Mo's Model: Week 3 Predictions

In Week 2, my model didn't do quite as well as in Week 1.  It was 33-8, though it correctly identified Georgia over South Carolina and Miami over Florida.  The funny thing is, one of my main motivations for creating the model was that I could NEVER seem to pick BYU right.  So I wanted this to take the subjectivity out of my BYU predictions.  It's 0-2 picking BYU so far!  Overall, it is 64-10, though after a retroactive change (I gave the service academies a small ratings boost: I had identified a systemic underrating of the academies in the preseason, but it was evident that, as I predicted outside of the model, they would beat Indiana, so I finally made that adjustment), it changes to 65-9.  I believe there are a few other teams I should apply that same logic too (Northwestern being the first and foremost).

So, here are the models Week 2 predictions:

Why Play These Games?  30+-point favorites
LSU over Kent State
Florida State over Nevada
Michigan over Akron
Northwestern over Western Michigan
Arkansas over Southern Mississippi
Kansas State over Massachusetts
Rutgers over Eastern Michigan
NIU over Idaho (at least it's a road game for NIU)

3-4 TD Favorites
Oklahoma over Tulsa
Arizona over UTSA
Boise State over Air Force
Rice over Kansas (my model likes Rice)
Middle Tennessee over Memphis
West Virginia over Georgia State
Washington over Illinois
Notre Dame over Purdue
Stanford over Army
Ohio State over Cal
Louisville over Kentucky (ACC exposing façade of SEC defenses this year)
Bowling Green over Indiana

10-20 Point Winners
Oregon over Tennessee
South Carolina over Vanderbilt
Auburn over Mississippi State
USC over Boston College
Ohio over Marshall
Penn State over UCF
Pittsburgh over New Mexico
Fresno State over Colorado
USF over Florida Atlantic
Alabama over Texas A&M

More than a FG, less than a FG/TD
Nebraska over UCLA
ECU over Virginia Tech
Utah over Oregon State (preseason, this was a toss-up)
UNLV over Central Michigan
New Mexico State over UTEP
Iowa over Iowa State

Toss-ups (within 3 points predicted)
Texas over Ole Miss (actually have two models predicting side-by-side, one has Ole Miss by 2, other has Texas by 1)
Arizona State over Wisconsin (same as with UT-Ole Miss)
Model had TCU edging out Tech: WRONG
Louisiana-Monroe over Wake Forest
Ball State over North Texas
Maryland over Connecticut