yards per play
In a past diary, I charted Michigan's offensive trends over the course of the 2011-2013 seasons. In that diary, I relied on a statistic that I will call YPPdiv. This is equal to the offensive yards per play (YPP) in a given game, divided by the average YPP allowed by that team on defense. Thus, for example, Michigan averaged 7.35 YPP against Ohio State, a team that allows on average 5.01 YPP. So, 7.35 divided by 5.01 = a YPPdiv of 1.47.
Space Coyote suggested that I track the entire B1G over the course of the 2013 season to get a trendline for the entire B1G against which to compare Michigan (and each team's) offensive performance over time. Now that the regular season is over, it seemed like a good time to put the numbers together. Without further ado, here are the charts (click to embiggen):
The story: Honestly I don't know. I think that Illinois might just be bad.
The story: Indiana ripped it up early in the year but didn't have much success against Wisconsin and Ohio State. I'm not sure why, and the Crimson Quarry blog didn't seem to have any good answers, either.
The story: Iowa was pretty Iowa all year.
The story: problems with the offensive line, Devin Gardner being beaten up, complaints about the coaching. The Mathlete has proclaimed that this is the offense with the 4th highest standard deviation since 2003 (I'm not sure what stat he's using to calculate this; presumably it's PAN). But calculating the standard deviation of YPPdiv also yields a very high number for Michigan (0.33) compared to the rest of the B1G (an average of 0.21, with a high of 0.28).
The story: Michigan State found an offense this year. Their first three games were awful, but they've shown clear improvement since then. Of course, they haven't played a really good defense, either.
The story: Minnesota is Minnesota. They like to run the ball.
The story: despite the many injuries to key offensive contributors, Nebraska managed to basically hold serve offensively. They ripped up MSU in yards per play, but they turned the ball over far too much.
The story: injuries to key offensive players, including Venric Mark, Kain Colter, and Tony Jones. There have also been complaints of conservative play calling.
The story: Ohio State's offense has improved over the course of the season. Between Miller and Hyde, the Ohio State offense has pounded away at their opposition, none of whom, it should be said, has a particularly great defense.
The story: Penn State played great against Wisconsin but has been sub-par most of the year on offense.
The story: Purdue is bad, and was mostly bad throughout the year. They're consistent, at least.
The story: Wisconsin's running game is fantastic, but Stave has been inconsistent.
So. The average offensive performance by B1G teams, week by week, is as follows:
Perhaps unsurprisingly, it's fairly flat, with consistent performances from one week to the next.
This suggests that an offense with a slightly negative slope isn't to be expected; we should perhaps expect a fairly consistent performance from week to week. This is something that Nebraska managed despite many injuries on the offensive side of the ball. The best teams in the B1G this year -- Ohio State and Michigan State -- both showed improvement on offense over the course of the year.
The Ohio State game taught us that Michigan's problem isn't a steady decline in offensive performance, but rather wild inconsistency due to youth and playcalling.
In my last diary, in which I tried to chart Michigan's offensive regression over the season, Gandalf the Maize suggested that I track Michigan's offensive performance over the course of the past three seasons. That seemed like a good idea.
For this diary, I only used one statistic: Michigan's offensive yards per play (YPP), which I then divided by the average yards per play allowed by that team over the course of that season. So, for example, Michigan averaged 7.38 yards per play on offense against Western Michigan in 2011, and 6.63 YPP against Ohio State. But Western allowed 6.1 YPP on defense over the 2011 season, whereas Ohio State allowed only 5.06 YPP. That means that our offensive output against Ohio State was more impressive (131% of their average YPP allowed) than our output against Western (121%).
Note: the black diagonal line is the trendline. Maize dots indicate losses.
Here are the charts:
And, all three seasons in one chart:
There is offensive regression in each season. As Gandalf pointed out, this is sub-optimal.
- It is most pronounced this season, of course.
- In 2012, it is the least pronounced; I took out the Nebraska game and the trendline was still slightly negative (y = -0.0063x + 1.205, R² = 0.011).
- The 2012 Iowa game was Michigan's best offensive output over the past three years by this metric. For those who forgot, that was the game where Devin played QB and Denard returned as a RB. Devin was 18/23, 314 yds (13.65 yards per attempt), and Denard rushed for 98 yards on 13 rushes for an average of 7.5 yards per rush.
- The 2012 Nebraska game was, unsurprising, the worst offensive output over the past 3 years. I don't think anyone needs to be reminded about that game.
I'm not sure how to diagnose this overall trend. Borges running out of ideas? Our quarterbacks getting banged up? Cold weather? On the other hand, perhaps it's not terribly significant -- the slopes for 2011 and 2012 are only slightly negative, after all.