VN: Stats, Eyeballs, and Underrated Alabama

VN: Stats, Eyeballs, and Underrated Alabama
VN: Stats, Eyeballs, and Underrated Alabama
Photo: USA Today Sports Images

by Bill Connelly

When I redesigned S&P+ this past offseason, I tried to look at basically every factor that goes into the ratings process through the course of a given season. I tried to more properly weight early-season ratings based on factors that were most likely to tie in to late-season rankings, and I tinkered quite a bit with how and when to phase out preseason projections.

The conservative process for phasing out those numbers -- generally, incrementally (and evenly) decreasing their weight for about seven weeks -- tends to work pretty well. But I wanted to see if there was a more aggressive way to go about this process. How quickly could I make in-season numbers stand on their own?

In the end, I settled on an approach that makes me both happy and a little bit scared. In simulating the 2005-14 seasons, I found that by quickly dialing projections down -- to 30 percent after three weeks and to 10 percent after four -- I was coming up with a system that was a bit more volatile in terms of week-to-week shifts but was, in terms of picks against the spread, about as accurate as the conservative approach. I don't know if it's better, but it doesn't appear worse. And if you've got two approaches that seem to produce results of equal quality, go with the one that better fits your philosophy.

So here we are. From week 3 to week 4 of the 2015 season, West Virginia jumped to No. 3 in S&P+, Michigan jumped to No. 10, Duke jumped to No. 19, Illinois jumped to No. 34... and on the flip side, Arizona State fell to No. 48, Oregon fell to No. 52, Auburn fell to No. 58, and Missouri fell to No. 79.

Even with conservative use of projections, S&P+ has always been pretty volatile, full of both prescient views of certain teams (catching on that they're good before other systems) and some false flags. That's probably going to be the case even more over these next few weeks, when the half-naked 2015 data is less covered by projections.

With that in mind, let's take a look at a few of the teams S&P+ disagrees about most in comparison to conventional wisdom (i.e., polls). Which ones are most likely false alarms, and on which teams are the numbers most likely right?

AP likes, S&P+ dislikes

Oklahoma State (difference: 37 spots)

AP Rank: 20
S&P+ Rank: 57

Why S&P+ Dislikes: OSU has beaten Central Michigan, Central Arkansas, UTSA, and (with help from a couple of awful calls and a heinous special teams miscue) Texas. None are ranked higher than 61st, and both the CMU and Texas games were a little closer than many anticipated. They have moved up to 20th in the polls simply because polls are a giant ladder. Win, and you move up. Lose, and you move down.

OSU looked truly awesome in disposing of UTSA but has otherwise just gotten by. They're awful at running the football and are giving up a few more big plays on defense than they should, given the competition level. I could feel comfortable with thinking of them as a top-40 or top-50 team, and they're young enough to keep improving as the year progresses, but they have in no way established top-20 credentials yet.

I side with... S&P+.

Northwestern (22 spots)

AP Rank: 16
S&P+ Rank: 38

Why S&P+ Dislikes: The Wildcats' defense has been one of the most surprising units in college football thus far, allowing just 3.9 yards per play to Stanford in a 16-6 win and giving up just 35 points in four games. But you have to play offense, too, and the only team Northwestern has scored more than 24 points on is Eastern Illinois. They currently rank 10th in Def. S&P+... and 105th in Off. S&P+.

NU's 24-19 win over Ball State was a little bit scary in that it was their worst defensive performance of the year, and the offense could barely do enough to keep the Wildcats ahead. This should end up a nice bounceback season for the Wildcats, something in the neighborhood of 7-5 or 8-4, but they're far more of a top-40 team than a top-16 team.

I side with... S&P+.

Utah (19 spots)

AP Rank: 10
S&P+ Rank: 29

Why S&P+ Dislikes: I think S&P+ dislikes Utah more than any other computer rating in the universe, and the major reason for the low rating is because, for as good as the Utes looked in mauling Oregon in Eugene last weekend, that was the first time all year that Utah out gained its opponent on a per-play basis. Michigan gained 4.9 yards per play to Utah's 4.8. Utah State gained 5.6 to Utah's 5.0. Miserable Fresno State gained 5.7 to Utah's 5.1.

Special teams and timely play have made a world of difference for the Utes, who certainly proved in Eugene that their ceiling is high. But while per-play yardage isn't a direct factor in the S&P+ ratings, that's a pretty big harbinger.

Still... 29th! And they've beaten Michigan and Oregon! That feels low to me.

I side with... AP. Utah probably isn't 10th, but it's closer to that than 29th.

Mississippi State (16 spots)

AP Rank: 21
S&P+ Rank: 37

Why S&P+ Dislikes: The Bulldogs can't run the ball and haven't been able to stop other teams from establishing the run to a degree. They rank 98th in rushing success rate allowed (unadjusted for opponent) and 105th in offensive Rushing S&P+ (which is adjusted). Yikes. And the passing game, while pretty efficient, hasn't produced enough big plays to make up the difference.

