The value of a statistical model comes partially from its boldness: few college football fans would be excited by, for example, a system that predicted LSU would wax UAB. Then again, bold predictions that turn out inaccurate don't have much value: college football fans would have even less interest in a model that thinks the Blazers of Birmingham could have knocked off the Bengal Tigers of Baton Rouge. With this in mind, I thought it'd be interesting to take a look at some of bold predictions Sagarin (http://sagarin.com/sports/cfsend.htm) made that came true, as well as the ones that...well, didn't come true as much per se.
First the hits:
- Home-field advantage would give Miami the edge over Florida. I still think Sagarin is overstating the home edge (right now pegged at an absurd 5.6 points), but sure enough the usually-absent Cane fans helped deliver the one-possession win that was predicted.
- Georgia would bounce back against South Carolina. The model still loves the Dawgs (even ranking them 7 points higher than the Clemson squad that beat them last week), and felt they would win comfortably over Spurrier's Cocks. Projected margin of victory: 7 points. Actual margin of victory: 11 points.
- Texas should have been on upset alert traveling to BYU. BYU looked awful in a loss to lowly Virginia, but the computers correctly priced in either the effect of weather on that first match-up or Mack Brown's ability to Mack Brown this game away. If anything, the model was too forgiving of Texas, expecting the visiting Longhorns to squeak out a 1-point victory.
And now the misses:
- USC should handle Washington State easily. Sagarin was not impressed by the Dread Pirate's performance against Auburn, expecting the Men of Troy to roll to nearly a 3-touchdown margin of victory. Instead, the Cougars claw and scrape to an ugly 3-point win.
- Northwestern is not good enough to blow out Syracuse. The model picked the right winner, but thought this game could last into the fourth quarter when it was in fact over by halftime.
- Arkansas will blow the doors off of Samford. OK, an FCS team leading an SEC team not coached by John L. Smith in the fourth quarter? No one saw that coming.
You can't win 'em all. Now, let's look at the Sagarin rankings for the SEC after two weeks of football:
SEC-EAST 1. Georgia 89.18 #4 2. South Carolina 87.06 #8 3. Florida 84.76 #13 4. Tennessee 77.51 #37 5. Missouri 77.35 #38 6. Vanderbilt 76.45 #43 7. Kentucky 68.81 #74 ______________________________ SEC-WEST 1. Alabama 99.19 #1 2. LSU 88.39 #5 3. Texas A&M 87.48 #6 4. Auburn 79.23 #29 5. Ole Miss 77.23 #39 6. Arkansas 74.90 #50 7. Mississippi State 73.09 #55
The big mover this week was Auburn, going from last in the SEC West to bowl eligible after one win over a Sun Belt team (and, maybe, because the team they beat in Week 1 just beat USC on the road):
SEC-EAST 1. Georgia 11-1 (8-0) Capital One Bowl 2. South Carolina 11-1 (7-1) Sugar Bowl 3. Florida 8-4 (5-3) Outback Bowl 4. Tennessee 6-6 (3-5) Music City Bowl? 5. Missouri 6-6 (2-6) Liberty Bowl? 6. Vanderbilt 5-7 (2-6) 7. Kentucky 2-10 (0-8) ______________________________ SEC-WEST 1. Alabama 12-0 (8-0) BCS National Championship Game 2. LSU 10-2 (6-2) Cotton Bowl 3. Texas A&M 10-2 (6-2) Chick-Fil-A Bowl 4. Ole Miss 6-6 (3-5) Gator Bowl?
5. Auburn 6-6 (2-6) BBVA Compass Bowl?
6. Arkansas 5-7 (2-6) 7. Mississippi State 5-7 (2-6) ______________________________ SEC CHAMPIONSHIP: Alabama (-10) over Georgia
The rules for bowl assignments are fairly clear, at least for the top 6 teams. There are four 6-6 teams in this week's projections, however, so picking bowl assignments from the Gator Bowl down was pretty tough. I figured Ole Miss would generate the most excitement of those four, Auburn and Tennessee could cash in on home state fans, and Missouri would be the odd man out. Your personal projections may vary.
Elsewhere, we now have two additional undefeated BCS conference champions: Stanford and (perhaps due to Sagarin overrating Notre Dame) Michigan. Stanford is projected to beat current #2 Oregon in Week 11, whereas Michigan won't unseat Ohio State until the final week of the season; between that an a lower preseason ranking for the Wolverines, it should be the Cardinal in the title game. Clemson and Louisville both remain one-loss conference champs, losing in the final week of the regular season, joined this week by Oklahoma State, who Sagarin believes will inexplicably derp a game away to Texas. Northern Illinois still gets to play BCS buster with an undefeated MAC run.
For the BCS bowls, this leaves a glut of quality, big-name two-loss teams once the AQ slots are filled in. You've got Ohio State, Notre Dame, Florida State (unlikely, as that would be a regular season rematch), Oklahoma, and Baylor (remember, the SEC and Pac-12 already have two teams in) to choose from; if you can tell me who the Orange Bowl would pick out of that crew with any kind of certainty, you're a more knowledgeable person than I am.
BCS Matchups:BCS National Championship Game: Alabama (-9) over Stanford
Rose Bowl: Oregon (-5) over Michigan
Orange Bowl: Ohio State (-1) over Clemson
Sugar Bowl: South Carolina (-6) over Louisville
Fiesta Bowl: Oklahoma State (-12) over Northern Illinois
Roll Tide, y'all.