Arkansas heads into the second of a brutal four-game stretch away from home, still needing to “steal” a win to help offset the BYU loss. The Hogs have plenty to build on from the LSU loss, but the game against Texas A&M isn’t going to be easy, even if it is at a neutral site.
Our Model Rankings
It’s time to debut or first model rankings for the season. Note that we’ve only had four weeks, and most teams have only played three games that count for the model, since we don’t do FBS-vs-FCS games. That’s a small sample size. We include no preseason projections whatsoever, so the numbers are usually unstable early, though I think they are starting to stabilize.
How are these numbers calculated? The formula is understandably complex, but there are two parts. First, the “noise adjustment”, based on this method from ProFootballFocus. Noise is anything that happens during the game that has a significant impact on the final score of that game, but isn’t necessarily predictive of future performance. Some “noisy” plays include breakaway touchdowns, turnovers, missed field goals, penalties, and third-and-long conversions. The ultimate “noisy” play was KJ Jefferson’s fumble in last year’s Texas A&M game. That play cost Arkansas 13.7 points in terms of EPA… the offense had 6.7 for having the ball right there at the 1, and A&M ended up getting 7 out of it the other way. It was a huge play that told the story of that game, but for predictive purposes, it was pretty noisy. Crazy events have their impacts tone down, so the “adjusted” drive points of each drive reward long, on-schedule drives that end in a touchdown.
To account for quality of opponent, we take a team’s adjusted drive points through a ridge regression, similar to what we use to create our xRAPM player rankings for basketball. The ridge regression considers the defense faced over every drive for every team, adjusting the number based on quality.
The output is what we’re calling Predicted Points per Drive. Here are the top 5 offenses, per our model:
- Washington +1.04
- USC +0.74
- Oregon +0.72
- Oklahoma +0.72
- LSU +0.69
And the top 5 defenses:
- Oklahoma +0.61
- Penn State +0.54
- Texas +0.53
- Tennessee +0.51
- Michigan +0.48
And the top 5 overall:
- Washington +1.39
- Oklahoma +1.33
- Oregon +1.14
- Michigan +1.03
- Miami +0.92
If you’re curious, our rankings almost perfectly match ESPN analyst Bill Connelly’s raw SP+ rankings.
Here’s how the SEC shakes out, with overall FBS rankings:
- Texas A&M 9th
- LSU 11th
- Tennessee 13th
- Georgia 16th
- Florida 19th
- Kentucky 24th
- Alabama 25th
- Auburn 28th
- Mizzou 40th
- Ole Miss 50th
- South Carolina 55th
- Arkansas 73rd
- Vanderbilt 105th
- Mississippi State 115th
So it’s out of the frying pan and into the fire for the Hogs, who go from facing the 2nd-best team to the best. Both A&M and LSU had high hopes and were written off due to big losses early, but the model says not so fast. It thinks LSU’s margin of defeat to Florida State was largely fluky (the noise-canceled final score was 35-34 Noles), while A&M’s loss can be explained by the fact that Miami is really good, ranking 5th overall in our model.
Early model pick is Texas A&M 31, Arkansas 18, with the Aggies recording an 83% chance to win. We’ll have a full preview of the matchup later this week.
Other Games to Watch
Florida at Kentucky (11 am, ESPN). Graham Mertz is a very mediocre quarterback, but Florida’s defense is no joke. Kentucky has waltzed through a weak schedule, so this will be the first test for Devin Leary and a Kentucky offensive line that has had issues in the last few years. Spread: Kentucky -2.5, Model Pick: Florida 23-21
Georgia at Auburn (2:30 pm, CBS). Georgia has sleepwalked a bit, failing to cover a spread once this season. Auburn, meanwhile, has a stingy defense but a catastrophic offense. Payton Thorne was truly horrid against Texas A&M, and I don’t think Hugh Freeze has any other options. Georgia won’t need much from their offense to win, but are they going to cover a spread of more than two touchdowns? Spread: Georgia -14.5, Model Pick: Georgia 25-24
Missouri at Vanderbilt (3 pm, SEC Network). It really is starting to look like Mizzou has found something on offense. Brady Cook has been solid, and Luther Burden is just so good. The run game is working, too. Of course, that’s all come just in time for Mizzou’s defense to take a step back this season, falling to 72nd after finishing top-20 a year ago. Spread: Mizzou -14, Model Pick: Mizzou 30-26
LSU at Ole Miss (5 pm, ESPN). The early, unreleased version of our model last week thought that Ole Miss’s offense was overrated, and they got exposed by Alabama. The Rebel offense now sits at 71st in predicted offensive points per drive, and they are 78th in EPA per rush. Now they get to try and hang with Jayden Daniels and an elite LSU offense. Spread: LSU -2.5, Model Pick: LSU 27-21
South Carolina at Tennessee (6:30 pm, SEC Network). Here’s a fun matchup between two disappointing offenses. Maybe the Gamecocks fixed some things against Mississippi State… or maybe they were just playing Mississippi State. Carolina knocked the Vols out of the playoff with a big win in this game a year ago, so Vol players and fans really want this one. Spread: Tennessee -12.5, Model Pick: Tennessee 30-18
Alabama at Mississippi State (8 pm, ESPN). Alabama only ranks 25th in our model… but pity poor State, currently the SEC’s worst team at 115th. Zach Arnett moved away from Mike Leach’s preferred strategies on both sides of the ball, and so far, it’s not working at all. Spread: Alabama -14.5, Model Pick: Alabama 27-18
Thanks for reading! Be sure to follow us on Twitter and on Facebook.
The latest from Fayette Villains, straight to your inbox
Enter your email to subscribe and receive new post alerts and other updates. You can unsubscribe at any time.