Week 10 EPA Preview

Like last week, in this space I'll take a look at what I subjectively decide are the 5 most interesting games of the weekend.  Specifically, I'll look at what EPA tells us to expect.

Texas A&M vs. Mississippi State (12:00 ET, ESPN)

CFBTN projection: 36-29, Aggies

Quarterbacks: Johnny Manziel (94.67 EPA, .234 EPApp) vs. Tyler Russell (37.31 EPA, .142 EPApp)

Running Backs: Christine Michael (7.51 EPA, .127 EPApp) & Ben Malena (3.37 EPA, .038 EPApp) vs. LaDarius Perkins (9.48 EPA, .057 EPApp) & Josh Robinson (-1.07 EPA, -.036 EPApp)

On offense, advantage A&M.  While Perkins has been the most productive of the backs on the field, Michael has been the most effective.  Malena's not far behind, and that doesn't even include freshman Trey Williams, who could factor in as well.  At QB, there's no comparison; I'm not suggesting Tyler Russell is that bad - far from it - but Johnny Football is that good.  If MSU expects to win, it's not going to be by out-offense-ing the Aggies.

Defensively, Mississippi State has allowed -32.53 EPA against the pass (-.121 pp) and -7.85 EPA against the run (-.047 pp).  These good stats are largely a function of the schedule they've played.  Against QBs in our top 50 in EPA, MSU allowed 1.39 EPA to Troy's Corey Robinson, -2.72 to MTSU's Logan Kilgore, 0.21 to Tennessee's Tyler Bray, and 2.13 to Bama's AJ McCarron.  Good QBs have been mostly average against Mississippi State this year.  I'd expect a great QB like Manziel to be about 5-8 points above average on Saturday.  Then again, when your EPA is over 90 at this point, I wouldn't limit his possibilities.  Against RBs in our top 50, they gave up around 3 EPA to Eddie Lacy and TJ Yeldon, and haven't faced any others. So, MSU remains somewhat untested on defense - I'd expect Michael & Malena to have a similarly productive day.

The Aggie D has allowed -40.78 EPA against the pass (-.122 pp) and -10.83 EPA against the run (-.060 pp).  Against the pass, the Aggies have been inconsistent, and the stats are buoyed a bit by their dominance over SMU's Garrett Gilbert.  He's been awful a lot.  However, our #35 ranked QB, Tyler Wilson, had his worst game of the season against the Aggies.  But Louisiana Tech's Colby Cameron had a great game against A&M.  Ole Miss' Bo Wallace wasn't bad either.  Like I said, inconsistent.  No team has had a lot of success rushing against A&M, but that's not to say they didn't have any success.  You can run on the Aggies - the problem is that with Johnny Manziel directing the other offense, you might not have the luxury of rushing late in the game.

All in all, like CFBTN's statistical projection, I'm thinking the Aggies win this one.  However, I don't necessarily think it'll be as close.  Bama exposed the Bulldogs last week for what they were - a solid middle of the pack SEC team that made a name by whipping inferior opponents.  Perhaps the depression of having to travel to Starkville could derail A&M just enough to keep it really close, but I think A&M wins by double digits.  They've got the better offense, and though it tends to be inconsistent, the better defense as well.  

Oregon vs. USC (7:00, FOX)

CFBTN projection: 36-30, Ducks

Quarterbacks: Marcus Mariota (54.32 EPA, .209 EPApp) vs. Matt Barkley (68.97 EPA, .228 EPApp)

Running Backs: De'Anthony Thomas (39.55 EPA, .412 EPApp) & Kenjon Barner (39.32 EPA, .238 EPApp) vs. Silas Redd (-15.28 EPA, -.117 EPApp)

54, 69, 40, 39, and -15.

Which of these is not like the other?  If you can answer that, you can figure out why one of the teams has been an offensive juggernaut and the other has sputtered at times.

The Duck D has allowed -60.66 EPA to passers (-.160 pp) and -12.64 EPA rushing (-.059 pp).  They held Fresno State's Derek Carr to his worst game of the year.  Ditto Arizona's Matt Scott.  Ditto Taylor Kelly, #11 in our rankings.  Had Kelly never faced Oregon, he'd rank #7.  They're good.

USC's D has allowed -60.06 EPA to passers (-.145 pp) and -15.34 EPA rushing (-.071 pp).  The defense looks every bit as stout as Oregon, but I think USC's quality of opponents isn't quite up to par.  Both teams have faced Colorado, so that's a wash, but USC has also faced Hawaii, Cal, and Utah (with Jon Hays).  None of these teams are particularly good at passing, and those 4 opponents make up -67.88 EPA against USC.  You read that right - USC is actually below average defensively against passers when not facing those 4 pretty bad passing attacks.  This will bode well for the Ducks.  And I'd look at the RBs USC has faced, but there's no point - no team can boast running backs like Oregon.  No team can truly be prepared for that.