Still, the Bulldogs proved themselves resilient in nearly coming back to beat LSU, and they have what a lot of other foes in the SEC do not: a proven quarterback. That could make the difference in quite a few games (i.e., any game not against Alabama, Ole Miss, or maybe Texas A&M).

I side with... AP.

[ad placeholder 3]

Baylor (13 spots)

AP Rank: 5
S&P+ Rank: 18

Why S&P+ Dislikes: Opponent adjustments have been quite cruel to Baylor's defense, which is allowing only 23 points per game (despite high tempo) but ranks just 94th in Def. S&P+. The Bears are first in Off. S&P+, so if the defensive ratings come around, they'll be fine. But 94th feels off to me considering they allowed just 3.7 yards per play to Rice and 4.7 to SMU. I think the D is fine... not great, but fine. The opponent adjustments are being pretty harsh based on one iffy half against Lamar.

I side with... AP.

S&P+ likes, AP dislikes

West Virginia (difference: 20 spots)

AP Rank: 23
S&P+ Rank: 3

Why S&P+ Likes: Third! West Virginia! The Mountaineers have quickly moved from a team with top-40 potential to, if they win in Norman tomorrow, projected Big 12 champions. The Mountaineers are as high as they are because they're one of the only teams playing excellent offense (14th in Off. S&P+) and defense (ninth in Def. S&P+) at the moment. Only four teams are in the top-16 in both categories, and they happen to be the top four in the rankings.

This defense is loaded for bear, but I worry about the offense. The run game isn't producing many big plays, and the passing game is leaning heavily on a freshman (Jovon Durante) and a sophomore (Shelton Gibson). It will probably regress at least a little bit. Plus, while the defense is awesome, it's benefiting a bit from the fact that WVU played Georgia Southern when the Eagles were without their starting quarterback.

I side with... AP.

Minnesota (20 spots)

AP Rank: 40
S&P+ Rank: 20

Why S&P+ Likes: Defense. The Gophers are playing D at a level similar to Northwestern, and while the offense hasn't been very effective (like their conference rivals from Evanston), they benefit from playing against Kent State, which has had one of the most strangely dominant defenses in the country so far -- the Golden Flashes have allowed just 3.4 yards per play this year, but allowed 4.1 to Minnesota.

Of course... Kent State's record has been bolstered by its utter domination of lowly FCS foe Delaware State (38 plays, minus-33 yards). The Golden Flashes clearly have a good mid-major D, but it's probably getting too much of a bump for that performance, and consequently, Minnesota and Illinois (5.3 yards per play) are probably getting an artificial bump of their own.

I side with... AP.

[ad placeholder 3]

Tennessee (19 spots)

AP Rank: 42
S&P+ Rank: 23

Why S&P+ Likes: Because S&P+ doesn't explicitly care that the Vols figured out ways to blow double-digit leads against both Oklahoma (eighth in S&P+) and Florida (21st) and lose at the wire. S&P+ only cares that, for all intents and purposes, they fought those two teams to a virtual draw. This is a good team, albeit one that has no idea how to finish games (and isn't getting particularly good guidance from its coach).

I side with... S&P+.

Duke (14 spots)

AP Rank: 33
S&P+ Rank: 19

Why S&P+ Likes: Like Northwestern and Minnesota, the Blue Devils are carried by defense. They rank a rather shocking seventh in Def. S&P+ at the moment, having allowed 46 points in four games; the next time they allow greater than 4.2 yards per play in a game will be the first. (And here's where I mention that they just held Georgia Tech, last year's No. 3 offense according to Off. S&P+, to 3.9 yards per play.)

This defense appears legitimate. But like Northwestern, the Blue Devils are struggling on offense. Averaging 7.7 yards per play against N.C. Central is fine, but they managed just 4.5 yards per play against Northwestern and Georgia Tech. Junior quarterback Thomas Sirk's passer rating against power conference opponents is thus far under 100.0. That's foreboding.

I side with... AP.

Alabama (12 spots)

AP Rank: 13
S&P+ Rank: 1

Why S&P+ Likes:Alabama is underrat--

Alabama is underrat--

Alabama is underrat--

Alabama is u n d e r r a t e d. Whew. There. My computer crashed every time I tried to type that properly. That's not a sentence it recognizes.

Alabama lost to a top-three caliber team because of a minus-5 turnover differential that included bad turnovers luck. With only a minus-4 differential, the Tide maybe win. With a minus-3, they almost definitely win. They're very good. They might have a second loss after this weekend's trip to Georgia, but that has more to do with Georgia than Alabama.

I side with... S&P+. If the Tide aren't the best team in the country, they're still a lot closer to No. 1 than No. 13.


1 comment, Last at 04 Oct 2015, 1:25am

#1 by Tomlin_Is_Infallible // Oct 04, 2015 - 1:25am

Alabama lost to a team that lost to a team that.....................


The standard is the standard!

Points: 0

Save 10%
& Support Bill
Support Football Outsiders' independent media and Bill Connelly. Use promo code WRITERS to save 10% on any FO+ membership and give half the cost of your membership to tip Bill.