Since I frequently procrastinate by checking stat sites, I might as well share the wisdom I glean with the world, right? This week, Northwestern manages a big leap, Nebraska a bigger one, and I have horribly irrational hope for the weekend.
There are three stat systems I will be looking at here: the Football Outsiders FEI and S&P+ ratings (as well as the composite F+ ratings) and Chase Stuart's SRS. FEI is a drive-based stat, S&P+ is play-based, and the SRS is score-based (similar to Sagarin's predictor ratings).
Northwestern's performance against Indiana managed to impress the FO measures: NU jumped 11 spots from 73 to 62 in S&P+ and 9 spots from 63 to 54 in FEI. The NU offense is in the top 30 of both; this is a testament to how strong the offense has been in conference play, as the BC and Army games are serious anchors on these numbers. The defense is mildly less impressive: 99th in S&P+ and 89th in FEI. Still, um, hooray for a top-100 defense?
Now for the more impressive change: Nebraska's thrashing of Michigan State led to a rise from 26 to 19 in S&P+ and an unbelievable rocket from 37 to 11 in FEI. The really interesting thing is that the two stats disagree about which side of the ball is Nebraska's strength: FEI sees them as a top-20 offense with an above average defense, while S&P+ sees them as a top-20 defense with an above average offense. I'm not sure what to make of this inversion.
The SRS has Nebraska at 13th with an SRS of 54.8 and Northwestern 67th with a 37.9; factoring in the 3-point home field advantage that Stuart uses, this translates into an expected 19.9 point Nebraska win. This is slightly higher than the lines that I have seen, though it seems fair to give NU a couple of points because, again, the BC and Army games are probably a less accurate indication of the way this team is playing right now than the Big Ten games.
What does it all mean?
It means that Nebraska is correctly a heavy favorite this week. Northwestern has improved tremendously on offense since Persa's return, but the defense is still the same. Nebraska, meanwhile, just played their best game of the year by any measure. Northwestern is an interesting team, recalling the 2005 team's dominance on offense and incompetence on defense, but Nebraska is a genuinely good team.
Still, some small part of me derives hope from this exercise. Nebraska looked less than impressive before last week, allowing some bad teams to hang with them and a very good Wisconsin team to thrash them. Furthermore, this gives further support to the idea that Northwestern's offense is actually good and last week's scoring explosion wasn't just an artifact of playing Indiana. If NU can match Wisconsin's output again (48 against Nebraska), they have a legitimate chance. I have no idea why the over-under for this game is only 64; Nebraska could score that on their own, and there is little reason to think that this will be the week that someone shuts down the Northwestern offense.
One last note on something I wrote earlier about the S&P+ covariance measure: after I wrote that, I realized that it was even less meaningful than I thought at the time. Basically what the off the charts covariance was saying is "Persa missed the three worst opponents to date" in a fancy statistical way. Had NU replaced any of the first three opponents with Michigan or Penn State, I am willing to bet that NU's covariance would have been more normal. Instead, what S&P+ saw was massive offensive improvement (with possible but lesser defensive regression) against much improved opponents, but the causal factor is most likely not a general tendency to play up and down to opponents so much as the return of a star player before starting a significantly harder slate of games.