The Weekly Six: The Descent of DG

Submitted by The Mathlete on October 16th, 2014 at 1:38 PM


NORFLEET! Bryan Fuller/MGoBlog

1. The Four Factors

  Expected Pts Conv Rate Bonus Yards Red Zone
Michigan 26.9 57% 1.8 3.0
Penn St 22.9 64% 1.0 4.3

So…that was not a good offensive game. But you didn’t need me to tell you that. Below average conversion rates, no action beyond the sticks. One touchdown in four combined red zone trips. Michigan won the field position thanks to winning the punt battle, stopping the fake punt a turnover in field goal range. Other years this would be cause for concern, right now, a win anyhow anyway is no time to complain about the lack of offensive success.

2. Individual Performances

Devin Gardner: –1.1 pts, –7% Win Pct Added on 31 plays

Deveon Smith: +0.1, +7% on 15 plays

Devin Funchess: +0.8, –2% on 11 plays

Amara Darboh: +2.6, +4% on 5 plays

Christian Hackenberg: +0.5, –4% on 36 plays

Bill Belton: +1.0, –4% on 20 plays

Not a lot of offensive stars in this one. Devin Funchess barely finished plus on the game after a huge opening drive touchdown. Amara Darboh ended as the only Michigan offensive player with a significant positive contribution on the night.

As for Devin Gardner, he is broken. Prior to last season, I wrote about his amazing run to end the 2012 season and that he had done things that very few college quarterbacks had every done. Based on 2012, he had the makings of a QB capable of adding an average of two touchdowns above a normal offensive output every game. He had practically done it already. And then 2013 happened. Below is his chart of opponent adjusted EV (expected value or points added) for every game he had at least 10 plays (rushes + passes – sacks).


2010/11 were pretty pedestrian. 2012 was incredible and 2013 started pretty well too. Akron was a bad performance by his high standards. Things were ugly against UConn before rebounding against Minnesota. In his next 14 games, there were two great games against bad defenses (Indiana and App St) and the heroic one-footed game against Ohio St.

There are six negative games and five more that were essentially zero (which is below average for a QB). It’s hard to say when Devin Gardner was broken but it obvious, even without the numbers, that he has been. I don’t know if it’s possible for him to be fixed at this point with this staff, but I sure hope so, because he is poised to be the biggest casualty of the Hoke era.

3. Game Chart


Hey, this one goes up!

6. –10.1% Russell Bellomy incomplete on third down (late Q3)

5. +10.2% Deveon Smith picks up a first down on 3rd and 1 (mid Q4)

4. +10.3% Michigan stops the Penn St fake punt attempt (mid Q3)

3. –11.1% Devin Gardner incomplete to Darboh on 4th and 3 (early Q3)

2. +12.6% Jourdan Lewis intercepts Hackenberg (late Q3)

1. +16.6% Jake Ryan forces Hackenberg into a 16 yard intentional grounding (late Q4)

The Blame Game is now the credit game, with a fair amount of blame as well. The results should not surprise.


1. Pass Defense: +49%

2. FG/PAT: +22%

3. Rush Defense: +8%


3. Opponent FG/PAT: –6%

2. Rush Offense: –9%

1. Pass Offense: –22%

4. Dumb Punt of the Week

David Shaw is poised to get a lifetime achievement award at this point. Stanford punted two more times from inside the opponent 40, bringing the total to 7 on the year, 2 more than anyone else.

Other Dishonorable mentions:

Washington State punted down 2 scores with two minutes left. This was a tough one because they were inside their own 10 and it was 4th and 33. But two scores in two minutes ain’t happening after a punt.

Coaching man-crush at Wyoming also punted down 10 with less than 3 minutes left.

All three were worthy candidates, without a doubt. But this week’s award goes to Coach Six-Pack, Larry Fedora of North Carolina.

Facing a Notre Dame team that would put up half a hundred on the day, Fedora called for a punt on 4th and 7 from the ND 33 down 9 points in the third quarter. UNC averaged 6.2 yards per play on the day and 32% of plays went for 7 or more yards (yes, MGoReaders, this is legal). The punt of course went for a touchback and field position gain of 13 yards, and North Carolina lost by a touchdown.

5. Fumble Luck & Last Minute Timeouts

Way back in 2011 when Brady Hoke was lucky, Michigan was the second luckiest fumble team in the country at +9.4*. Since then, Michigan has been –1.7, +1.0 and so far this year, –5.2. There is a reason they call it fumble luck. Mattison didn’t have some secret voodoo magic that results in a multitude of fumbles and recoveries, because no one does. Fumbles are lucky and Michigan been extremely unlucky on the fumble side (especially on defense) so far this season.

