Mid-Week Metrics comes back without compare Comment Count

The Mathlete

Thanks to user Feat of Clay for helping name this segment. This one comes with  lots of fun numbers to look at from an instant classic but also with lots of question marks. Will get to the good stuff first.

Win Percent Added

Let’s go straight to the chart:

image

Pretty nice job by BlueSeoul in his estimate on this. Michigan saw two steady descents in win probability over the first three quarters with only the first bomb to Hemingway to nudge the numbers up. It wasn’t until Michigan pulled within a field goal that the odds got above 30% for the first time since the opening Notre Dame drive.

WPA Detail

The WPA graph shows the flow through time but has a lot of approximations. Since this game was so spectacular, I decided to go back to all of the major lows then highs to see what the odds of winning were for each specific situation based on the last 9 seasons (the numbers will differ slightly from above in some cases):

Notre Dame scores first (0-7): 29%

Notre Dame extends lead to 14 (0-14): 9%

Michigan gets on the board (7-14): 28%

Half-time (7-17): 17%

Notre Dame Scores Again (7-24): 6%

Michigan scores to start the 4th (14-24): 13%

Michigan cuts it to a field goal: (21-24): 21%

Denard throws a pick in the end zone (21-24): 17%

Michigan gets the stop (21-24): 28%

Michigan completes comeback #1 (28-24): 77%

Michigan allows comeback #2 (28-31): 8%

Epic comeback #3 (35-31): 99.999% (allowing for Stanford-Cal)

Since 2003 I have 38 cases where a team was down 3 with the ball with less than a minute left and first ten at their own 20 (give or take ten yards). Of those 38 cases, 35 times the team failed to score or lost in overtime. In 3 of those cases, the team was able to kick a field goal and win overtime, including Dooley Premature Handshake Pt 2 last year against North Carolina. Michigan is the only team that was able to win in regulation under these conditions. The closest I could find were three times when teams received the ball down by 3 with 1:15-1:20 left on the clock and went on to score.

Arizona State actually came close, scoring in 31 seconds and giving the ball back to Purdue with 43 seconds left in the 2004 Sun Bowl, but they had more time if they needed it. Strangely enough, the other two times were both done by Arkansas State in 2006 and 2007.

Turnovers

Factoring in fumble and interception returns, Michigan is currently first in the nation in turnover PAN with 14.4 pts/game net between the offense and defense. This is obviously unsustainable. Last year UConn finished first overall with a season average of +4.9 points per game in turnover value.

The good news is obviously that Michigan is generating a lot of points off of turnovers through two games. The bad news is that it is unsustainable and Michigan will need to find other ways of generating points than what they have so far.

The turnover-prone Irish sit at –15.2 PPG through two games. It should be noted that for all the mileage gotten out of mocking the Irish’s turnover failures through two games, Michigan has been nearly as strong to the positive. Neither trend can continue at their current magnitude, but let’s hope they do.

 

 

Eastern Michigan Preview

Eastern Michigan is one of four teams that have yet to play a game against an FBS team this season. All numbers for them will be based on 2010. Michigan’s numbers for 2011 are still not opponent adjusted.

Michigan Rush Offense:

Michigan 2010: +6

Michigan 2011: +3

E Michigan defense 2010: -4

Michigan Pass Offense:

Michigan 2010: +3

Michigan 2011: +6

E Michigan defense 2010: -8

E Michigan Rush Offense:

Michigan defense 2010: –3

Michigan defense 2011:

E Michigan 2010: +0

E Michigan Pass Offense:

Michigan defense 2010: -3

Michigan defense 2011:

E Michigan 2010: -1

Special Teams:

Michigan 2010: -2

Michigan 2011: -2

E Michigan: –3

Prediction: Michigan by 21, 98% chance of victory (although my numbers typically underestimate blowouts)

Luckily for Michigan they have three games in a row that are at home and against teams that they should be able to win while working out some kinks. As you’ll see in my Big Ten team rankings below, the numbers are not kind to Michigan so far. The bad news is that the offense is down, the defense hasn’t show much bounce on a down by down basis and they are heavily dependent on an unsustainable turnover margin.

