Comments on: Five weeks in, there’s no stopping the Crimson Tide http://www.footballperspective.com/five-weeks-in-theres-no-stopping-the-crimson-tide/ NFL History and Stats Fri, 05 Oct 2012 22:22:25 +0000 hourly 1 https://wordpress.org/?v=6.9.4 By: George http://www.footballperspective.com/five-weeks-in-theres-no-stopping-the-crimson-tide/#comment-2178 Fri, 05 Oct 2012 22:22:25 +0000 http://www.footballperspective.com/?p=3085#comment-2178 One last look at this – one game I am now really interested in Oaklahoma (-4.5) at Texas Tech – SRS, Least Squares and Weighted (90%) Least Squares have Texas Tech winning outright and comfortably (by a touchdown and not accounting for homefield advantage as well), so how comes they are getting points? Really can’t wait for tomorrow now to find out how the numbers stack up as this just seems wrong.

]]> By: George http://www.footballperspective.com/five-weeks-in-theres-no-stopping-the-crimson-tide/#comment-2086 Mon, 01 Oct 2012 21:21:02 +0000 http://www.footballperspective.com/?p=3085#comment-2086 Hi Chase,

Great post as usual. Re: Richie above, UCLA were expected to hammer Colorado by about 30 so that wasn’t a better than expected win (as odd as that may sound). I’ve done the least squares thing whilst I still can (199 teams in the mix, so I can’t adjust the homefield advantage – due to Solvers variable limit – which I’ve left at 2.89). Average is about 27.78 which you could add to my ratings to get to around the SRS scores. I still have the Youngstown State anomaly, and have constrained all the teams to an average of 0 (or this week an Eastern Washington) – anyhow my top 20 straight up:

Youngstown State 43.36167109
Alabama 43.12574177
Texas 41.18096787
Texas Tech 39.31195051
Oregon 38.46409007
Texas A&M 38.32974699
Florida 35.50061599
Cincinnati 35.32575234
South Carolina 34.37559472
Georgia 33.8005889
Kansas State 33.31412715
Oklahoma State 31.86761154
McNeese State 31.63943189
Florida State 30.33280959
Louisiana State 29.65177962
Baylor 27.75956199
West Virginia 26.76016787
Pittsburgh 26.30681457
Notre Dame 26.18460225
Arizona State 25.63353533

Now same constraints as above but adjusting and giving the games a 90% weighting (e.g. week 4 only counts 90% as much as week 5) – and we have Texas at the top clearly:

Texas 50.92345278
Youngstown State 47.76876156
Florida 44.71067652
Oklahoma State 43.90033306
Texas Tech 40.0296585
Texas A&M 39.00769511
Alabama 38.31862204
South Carolina 37.84048124
Oregon 37.04289653
Georgia 34.96139195
McNeese State 32.362075
Pittsburgh 30.86193181
Florida State 30.77633278
Kansas State 29.1823608
North Carolina 29.0866791
Notre Dame 28.16117873
West Virginia 27.47133299
Cincinnati 27.27788048
Baylor 27.04118227
Louisiana State 25.06161428

Incidentally the model is closer to the previously observed/verified stochastic element for College Football this week with a standard deviation of error around a predicted margin of victory of about 8.95 for the unweighted and 10.75 for the weighted – against an expected 16 (courtesy of Jeff Sagarin apparently).

]]>