Once again, I agree with CFBTN's projected winner, but I think the margin will be wider.

Alabama vs. LSU (8:00, CBS)

CFBTN projection: 30-15, Tide

Quarterbacks: AJ McCarron (40.28 EPA, .190 EPApp) vs. Zach Mettenberger (-17.46 EPA, -.075 EPApp)

Running Backs: TJ Yeldon (27.21 EPA, .267 EPApp) & Eddie Lacy (15.65 EPA, .125 EPApp) vs. Kenny Hilliard (6.69 EPA, .090 EPApp) & Spencer Ware (-1.72 EPA, -.050 EPApp)

Offensively, it's not close.  Bama's offense > LSU's.

Bama's D has allowed -82.07 EPA passing (-.312 pp) and -34.10 EPA rushing (-.202 pp).

LSU's D has allowed -86.08 EPA passing (-.299 pp) and -30.52 EPA rushing (-.182 pp).

We know what LSU's hope is - another game like last year's memorable regular season matchup.  They've got the defense to do it, too.  If they can hold Bama to a low-scoring game, they could potentially overcome their distinct offensive disadvantage.  Think of last weekend's Georgia-Florida game.

I think Bama wins, but I'm expecting a closer game than CFBTN's projection.

Kansas St vs. Oklahoma St (8:00, ABC)

CFBTN projection: 45-31, Wildcats

Quarterbacks: Collin Klein (109.19 EPA, .369 EPApp) vs. Wes Lunt (22.20 EPA, .196 EPApp)

Running backs: John Hubert (8.22 EPA, .058 EPApp) vs. Joseph Randle (-2.60 EPA, -.014 EPApp)

Earlier this week Scott Albrecht detailed why Kansas State is the best offense in the country this year.  OSU is pretty good, as usual, but it's not close.

K-State's defense is good, too.  Totaling the EPA allowed can't illustrate how good as well as this: they've held Seth Doege and Geno Smith to below-average performances.  Not below their average, but below THE average.  That's pretty remarkable.  OSU should be able to handle Hubert, but they've struggled against QBs a couple of times this year - Matt Scott and David Ash both had good games against the Cowboys. 

If this is the ABC game of the week, this is just a subpar week for college football.  My pick: K-State by multiple TDs.

Georgia vs. Ole Miss (3:30, CBS)

CFBTN projection: 40-28, Dawgs

Quarterbacks: Aaron Murray (38.92 EPA, .147 EPApp) vs. Bo Wallace (22.44 EPA, .074 EPApp)

Running backs: Todd Gurley (18.71 EPA, .127 EPApp) & Keith Marshall (6.65 EPA, .050 EPApp) vs. Jeff Scott (8.52 EPA, .061 EPApp) & Randall Mackey (2.87 EPA, .038 EPApp)

Georgia has allowed -29.26 EPA against passers (-.091 pp) and -27.45 EPA against rushers (-.116 pp).  For most good teams, the passing EPA allowed is much, much higher than the rushing.  They have big leads, other teams get into predictable passing downs, and they can stifle the riskier half of the offense.  Georgia doesn't enjoy such an advantage.  For starters, UGA's stats include half a season without two starting defensive backs, and a key game against South Carolina with an injured Jarvis Jones unable to do much of anything despite being on the field.  The D has been good.  They just haven't been as great as expected.  And they haven't been consistent.  However, in the week leading up to the Florida game, S Shawn Williams called out his defensive teammates, and much was made over how much more cohesive the unit would be.  Normally I take these things with a grain of salt - it's akin to the 'best shape of my life' comments you hear from MLB players every March - but after the ridiculously dominant performance against Florida on Saturday, I'm a believer in this defense again.

Ole Miss has allowed -10.92 EPA against passers (-.038pp) and -2.91 EPA against rushers (-.015pp).  They dominated Tulane but got dominated by David Ash of Texas.  AJ McCarron and Tyler Wilson both had subpar days against the Rebels.  Really, no one has been great agaisnt them except Ash.  But he was really great.  What does that mean for Murray?  I'm not sure.  My bet is he falls in around McCarron and Wilson, both of whom, while technically subpar, were close to average against the Rebs. 

I could see this one being a lower scoring game than CFBTN projects, but I do think Georgia wins.  The spread is 14, and Georgia never covers (it's not the Christian thing to do), so I'm thinking something like 30-17 for the final.  Scott's projection will be closer if last week's defense didn't make the trip home from Jacksonville.



Brent Blackwell compiles the EPA rankings for cfbtn.com.  Follow Brent on Twitter by mashing the pretty button below.

Follow @brentblackwell