexico, #118 Western Kentucky, #119 Miami (OH), #120 Rice

Conference Rating: #1 Big 12 (2.795), #2 SEC (1.982), #3 Big East (.811), #4 Big 10 (-.189), #5 ACC (-2.429), #6 Pac-10 (-2.436), #7 MWC (-9.326, leader #28 TCU), #8 WAC (-12.590), #9 C-USA (-16.583), #10 MAC (-18.081, leader #60 Central Michigan), #11 Sun Belt (-19.835, leader #79 Middle Tenn. St.)

Best game of week: UCLA @ Stanford, 12:30pm PT, ABC

All logos taken from Sportslogos.net, and are the trademarks of their respective schools, used without permission under fair-use clause of United States copyright law. Some logos may be out of date. Use of logos for commercial purposes without consent of the respective schools is prohibited.

Finally, it’s time to unveil this year’s edition of the college football rankings! I explain the formula behind the C Ratings, designed to factor in both margin of victory and strength of schedule in a sane fashion, in more detail here; an abbreviated version is below. The basic principle is that margin of victory and strength of schedule are interrelated: who you beat and how you beat them are important, but the Morgan Wick Ratings are unique in that your margin of victory against a given team is associated with that team. Two innovations protect against running up the score: diminishing returns in the A Rating for RUTS, and because margin of victory is associated with the team the margin of victory came against in the B Point formula, you don’t get much benefit for beating a scrub team big. (Ignore the part in the post about the B and C Ratings being overly high; I fixed that problem last season.) If I’d change anything, I’ve considered making C Ratings dependent on all a team’s opponents’ B Ratings, not just the rest of their conference regardless of whether they’ve played everyone or not, and fixing my database so FCS games are counted as being against a team with an A Rating of 0, instead of just not counting for B Points.

It’s so early that we’re going to get some really wacky results in the ratings, and you’re probably going to have to take everything with a grain of salt. We are, however, far enough into the season to make some early judgments, and one of them is that the Big East just may be due for another season like 2006, when – in contrast to the conference’s usual mire of mediocrity – at least three different teams were in the national championship conversation very late into the season. The Big East has had a fantastic nonconference slate so far, and their reward is that the team that looks to be the biggest diamond in the conference, Cincinnati, takes the top spot in this first edition of the rankings.

How the C Ratings are tabulated: First, A Ratings are tabulated by multiplying the total score ratio, which is expressed by (points-opponents’ points)/points, by the winning percentage. Score ratio minimizes the effect of running up the score. Next, B Points for each game are tabulated by (margin of victory)/(opponent’s A rating)+/-1 for wins, and -(margin of loss)/(1-opponent’s A Rating)+/-1 for losses. The “+/-” is + for road games and – for home ones. The total number of B Points is multiplied by the A Rating to get the B Rating. Conference Ratings are tabulated by averaging the B Ratings of all teams in the conference. (Independents are counted separately, and Army and Navy are counted as one conference.) Finally, the C Rating is tabulated by taking the difference between the team’s B Rating and his conference’s rating, taking a fraction of that equal to the fraction of Division I-A the conference makes up, and taking the result off the B Rating. The three ratings go A, B, C across. Click here to see the complete ratings.

39 teams total with positive C Rating (1 with negative B Rating (South Carolina))

2006 Boise State Title: #41 Oregon (2-1), .299, .070, -.018

Unbeaten teams not on top 25: #26 North Carolina, #27 Michigan, #29 South Florida, #30 Wisconsin, #34 Kentucky, #40 TCU, #42 Colorado State, #44 Texas A&M, #54 Southern Miss (last four not in positive C points, and Texas A&M not in positive B points)

Rest of Watch List (both 2-1): #28 Washington, #31 Arizona

Other Positive B Ratings (all 2-1 unless otherwise noted): #32 Duke (1-2 but Richmond doesn’t count and Army win outweighs loss to good Kansas team), #33 Ohio State, #35 Louisville (1-1), #37 Hawaii, #38 USC, #39 Minnesota, #41 Oregon, #55 Northern Illinois

Bottom 10: #111 Washington State, #112 New Mexico State, #113 Toledo, #114 Florida Atlantic, #115 Miami (OH), #116 New Mexico, #117 Rice, #118 Tulane, #119 Western Kentucky, #120 San Jose State

Conference Rating: #1 Big East (2.005), #2 SEC (1.604), #3 Big 12 (.787), #4 Big Ten (.693), #5 ACC (.186), #6 Pac-10 (-.980), #7 Mountain West (-9.883), #8 MAC (-11.405), #9 C-USA (-13.158), #10 WAC (-14.413), #11 Sun Belt (-15.238)

Best game of week: Iowa @ Penn State, 5pm PT, ABC

All logos taken from Sportslogos.net, and are the trademarks of their respective schools, used without permission under fair-use clause of United States copyright law. Some logos may be out of date. Use of logos for commercial purposes without consent of the respective schools is prohibited.