clock menu more-arrow no yes mobile

Filed under:

Fun with Numbers: Week 1

Stats are for losers. In other words, they're right up my alley.

A badly underused Fitz picture
A badly underused Fitz picture

Northwestern's game against Cal produced some interesting drive stats:

  1. Cal and Northwestern each had 16 possessions (that's a lot).
  2. Northwestern ran 0 offensive plays on 4 of those possessions: the first half expired during a kickoff return, NU fumbled another kickoff return, and Collin Ellis scored twice on interception returns.
  3. Northwestern's offense went three and out twice. Both of Trevor Siemian's interceptions came on or before the third play of a drive, though.
  4. Cal's offense went three and out three times; two of their interceptions came on the third play of a drive or sooner.
  5. On drives where the NU offense saw the field, they gained an average of 42.3 yards.
  6. Cal gained an average of 34.25 yards per drive.
Those last numbers are a big deal. While both offenses did well, the Northwestern offense was much more efficient on both a per drive and a per play basis (7.3 YPP to 5.5). Between kicking off as time expired in the first half and Northwestern fumbling a kickoff, Cal effectively ended up with a 2 drive advantage in the game. Combined with a slight field position advantage (Cal's average starting field position was their own 27, to Northwestern's 23) and a spectacular goal line stand in the third quarter, this allowed Cal to match the NU offense point for point. I've been gloomy about defensive scores, a fickle source of points, providing NU's margin of victory, but this game was different from last season's opener, in which NU needed a defensive score, a special teams score, and two extremely short fields to pull out a 1-point victory. Northwestern was a bit more efficient in the offense vs. defense part of the game; the pick-sixes turned that modest edge into a solid margin of victory.

Stat Tracking

I'm interested in tracking a few numbers about NU's team performance that are useful but hard to find: sack rate (sacks/(sacks+pass attempts)), pass defense rate (passes defensed/pass attempts), interception rate (interceptions/passes defensed), and fumble recovery rate (fumbles recovered/total fumbles). I will also track Northwestern's Adjusted Turnover Margin, calculated by assuming that both NU and opponents recover 50% of fumbles and intercept 22% of passes defensed. All data is drawn from

Sack Rate: 5.9%; that's a slight improvement over last season's 5.3%.
Pass defense rate: 15.6%; last season this was 15.3%.
Interception rate: 30% of passes defensed; last season was 17.1%.
Fumble recovery rate: 0%.
Opponent pass defense rate: 16.7%; last year was 15.7%.
Opponent interception rate: 40%; last year was 12%.
Opponent sack rate: 3.2%; last year was 4.2%.
Adj Turnover Margin: 1.1
Difference from actual turnover margin: -1.1

Overall, NU was a bit better both at getting to the quarterback and playing balls in the air than last season. Both of those differences were pretty small, though. The improved offensive sack rate would be interesting if it continued; NU's offense last year leaned more heavily on quick throws, which kept sack rate down by design. Finally, NU was a bit unlucky with turnovers; while both teams picked off more passes than expected, Cal was a bit luckier in this regard and significantly luckier with fumbles, recovering all 4 (though NU only forced 1 of Cal's 3 fumbles, so it's hard to complain).


Last year, the intermediate passing game that was Dan Persa's specialty virtually disappeared. This resulted in extremely low yards per reception; Christian Jones led players with more than one catch with only 11.8 yards per reception. After one week, four Northwestern receivers are ahead in that stat: Dan Vitale (20.2), Christian Jones (18.8), Tony Jones (14.0), and Mike Jensen (12.0 on only one catch). The more Siemian plays, the more likely the receivers are to keep these numbers up. Collin Ellis defensed a total of 5 passes (2 INTs and 3 PBUs), which makes him the early national leader in that stat.

Advanced Stats and Computer Rankings

Football Outsiders hosts three statistical ranking systems; FEI is based around drive data, S&P+ considers individual plays, and F/+ combines the other two measures. For the past few years, FEI has been significantly more positive about Northwestern than S&P+, a trend that showed up in their preseason projections: FEI projected NU as the 26th team in the country, while S&P+ projected NU as the 49th team. After week one, NU moved up in both, from 26th to 25th in FEI and from 49th to 46th in S&P+. In spite of moving up in both of its components, NU dropped two spots, from 35 to 37, in F/+. Cal dropped from 66 to 71, while Syracuse moved from 55 to 53.

In Jeff Sagarin's rankings, NU ranks 34th in the BCS-component ELO ranking (which considers only wins and losses) and 37th in the more accurate Predictor (which considers only points scored and allowed). Sagarin's predictor ranking pegs NU as about a 7.5 point favorite over Syracuse on a neutral field; that should be about a 70% win probability.

Finally, OTE commenter hoegher is willing to let us see the madness of rankings that minimize the importance of preseason projections. Early results say Washington is the best team in the country; that wouldn't be a bad world. NU is 41st nationally, third in the Big Ten, in his rankings.

If you have a stat or ranking system you want me to track, let me know in the comments.