Moving the (Stati)Sticks: Week Eleven

[Fuller/MGoBlog]

I don’t have an artsy intro this week because I spent a bunch of extra time making some charts, so instead I’ll refer you to BiSB’s intro to Opponent Watch, which is excellent.

The Mathlete’s Four Factors:

Again, a quick reminder of what the factors mean:

Field Position =  Expected team points based on starting field position. This accounts for all elements of field position: turnovers, special teams, drive penetration etc.

Conversion rate = [1st Downs gained]/[1st Down plays (including first play of drive)]. A three and out is 0/1. A one play touchdown is 1/1. Two first downs and then a stop is 2/3, etc.

Bonus Yards = [Yards gained beyond the first down line]/[Total plays from scrimmage]

This is an adjustment to how I have previously calculated, to account for the plays a team runs.

Red Zone: Points per red zone trip (TD’s counted as 7 regardless of PAT)

Offense:

Defense:

[Hit THE JUMP for 100% more charts]

*all numbers above are cumulative, not individual game data

There’s nothing in this week’s results that stand out to me as being really unexpected. Indiana’s average starting field position was the 33.5 yard line, so it’s no surprise that Mathlete’s defensive Field Position number has gone up again this week; they were able to move the ball well and had a shorter field to do so than Michigan’s typically been ceding this year. Defensive Conversion Rate took a huge jump for a single week, though that’s not a huge surprise; Indiana picked up 32 first downs. (What is surprising is that 20 of them were via the run.) Michigan held their own offensively, registering 28 first downs (20 via passing) and seeing a bump up, albeit a bit smaller than what they gave up defensively, in Conversion Rate.

Michigan edged Indiana in the battle of Bonus YPP, with their national defensive ranking actually going up three spots despite an aggregate increase of 0.03 yards per play. Compare that to Michigan’s offense, which jumped 10 spots in the national rankings thanks to an aggregate 0.11 yards per play increase. The explosive plays were finally there for Michigan; if there’s one image I’ll remember from this game, it’s two Indiana defenders taking each other out at midfield while Chesson saunters toward the left sideline and into the endzone.

Advanced Box Score

Indiana’s been tempo-ing people for a few years, so it’s not much of a surprise to see they ran 96 plays to Michigan’s 78. Michigan managed to win the yards-per-play battle, though, averaging 7.15 to Indiana’s 5.28.

Michigan was also able to stay on track more often than Indiana; their Success Rate was 45% while Indiana’s was 40%. Average starting field position was within a yard (Michigan’s average start was the 34.6, Indiana the 33.5), which heightened the importance of Success Rate and YPP. Scoring opportunities (keep in mind that in this box score these are counted as trips inside the opponent’s 40) were 10-7 in favor of Michigan, with M averaging 4.8 points per opportunity to Indiana’s 4.71.

Bill Connelly posted an expanded box score, so we’re able to parse things out a little more than we usually can for a single game. I made graphs of the three factors I thought were most interesting, but I highly recommend the whole post; there’s a win probability graph at the top that’s definitely worth your time.

Connelly’s advanced stats glossary is also worth your time and something I should probably be linking every week, as is the adv stats glossary at Football Outsiders. Connelly developed a system for assigning point values to every yard line; another way to think of it is how many points you can expect to score from having the ball at a certain place on the field. You can then assign a point value when the ball is moved based on those aforementioned points values that are tied to field position. Basically, football’s a weird game where there’s little to no certainty, but if we look at things in the aggregate it’s possible to estimate how many points should have been scored based on ball movement.

Michigan’s offense had a better day overall per EqPts, as they were expected to score 56.9 points to Indiana’s 48.9. Going back to the “football is just weird” trope, Indiana’s very good passing offense put up only 18.7 EqPts while their ground game posted 30.2 on Michigan’s previously stellar run D. Likewise, Michigan’s meh passing attack went nuts, registering 43.5 EqPts.

Success Rate, the stay-on-track stat, looks at how often a team got 50% of necessary yardage on first down, 70% on second down, and 100% on third and fourth down. As with EqPts, Michigan had the overall edge despite Indiana’s run game success. Michigan’s passing game was better than it’s been all year (with the possible exception of Rutgers, but I don’t have similar data for that game); you can see how far above the national average it was against Indiana.

It had to be, though, as Michigan’s success on standard downs was below both Indiana’s and the national average. Passing downs are considered 2nd-and-8+, 3rd-and-5+, and 4th-and-5+, and Michigan successfully gained necessary yardage on those downs an insane 60.9% of the time; the national average is 30.5%. That’s not a very sustainable way to have success on offense, but it makes for some fun gifs.

