Rush Offense vs S&P Rush Defense.

Submitted by swdodgimus on November 14th, 2017 at 7:33 PM

Wisconsin's run defense is 25th on Football Outsiders Rush Defense metric. 

Michigan ran for 140 on the 5th best one, which is Purdue.

They ran for 271 on the 21st best in Indiana. 

215 on the 29th in Florida.

Add in two 300 plus days in the last three weeks, and I have to ask: 

Why can't we run on Wisconsin?




November 14th, 2017 at 7:37 PM ^

It's not that we "can't" it's more that they probably won't respect our pass offense and stuff the box like PSU did. Those other teams weren't as good overall on D as Wisconsin is either.. soo, grain of salt and all. 

Not saying we won't be able to have success though. 


November 14th, 2017 at 8:42 PM ^

Every team on the schedule has loaded the box to try to make Michigan throw so that's nothing new.

A big driver why Michigan has been stuffed a fair amount (and has the S&P success rate to match) but also why they've broken so many long runs.


November 14th, 2017 at 9:25 PM ^

But this is an undefeated Wisconsin team that makes plays when they need to.

The TEs making plays is key to winning this game. As well as our run defense showing up strong.

The last three games are a testament to that has been working for us.

Gentry, please have a monster day.


November 14th, 2017 at 7:41 PM ^

But there was no chance in hell at passing on them with O'Korn, so they loaded up the box. Michigan still paved them. I'm just pointing out that much worse run offenses found room to run on Wisconsin, which is why they aren't ranked super high in advanced stats on run defense.


November 14th, 2017 at 7:47 PM ^

I like our front seven on the ground against their front seven. We should be able to move their front seven, the key will be keeping the DBs away from the box. If they can put 8 in the box with no consequence, then it will be tough sledding. If Peters can hit just a few PA shots, then we should find success. I'm confident in our OL run blocking and in our RBs getting all available yards. We just need a modicum of a passing game to win.


November 14th, 2017 at 7:54 PM ^

Something I found interesting: None of their DL have a success rate (measures Offense success when that player made the tackle, meaning that the lower the number the better) below 34%. For comparison, our DL go 12.7% (Chase), 15.4% (Aubrey), 15.7% (Mo), and 17.4% (Gary). Our worst DL (Rashan Gary lmao) has nearly twice as good of a success rate as their best DL (Conor Sheehy). 


My deduction: our RBs are going to get to the second level quite often. Have to like Evans and Higdon getting chances to make LBs miss.


November 14th, 2017 at 8:50 PM ^

I agree and I considered this, but I think you would still hope that 3-4 DL would be more successful at stuffing plays than...Khaleke Hudson (who has been great but is generally 5+ yards downfield tackling someone when he isn't blitzing). Additionally, their LBs aren't under 20%. Their best is Dooley at 21%. This means that their DL is not going to stuff guys and we are going to get to see our improved run blocking matched up against their LBs a lot. I still think this means that we are going to get our RBs to 3-4 yards with a chance to make a guy miss with regularity. 


November 15th, 2017 at 2:27 AM ^

I think that the offensive brain trust has been holding things back  for the past few weeks just for these two games to be honest. We have literally ran plays out of a total of 4 formations after we went up big on Rutgers...

I hope I am right, we will need a few plays breaking big to win this game.


November 14th, 2017 at 8:18 PM ^

This is going to be a rock fight. Early weather forecasts show winds 15-25 mph with temps in the mid 30s. It will make passing difficult. However, I do like Peters more than Hornibrook in those conditions. He has a bigger arm and a tighter spiral when he throws, he should have an easier time with the wind. Still tbd on the decision making/pass protection aspect of the passing game. 


November 14th, 2017 at 8:42 PM ^

I'll respond to your concern, but first, want to hear something supremely stupid? Northwestern, maybe the team that came closest to beating Wiscy this season, had the ball with a chance to tie the game in the final minutes despite running for 25!!!!! yards. TWENTY. FIVE. WTF. Someone say the Big Ten west is boring, but I say that it is wonderful for its cold and concentrated stupidity.


But to your point, the two best rushing performances against Wisconsin this year have come from...Maryland....and Illinois??? So neither of these two performances were stellar (143 yards at 4.1 ypc and 134 yards at 4.1 ypc). Maryland has a 21 spot gap in their rushing and passing S&P ranks, Illinois has a 37 spot gap. So no, no such gigantic gap between the two most successful teams on the ground against UW. However, the best rushing attack against Wisconsin so far is currently ranked 24 spots behind us in rushing offense. I trust that we will be the best rushing performance against Wisconsin's D this season.

Real and Spectacular

November 14th, 2017 at 8:33 PM ^

Do we really think Purdue and IU have better run defenses than Wisconsin all things considered? I don't care what the fancy stats say. We have played 2 decent teams and struggled to run on both. We were "ok" vs MSU.


November 14th, 2017 at 10:04 PM ^

I'd say the opposite.  They are not infallible, as nothing is, and when they fail a bunch of dads start yelling "See, it's all lies!" like bad cartoon villains. 

I understand they're not infallible, I don't expect them to be.  Are they better than the collective eyeball test of a bunch of engineers and HR reps who aren't quite sure what inverted veer is?  Yes.  


November 14th, 2017 at 8:36 PM ^

It's amazing dpj and some of our other highly rated wrs can't do shit on offense, then Wisconsin trots out scrubs with a trash qb and can still make plays even if they don't throw for many yards.

It's just rare to see any of our wr make any kind of a play that creates an impact.


November 15th, 2017 at 2:33 AM ^

You do realize that WR is one of the hardest positions to transition from HS in to college ball, right?

When DPJ was in HS, like all the other 4/5 star receivers, all he had to do was shuffle his feet for a solid 3 seconds off the line and then sprint past the conerback...boom TD. He has never been shown how to properly run a NFL style route tree, or read defenses on option routes, etc. These things come with time, its not as simple as you think.

Once he and the other receivers start to ingrain these techniques, we will have one hell of a WR group.


November 14th, 2017 at 8:48 PM ^

A good question, as it cuts to the heart of the matter. If Michigan outgains Wisconsin on the ground they win. But Michigan will only be able to run the ball if they can effectively threaten to throw, at least. 


November 14th, 2017 at 11:46 PM ^

whichever team has the better game passing is going to win.  I'm not expecting either teams running game to be super effective.  The team that's able to complete some passes to move the chains, put themselves ahead of the down and distance curve and possibly pick up a chunk play or two will probably win.  I'm not expecting either team to carve the other up through the air, but if one team can get something like 15 for 23 for 180/190 yds that is going to be a huge boost for that teams offense.

The Denarding

November 15th, 2017 at 1:47 AM ^

There is a difference between statistics and analytics. Statistics can be misleading but analytics especially if indexes to an outcome rarely are. You can speak to correlation versus causation and how predictive regression is once you assume more volume but overall if data is clean and there is enough available longitudinally it is a very goood guide of performance. The one thing I will say is regression whether logistical or mutivariate has a meaningful flaw in that it attempts to normalize for variance. You have one equation that defines all outcomes and then scores the relative expression of variables that determine the outcome. You don’t actually create an equation for each possible outcome individually. Hence you can never predict variance. Only binary segmentation predicts variance.

In the case of college football, you want to find variables that define variance from a regression outcome. 70% of the time the favorite wins and the variables that define that winner are easily expressed. It’s the 30% that lose and their variance from the other 70 that is valuable to know. In football variance from regression usually occurs because of turnovers or explosive plays. If Michigan wins, that will likely be why because by all other metrics they should lose. Statistics may be fallible but analytics with informed data is almost always at least directionally sound.