clock menu more-arrow no yes

Filed under:

Tuesday Morning Brews: Lightning crashes

Using stats to diagnose Michigan’s offensive struggles

NCAA Football: Michigan State at Michigan
Under the Lights IV
Rick Osentoski-USA TODAY Sports

Happy Tuesday, folks. Welcome to Morning Brews. Well, you all watched on Saturday night so I’m not going to rehash the results this morning. There has been plenty of time now for disappointment, anger, and other emotional reactions, but all the recriminations in the world won’t change the fact that a loss is a loss and there is a game against Indiana on Saturday. In that spirit, this morning I’m taking a look at the offense’s woes and what I think is causing them. I will also say that this is a much longer and more stat heavy Brews than normal, so don’t say I didn’t warn you.

Without further ado, let’s get to it:

NCAA Football: Michigan State at Michigan Mike Carter-USA TODAY Sports

Ghosts in the machine, part two

Those of you who read Morning Brews regularly may remember that three weeks ago, I put forth my conclusion that the offense’s difficulties didn’t rest squarely on the shoulders of Wilton Speight and that there were larger issues at work. Specifically, I closed that article by saying, “what I do know is that there are going to be close games this year. And if Michigan struggles to finish in the fourth quarter, there will be close losses as well.”

Unfortunately, instead of refuting my conclusion Saturday’s game provided another data point in the broader picture. This morning I’m going to utilize three graphs to illustrate the deficiencies of the offense so that readers and fans can get a better handle on where exactly the Wolverines are coming up short.

I need my charts, Jack!

This first picture (click and zoom if you must, graphs tend to not render well on the platform) is a snapshot of offensive production. The Y axis is numbered 0-100, and the X axis is numbered one through four for the quarters of a game. The blue line shows yards per quarter from Saturday’s game, the orange line shows completion percentage by quarter for the season, and the gray line shows average rush yards per quarter for the season. There are four things I want to highlight here.

First, Michigan’s offense struggled in the second and fourth quarters relative to the first and third quarters. Second, the inverse relationship between offensive production and what quarter it is that I pointed out three weeks ago remains intact (less production later in the game). Third, while stable when compared to overall production and rushing production, completion percentage reliably decreases as the game goes on. Fourth, rushing production falls off a cliff in the second quarter.

To illustrate those last two points more, let’s look at our second graphic.

This is a breakdown of play calls by type (run or pass) by quarter for the season thus far. The Y axis is numbered 0-60 and represents the total number of plays called, and the X axis again represents the quarters of a game. The blue line represents rushing plays and the orange line represents passing plays. What I think is most interesting here is that the crater in second quarter run production exhibited in the first graph is not a product of a reduction in run play calls in the second quarter. In fact, the second quarter has seen the second most rushes called. It’s not that the offense is getting away from rushing in the second quarter, it’s just that the yards per carry drops to a shocking 0.93.

My other observation here is how play calling changes throughout the game, and specifically the convergence of rush and pass plays in the third quarter and the divergence in the fourth quarter. Both of these are interesting because they coincide with Michigan’s most productive quarters on offense. The Wolverines have scored 87 of their 136 points in the second half, 44 in the third quarter and 43 in the fourth quarter, and seven of their 11 offensive TDs have come in the third and fourth quarters (four in the third, and three in the fourth). I suspect these play calling differences are a result of shifts in strategy, i.e. throwing to retake or extend the lead in the third quarter and, before Saturday, running to kill clock in the fourth quarter.

This final graph is representative of offensive production per quarter. The blue line shows yards gained per quarter on Saturday, the Orange line shows average yards gained per quarter across the first five games this season, the Yellow line shows average yards gained passing, and the gray line shows average yards gained rushing (yellow plus gray equals orange).