IsoPPP is Connelly’s explosiveness stat; it looks at how successful you were when you were moving the ball successfully i.e. when you’re staying in favorable down and distance situations as measured by Success Rate.

Generally speaking, this was a more explosive game than average. This is also where anomalies come out; we’ve established that Indiana had a good day running the ball and Michigan had a good day passing it and that Indiana was more successful on standard downs and Michigan on passing downs, yet Indiana was more explosive on passing downs and Michigan more explosive on standard downs. Basically, Indiana’s successful passing-down plays went for more yards than Michigan’s, but Michigan picked up the necessary yardage to move the chains more often on passing downs. It’s the inverse for standard downs.

Five Factors

Everything got a little bit worse for Michigan’s defense after playing Indiana. They dropped two spots in Explosiveness to 67th; their Success Rate against went from 29.9%, which was third overall, to 32.1%, which is seventh; points per trip inside the 40 rose from 3.24 to 3.70 and tumbled six spots to 11th; average starting field position dropped seven spots to 18th overall in going from the 26.2 to 26.8 yard line. The defense is now third overall in S&P+.

The silver lining came on offense, where Michigan’s Explosiveness again took a huge leap, jumping 20 spots to 28th overall after a 24 spot jump after the Rutgers game. Their Efficiency is basically the same, having dropped one spot to 39th. Finishing Drives (measured as points per trip inside the 40) increased as well, though just barely; they moved from 46th to 44th.

FEI really liked Michigan’s performance, showing gains in every category. The ones that stand out to me are in explosive drives, where Michigan went from 82nd to 46th, and methodical drives, where they went from 102nd to 82nd. The offense is also gaining more available yardage, going from 49.8% (36th overall) a week ago to 51.8% (32nd) after Indiana.

Opponent’s Run Game: Well, this got ugly in a hurry. Michigan’s rushing S&P+ dropped from third to 15th. Rushing Success Rate dropped from third to fifth, with the actual SR going from 27.3% to 31.7%. Michigan was ranked fifth in Opportunity Rate, which measures how often the O-line produces at least five yards for the rusher, but is now ranked seventh; they were allowing the O-line to do so on 28.3% of carries a week ago, and that’s now up to 30.9% thanks to a veteran (and very good) Indiana offensive line.

Michigan’s Run Game: Not a great day. Overall rushing S&P+ fell 17 spots to 53rd, while Success Rate fell 13 spots to 56th. They did somehow get a little more explosive, going from 77th to 69th in IsoPPP.

Opponent’s Pass Game: Overall passing S&P+ held serve at 13th, which is no small feat against an Indiana pass game that was one of the top 20 passing attacks in the nation. Passing IsoPPP did drop 11 spots to 36th, but that number has fluctuated a lot recently and I don’t put as much stock in worrying about that as I would if the overall value was falling.

Michigan’s Pass Game: Michigan’s faced two shaky secondaries in a row and laid waste to both. Before the Rutgers game they were 72nd in IsoPPP; they jumped 10 spots this week to 28th. Their overall passing S&P+ is now 29th; one of the many reasons I’m going to start putting these things into a table that allows me to track weekly progress is that I’m positive the graph for this would show huge leaps, and who doesn’t like graph that’s trending in the right direction.

Looking Forward, Looking Back:

What about Saturday?

Alum96 has a nice look at where Penn State ranks in the stats the NCAA keeps. What’s your takeaway, alum96?

PSU is bad at offense

What about defense?

As noted earlier PSU has a very good D - the one area to exploit is in the red zone which just so happens to be an area UM O excels.

Looking through their advanced stats profile, a couple of things stand out to me. They’re eighth in Explosiveness but 113th in Efficiency. When they’re successful they’re successful in a big way, but they aren’t successful very often. Perhaps that has something to do with an adjusted sack rate that’s 121st overall (124th on std downs, 106th on passing downs).

Their defense is certainly good. They’re 13th in S&P+ with a passing defense that’s fourth overall. Their run D is a step down (though still very good) at 37th; that might have something to do with a 112th-ranked rushing IsoPPP, which may stem from a defensive line that’s first overall in havoc rate. (Havoc rate is TFLs, passes defended, and fumbles forced/total plays.) Their D-line might occasionally vacate a run lane, but when they don’t your offense is in trouble.

You should also check out Ecky Pting’s work before the game. He’s got a nice, color-coded schedule rundown and a post on win probabilities.