The sane football fan knew that Hoke’s end of half timeout was idiotic. It is my understanding that there are some that think it was a good idea based on a defensive TD potential. Some quick numbers to put this to bed.

I looked back to 2003 and found 7 cases of a half ending interception return for a touchdown, the only case that could justify the timeout. Of those 7, three cases came when the offense was within ten yards of scoring a touchdown. Another three were on returns that began close to the line of scrimmage which I guess could be appropriate to this situation. And only one on a Hail Mary returned 100 yards for a TD and that was from a 2010 matchup between Tulane and UCF that was a 41 point game at the time. Compare this with 25 offensive touchdowns on end of half Hail Mary’s of at least 40 yards. That is between a 6 to 1 and 25 to 1 ratio of bad to good depending on how you want to count it.

* Fumble Luck is calculated based on this article assuming 1% lost fumbles on most plays, and 6% on sacks.

6. Predictions

Michigan’s four factors for the season [Value (national rank/B1G rank)]

  Expected Pts Conv Rate Bonus Yards Red Zone
Offense 24.0 (101/13) 65% (84/8) 1.8 (112/13) 5.2 (55/6)
Defense 26.1 (50/7) 62% (19/4) 1.9 (28/3) 5.4 (79/10)

Michigan is 90th overall in net field position, only Penn St is worse in the Big Ten. The offense is below average at generating first downs and truly dreadful at pushing the ball down the field in big chunks. On defense, the news is better, as they crack the top 20 in conversion rate allowed and they are making opposing offenses almost as bad at generating yards beyond the sticks as Michigan’s offense is at getting them.

With a bye week upcoming, no game predictions. For the season, my numbers have an average of 1.8 wins left on the schedule with 3 wins and bowl eligibility at a 25% likelihood.



October 16th, 2014 at 1:45 PM ^

DG's decent is caused by the decent of the offensive line more than anything. Look at any team, it doesn't matter how good the QB is if no one blocks for him.

Glad we replaced the offensive coordinator but at this point the obvious person who should have been fired was Funk.


October 16th, 2014 at 2:51 PM ^

His technique is seriously lacking.  For example, on rollouts, DG had/has a horrifying tendency to entirely turn his back to the LOS facing the opposing endzone for a second or two, only to turn and suddenly face an oncoming LB or DL which then causes a chaotic scramble. Defenses picked up on that last year for sure. It appears Nuss has put a stop to that baffling practice, but it indicated a lack of basic technique by DG that should not have existed in a junior QB.  And this year, I agree the Oline is still struggling, but DG's TD/INT rate, for example, is more a reflection of poor decision-making than protection. 


October 16th, 2014 at 4:38 PM ^

But I can't help but notice it is because he never feels comfortable in the pocket and it closes so fast on him!

Play calling on 3rd and long has to improve but the offensive line cannot stop a blitz. In Hoke's tenure, I have not seen the Oline doing anywhere remotely decent against the blitz. The line hasn't even been good in other situations. Too much inconsistency is going to impact your QB.


October 23rd, 2014 at 3:16 PM ^

And there's where play calling steps in to adjust - at least at some level - for lack of protection.  Quick slants, screens, designed scrambles, etc start pressing the D-line back.  That assumes you actually "execute" which has been lacking.  So I agree in part with your position, it just gets into one of those horse/cart analyses.

Let's just admit it - Hoke and Funk just SUCK BALLS on offense.


October 16th, 2014 at 2:12 PM ^

Well, that was a little depressing.

I would add as a minor argument for the end-of-the-half TO (which I guess over time I've soured on a bit more), there was also a chance to get another hit on Hackenberg, and you also had a chance at a bad snap/fumble to boot.  Against a normal, competent offense it was dangerous, but against this crappy PSU line I thought the decision wasn't as horrible as some think.  In fact, I'm surprised Franklin even tried to throw the ball; might as well just run it and save your already-rattedl QB taking another hit.


October 16th, 2014 at 3:23 PM ^

I was surprised he didn't punt either, though I guess the concern would be a blocked punt or a bad snap.  And based on how his punter was during the game, he probably didn't need to tell him to kick it out of bounds; letting him just punt away pretty much assured it wasn't going to go very far or anywhere near people.

Peter Nesbitt

October 16th, 2014 at 2:26 PM ^

I was wondering about how much risk is really involved in Hail Mary passes like that. It's amazing to me that in a sports world so full of statistics that numbers like this aren't sticking way the hell out at coaches.


October 16th, 2014 at 2:45 PM ^

I was at the Stanford game. The funny thing was, I loudly expressed disagreement about both 4th & 3 punts and everyone around me was "Ehh, you know Shaw." Especially in this game, Hogan, Montgomery and the TEs all looked impressive. To punt and leave points on the field in a close game was terrible. I enjoy Stanford football, but Shaw needs to address his punting philosophy .