The good news is that they have Denard Robinson, are 2-0 and have two accomplished coordinators behind the wheels of the offense and defense. It will be critical to come away from the next month with wins in hand and a lot of holes fixed. If the coordinators live up to their reputations Michigan should be looking at a strong first year under Hoke and a stepping stone to a high expectations 2012. If they don’t it could well turn into 2009 Part 2 with a quick start thwarted by an offense that is good but incapable of hitting Ludicrous Speed and a defense that is undermanned.

Quick Hits

Denard Owns The Irish: Denard has almost a thousand yards in two games against the Irish. The number one yardage total for one player against one team in my DB is Dan LeFevour at Central Michigan against Ball St with 1,575 yards. Brady Quinn against Purdue is second. Adam Weber at Minnesota holds the top two within the Big Ten with nearly 1,200 yards against both Northwestern and Wisconsin.

Rush Defense by half (via @MeanChuckieB): Big swing from first half to second half. –4 in the first 30 and +4 in the second 30 minutes with a couple of big 3rd down stops.

Short Drives (via @drboud): Michigan’s longest drive of the day was 5 plays. 15 teams have done that in a game since 2003. Of them, only four have scored more than a single touchdown with Michigan’s 35 points far exceeding the previous high set by Boise St against Hawaii in 2003. Coincidently, the only other time the 5 plays or less on every drive has happened this year was also in Michigan last week with Michigan State’s domination of Florida Atlantic.

ND offense 2011 vs 2010 (via @doughelmreich): Eliminating the turnover portion and this year’s version of the Irish offense is far surpassing last year’s performance. Without adjusting for opponents and removing turnovers, last year the Irish were +5 and this year they are at +13. Most of the jump has been in the passing game with has gone from +4 to +10. They are identical to last year on first down, but significantly higher on 2nd and 3rd downs, especially in medium and long yardage situations.

My Top 5: 1. Oklahoma … State 2. Alabama 3. Boise St 4. Oklahoma 5. Wisconsin

Rest of the Big Ten: 9-Nebraska 13-Ohio St 14-Michigan St 23-Illinois 42-Penn St 58-Michigan 77-Northwestern 79-Iowa 88-Purdue 100-Minnesota 106-Indiana

Comments

jlvanals

September 15th, 2011 at 1:41 PM ^

Last time I cared was Antwaan Randle-El.  Best college football player no one remembers, first IU player to make buying season tickets worthwhile in my lifetime (I was too young to really remember Anthony Thompson, even then he wasn't as exciting as Randle-El). 

And I hated Miles Brand.  You fire Bob Knight, I will refuse to say nice things about you even when you're dead.

Joined: 09/21/2009

MGoPoints: 707

Mengin06

September 15th, 2011 at 12:18 PM ^

While the turnover margin is unsustainable, it seems like the defensive staff is putting extra emphasis on creating turnovers this year which is encouraging.  While we may drop in that number, I think it will be much better than the previous few years.

Joined: 02/02/2011

MGoPoints: 144

joeyb

September 15th, 2011 at 12:35 PM ^

You know, the last 2-3 years, opponents have scored at will. We have had to just outscore the opponent to win games. So far, things don't look different this year. However, the one thing that is different, so far, is that our turnover margin is positive instead of negative. We did start out this way last year, but I think it's more sustainable this year than last. Last year, the turnovers came to us. There were a couple like that in the ND game, but there were just as many that were because of our players making plays in both the ND and WMU game. The other end of this is that we aren't fumbling (knock on wood) and when we are they bounce our way when they always seemed to bounce the other way last year. Even if you only count the two INTs against ND and the two scoring turnovers against WMU as forced turnovers, that still puts us at +1 on the season, which I think is sustainable. I don't know where that puts us in terms of PAN, but I'd assume that's a hell of a big improvement over last year's PAN from forced turnovers.