My sole observation from this graph is that Saturday was a lot like the offensive production for the season so far. The slight downward divergence in the first and second quarters from the average to Saturday amount to roughly one or two less successful plays and so aren’t really that notable. The more significant divergences in third and fourth quarter production could be attributed to the weather (MSU only gained 64 yards in the second half), but - again - the trend of offensive production remains.

So, what do the pictures mean?

Now that I’ve illustrated the problem, let me try to put it into words while augmenting the graphics above with some additional stats. First, ignoring the variations, offensive production is too low. Michigan is averaging 5.43 yards per play, or 385.8 yards of total offense per game, this year. Last year it was 5.85 yards per play, or 460.3 yards of total offense per game. Considering that the early part of the season is when such numbers are usually inflated before conference play begins, a large deficiency at this point is very concerning. This decrease in production in contributing to Michigan scoring nearly two fewer TDs per game this year than last year (27.2 ppg vs 40.3 ppg).

Second, the variations in offensive production are killer. Production dropping off in the second and fourth quarters is a recipe for frustrating games where the Wolverines just need a little bit to win, but don’t get it (see, last year against Iowa and OSU and this year against MSU). The fourth quarter is when you want your offense to peak in close games. Instead, it’s the second worst quarter for Michigan’s offense.

Third, there’s something weird (read: bad) going on with the rushing attack during the second quarter. Going from an average yards per carry figure of 4.5-5.5 in quarters one, three, and four to a 0.93 in quarter two is absolutely baffling. The dip in production is so significant and so sustained that it’s dragging down the overall yards per carry figure.

Fourth, overall running production is down. Despite being a roughly 60% running team, yards per carry and yards per game have dropped from last year (3.98 vs 4.82 and 167.8 vs 212.9). Again, considering that teams like Michigan usually beat up on defenses in the early part of the season, this sort of divergence is troubling.

Conclusions

My conclusion is that Michigan’s offensive woes stem from two sources: first, play calling; and, second, the offensive line. Aside from a few obvious examples (like a five-wide formation on 3rd and three in the driving rain), play calling is hard to evaluate. It’s a subjective art and the “correct” call almost always depends on a person’s individual offensive mindset, tolerance for risk, and a healthy dose of 20/20 hindsight. Play calling must also be informed by the strengths of your team, and, well, those are a little hard to identify for Michigan’s offense right now. With that being said, it’s the job of the coaches to identify their strengths and call plays appropriately - and that hasn’t really happened yet.

The second problem, offensive line play, is easier to quantify. The first indication of poor line play is the marked decrease in rushing production I mentioned above (a loss of 0.84 ypc year over year). This is due in large part to the horrific 0.93 ypc that rushing plays are averaging in the second quarter, which is likely produced by a combination of poor line play, subpar play calling, and some bad luck. Aside from the rushing woes in certain circumstances, sacks allowed are up - WAY up. Through the first five games, Michigan’s QBs have been sacked 16 times. They were sacked only 20 times all last season.

Of course, we’re not the only ones who notice these issues. During Harbaugh’s Monday press conference he addressed both. When asked about the line, he said, “we’ll look at that during the week.” When asked about play calling, the exchange with the reporters in the room was chilly - but he did say, “I want to . . . look at the film, I want to learn from it . . .” and, “As a coach you go through it, and uh, you look at where you can make corrections and where you can make improvements.” At this point, all fans can ask for are improvements in the play calling and line play - because the Penn State game, and even this week’s game against Indiana, loom larger following Saturday night’s performance.


In non-Football news, Michigan’s other fall sports continue to perform well. Field Hockey bested No. 20 Rutgers 5-0 and No. 12 Stanford 3-0 last week for their seventh and eighth straight wins in a row. Ice Hockey opened their season with a win to open Mel Pearson’s tenure, but unfortunately followed it with a loss. Women’s Soccer dropped a tough game against Penn State 0-2, while Men’s Soccer picked up a victory over Rutgers 3-1. This evening, Men’s Soccer takes on Notre Dame at 7 p.m. and Field Hockey travels to Indiana on Friday.