October 16th, 2014 at 2:54 PM ^

Despite Tevin Coleman being a terrifying monster-beast, it may be time to believe that the Indiana game is a win. Their defense is truly horrendous, and the QB situation is even more disastrous after Sudfeld's injury as the quality depth transferred (Roberson and Coffman) and the next guy up is now injured (Covington).

The most likely outcome then is a split with Maryland/Northwestern and two losses to the rivals. But this is still the BIG TENNN so who knows. I desparately hope we can reach a bowl to give the players (mainly the seniors) a tiny consolation for the lost season and a little more momentum for next year.


October 16th, 2014 at 3:02 PM ^

Hackenberg basically looks like Devin half way through a year of horrid OL play . Same PTSD symptoms.  Same "a few plays of brilliance" mixed between a lot of bad.   Same hearing footsteps way before they get there because one has it stamped in his mind they will be coming.

Cook for example has been sacked 3 times ALL YEAR.  Fathom that for a moment.  That's a half of 1 game for Devin.  Makes a huge difference. 

Now I don't think he'd ever be an accurate, downfield passer but it would have been nice to have seen Devin behind competent lines for 2 years.  The 2012 year you spoke of... that was the most competent OL he had.

What's sad about this year is there has been no moments of true brilliance outside of a horrid App State - last year he had OSU, ND and a bad Indiana defense games where he did shine.  This year none of those type of games where he chucks the ball up for 40 yards over and over and Gallon and Funchess go get it repeatedly. 

UofM Die Hard …

October 16th, 2014 at 3:18 PM ^

Looks like no bowling this year..if that aint a dagger to the heart for good ol Hoke than I dont know what is. 


What is we beat MSU and Ohio State but those are the only two wins we get the rest of the year  What say you guys?


October 16th, 2014 at 3:55 PM ^

To me, it seems much more realistic than the 17% chance reported in SB nation which in part used the Massey ratings.  I wonder if you just used different polls or if you considered other factors, such as the lack of independence in game outcomes apparently assumed by SBNation.  In any case, I think that the chances of UM going to a bowl will be higher if one considers a likely correlation between the outcomes of successive games. 

Even if momentum in an increasingly optimistic or pessimistic team were not a reason for correlated results--which has been subject to debate-- there are other reasons to expect correlations.  For example, if a key player is injured after 3 of the next six games, then the team is not likely to succeed as often as expected, based on results from the prior games.  Conversely, if a key player comes back from injury after 3 games, the chances of winning will be better than they otherwise would have been.  

At first glance, the effects of such correlations might seem to neutralize one another when the outcomes are symmetrical ie considering whether UM will make a bowl--ie win (0 to 3) of 6 games or (4 to 6) of 6..  But correlations will not affect the chances of win or losses eqally, depending on how great a favorite or underdog UM is.  They will give a greater bump up in the chances of beating Ohio, if correlations exist and we won the previous game.  The reason is that there is a lot of room to improve (ie a low win probability of 17% in theory has a potential improvement of up to 83%).*  . By contrast, if we lost the game prior to Indiana ( MSU), the correlation with Indiana game outcome will not have as great a potential effect.  If we are favored to win 54% of the time vs Indiana, then the chances of another loss can fall by only 46% at most.*

*In a simple two game scenario,  the degree of improvment (or deterioraton)  in the chances of a duplicate outcome will vary linearly with the correlation. 



October 16th, 2014 at 4:03 PM ^

I decided to watch U-M/Northwestern from 2003; this is the one-handed Avant catch game.  Navarre's senior year.  Everybody remembers how much hate Navarre got, and obviously he got better as time went on, but just in this highlight reel he makes three throws that I doubt Gardner can make.  Edwards and Avant catchability rules apply.  OL pass protection ability is obviously better as well, but Navarre still gets rid of the ball much quicker.  Makes the drop, gets rid of the ball.

0:33 -- doubt Gardner can make that throw.  Not a small window there, there's plenty of room, but it required some zip on it and I haven't seen Gardner throw a regular run-of-the-mill rope in some time.  

2:49 -- Doubt it, but not impossible for Gardner because it's a floater pass.  He could get lucky, I guess.

5:07 -- Not a chance in hell.  I doubt I've seen a pass like that in five years


October 16th, 2014 at 7:10 PM ^

Forcing fumbles does correlate (somewhat; lots of noise due to scheme, players and coaches all changing) from year to year, as does not-fumbling yourself, but recoveries are not under your control to any significant degree, even on sack-fumble plays.

Our luck in 2011 was due to an inflated recovery rate, but we didn't force turnovers as the elite defenses did.  When that came back down, so did we.