Monday, September 12, 2011

Notre Dame: Where 11/24 is good enough


My roommate often tells me that my analysis and understanding of sports is too statistically dependent. It's the old "that's why they play the game" argument. Human error (or human triumph, as is the case for one dreadlocked Florida-born quarterback) plays a far more significant role in determining the outcome of a game than do the statistics that say, for example, that Notre Dame should have won on Saturday in Ann Arbor or that possibly, that they were a more talented team before the game began. But for me, statistics help me rationalize things. They both confirm what I see on the field and help me reconcile how my beliefs differ from those of others.

So when trying to figure out what to write about after one of the most thrilling, unlikely football games I've ever witnessed, my mind kept coming back to this graph from Advanced NFL Stats:

That graph is intended as a baseline estimator for a team's real-time win probability and is independent of situation, but the site also offers a crude win probability calculator, which, while it's calibrated to an NFL scale, can at least give us a decent estimate of how unlikely Michigan's victory was: four percent, Michigan's win probability after Notre Dame's slot receiver scampered into the endzone without a defender in site. Denard Robinson laughs at your probabilities and says, "Really? Oh man, that's crazy," and throws the ball to Jeremy Gallon standing alone in the Notre Dame secondary.

Michigan's win on Saturday is a testament to the team's resilience and dumb luck. Luck that Tommy Rees got butter fingers in the red zone. Luck that half of Denard's 50/50 jump balls ended up in the hands of Michigan receivers. Luck that the Notre Dame secondary is as prone to crippling mental mistakes as Michigan's. But it's also the kind of game you secretly expect Brady Hoke's team to win because they just want it more, or maybe only he does, but it's something that resonates with the team.

Four percent is bad odds but it's also not a death sentence. Two years ago, Rich Rodriguez and Tate Forcier knew it. Last year, Denard Robinson knew it. And this year Roy Roundtree knew it and asked Denard to give him a chance on the final play. Brady Hoke might not know what Michigan's win probability is at any time during the season, but he also doesn't care because, and this might sound arrogant, this is Michigan.

Bullets
  • Welcome back to earth Denard. I've been skeptical that Denard can excel in a system that doesn't feature his legs like RichRod's did, and this game, in addition to his lackluster day against Western Michigan, is further evidence of that. Despite being praised for leading Michigan's winning touchdown drive, Denard was certifiably awful in this game, completing only 11 of 24 passes and throwing three bad interceptions. In spite of all of that, Denard's pass efficiency rating for the game (194.1) was astronomical because when he wasn't throwing the ball to Notre Dame defenders, he was tossing up jump balls to any receiver that had single coverage. His stat line was eerily similar to those of Georgia Tech quarterbacks of the last few years: option quarterbacks that can't complete many passes but when they do, they always go for a big gain. Denard is going to struggle mightily throughout this season, as expected. Let's just hope he has enough magic in him to get the team bowl eligible.
  • Brian Kelly had Greg Mattison's number. It's clear Notre Dame watched a lot of film from the Western Michigan game. The Irish offense was able to counter Michigan's zone blitzes with running back draws up the middle of the field as Mike Martin ran away from the play, and identify single coverage all over the field. Were Tommy Rees not Favring the game away with irresponsible turnovers, this game ends much more unpleasantly and with Mattison taking a lot more heat.
  • That said, this marks the second straight game that Michigan has won the turnover battle. Michigan's aggressive defense is causing turnovers and should continue to do so all season. It's clear this is going to be a high-risk, high-reward defense. Fortunately, Michigan doesn't face any athletes like Michael Floyd, who can so easily exploit those risks, for the rest of the year.
  • Michigan corners continued the disturbing trend of only face guarding in man coverage. Courtney Avery was called for an iffy pass interference call in the endzone late in the game, but referees make a lot of bad pass interference calls because corners are face guarding. If the coaching staff can't correct this before Big Ten play, expect teams to aggressively attack corners in man coverage.
  • Michigan's running game was terrible. Non-Denard rushers carried the ball eight times for a total of 10 yards. The offensive line was not built for this kind of offense. Moar zone blocking.
  • Denard, however, was set loose on a zone read play that we haven't seen from this or the prior coaching staff. I'll likely picture page it later in the week, but the offensive line would create a hole in the B gap and entice the playside linebacker (usually Manti Te'o) to plug that hole. When Te'o did, which he did frequently, Denard would pull the ball and run outside of the tackles. This is going to be Michigan's best running play this year. It'll also be interesting to see if Al Borges designs some counters to this play that will exploit teams trying to stop it.
  • Michigan's run defense was porous for the second straight game. Cierre Wood averaged 5.4 YPC on 24 carries. While large chunks of those yards came on long runs as the result of zone blitzes, that's one of the risks Michigan will take this year. Mattison needs to be less predictable in the future.
Going forward, I'm still leery of this offense. It had trouble finding a rhythm (Michigan's longest drive was only five plays) and doesn't really have anything that it does particularly well. Denard will need to take major steps to lead this offense to something resembling competitive. Conversely, the defense, though it has a few issues, once again showed promise.

Next week, Michigan takes on Eastern Michigan and should be able to experiment with new plays and work on perfecting the base offense. If nothing else, it'll be a virtual bye week for Michigan to get healthy.

1 comments:

Anonymous said...

great analysis. i thought the same thing about kelly vs. mattison. seems like the glow of winning is hiding the warts....


Post a Comment