Joined: 10/12/2008

MGoPoints: 16845

Phil.engin2011

September 15th, 2011 at 12:45 PM ^

we looked poor, I think, in the first half, getting repeatedly gashed on ND runs.  

 

Second half was better, according to the stats and my eyeballs, but mostly I remember more than a few third and short plays getting stuffed.  My question is:  was this due primarily to adjustments we made or ND's lack of execution?  Also, besides the stops on third and short, how did the run defense look the rest of the second half?

Joined: 06/29/2011

MGoPoints: 147

biakabutuka ex…

September 15th, 2011 at 12:48 PM ^

Interesting stuff. But why do I feel like this defense is better than last year's? Maybe it's the turnovers, or maybe it's the fact that they improved as both of their games progressed? Or perhaps it's good old-fashioned self-delusion. At any rate, I'm never a gut feeling kind of guy, but I can't help it in this case.

Joined: 04/17/2009

MGoPoints: 2849

wile_e8

September 15th, 2011 at 1:29 PM ^

I think it's clear that the coaches are still figuring out who can play and what they can execute. So while that leads to problems as they shuffle through different player combinations and try new packages, it is also clear that it should work when executed correctly. As opposed to the wreck of a scheme last year, which never had a chance.

Joined: 06/30/2008

MGoPoints: 7381

Farnn

September 15th, 2011 at 1:39 PM ^

Well ND was 8/14 on 3rd downs, don't know what teams averaged last year, but considering NDs offense I think that's pretty good.  Also I would assume the run defense struggled because Mattison was so worried about the pass.  When it was 3rd and 1, the run defense was pretty solid, with ND failing to convert 3 times.

Joined: 07/07/2008

MGoPoints: 12788

Blue in Seattle

September 15th, 2011 at 1:49 PM ^

I no longer see people hit the ball carrier and bounce off harmlessly.  Nor do I see a plethora of weak arm tackles escorting the ball carrier into the end zone.  I think this coaching staff takes a very long term approach to coaching.  The team was way behind on fundamentals, which is the foundation of everything else you try to do.  So teaching of fundamentals was a priority over "installing schemes".  Considering that a lot of schemes are being inserted, I think it's amazing these players aren't doing worse schematically (player in wrong position, or running to the wrong spot, or just standing still dazed) than they are after 1.75 games.  The first quarter against WMU was very indicative of this long term approach.  What is so amazing is how well the players correct the mistakes they have made.  I'm completely trusting the wealth of analysis done by everyone else on this blog, but my summary conclusion is that while mistakes are happening, they are not being repeated, or if they are, they are done screwed up in the opposite direction (ahh, but give TRUE FRESHMAN Brennen Beyer a break, for godsakes).  This tells me that the learning is happening in realtime, and the trend is upward.  Yes we had what appears to be the luck of the Irish last game (and we're not even the Irish, go figure?!?) but now we have three exhibition/training like games ahead of us at home before the First Exam in East Lansing.

As far as turnovers go, I do believe in the strong random component that statistics tell us does not tolerate people who try to stand in the tails of the bell.  BUT there are two things I have noticed.  1 - we've been on the bad tail of the bell curve for three years, maybe more, so while Karma is a bitch, this year the Bitch better have my MONEY!!, and 2 - I think the fundamental first teaching approach is making sure that IF a turnover occurs, there is a much greater likelihood that a Winged Helmet is there to capitalize on Lady Luck's offering.  Exhibit A - Will Campbell was involved in two of the fumble recoveries.  The first from Cierra Woods, he was Johnny on the spot.  The second is harder to notice, but it's Will Campbell's hand that sweeps the fumbled pass over to RVB, and Cambell is in position because he penetrated to the QB. 

 

 

Joined: 07/02/2008

MGoPoints: 1716

cclittle

September 15th, 2011 at 3:39 PM ^

"Neither trend can continue at their current magnitude, but let’s hope they do."

I don't really need ND to keep losing in tragic fashion- they've already suffered enough.  This especially applies to next week.

 

Joined: 10/04/2009

MGoPoints: 956