No RA knew what the hell they were talking about, but get a psych major or something, and they THOUGHT they knew what they were talking about.
Hokepoints: Michigan's Running Game, A Diagnosis
The dorm room has a shrine to Fred Jackson they call "Like Borobudur but more majestic"
Seth, doctor of blogging, is acting residential advisor for South-of-South Quad Residence Hall, Floor 1. Having heard reports that the occupant of Rm. 219, registered as "Michigan's Running Game" has 'not been himself' lately, the good doctor attempts to ascertain the source of his charge's recent morosity. He knocks on the door…
So, hey Michigan's Running Game.
Alright if I come in?
You've been kinda quiet this semester.
… Look, I haven't known you to be the kind of dude to go into a shell. Not since you broke up with DeBord, anyway. Um, you okay there man?
Oh you know, things.
|Year||Rush YPC*||Rush S&P+||Rk|
* (Called running plays when Denard is QB, no short situations. 2012 stats are through Nebraska because this is from my UFR database and Brian hasn't UFR'ed Minnesota yet. S&P+ is a
Fremeau EDIT: Bill Connelly (they're all football outsiders) stat that measures success based on down, field position, and strength of opponent. Higher is better. FWIW these stats have been screwy this year but I think rushing yardage is the part that's actually working.)
That's…that's not so bad man, 5.77 YPC is pretty respectable.
Yeah but I'm supposed to be much better than 'pretty respectable.'
Cause Denard and Toussaint and most of the line back?
That and I'm MICHIGAN fergodsakes. Plus I think a lot of that 5.77 is Denard shooting off long runs against Air Force and Purdue. Here's a table so you know what I mean.
I see. Wait, what the hell is that?
A table of all the runs charted in that stat. So like that gray peak is the 2010 offense getting lots of 4-yard runs, and the yellow peak is the 2011 offense getting stopped for just 2 yards a lot. And the lines at the bottom are polynomial trend lines.
Poly—? Dammit man, I'm a doctor, not a physicist.
See how the yellow and gray lines follow the same trend but the blue one doesn't? The 2012 offense is ripping off big runs more often, but not getting those 5- to 12-yard runs with the same regularity.
I understand. I'll see if I can find what's going on. You mind if I ask some questions?
[After THE JUMP: Is it for want of play calling, tougher competition, or Molk?]
1. Is it the other kids?
Michigan has now played the #2 (Alabama), #18 (Notre Dame), and #31 (MSU) rushing defenses by straight-up YPC. If you want to believe Fremeau's S&P+ for defense, Alabama has the top rushing defense in the country, Michigan State's is 2nd, and Notre Dame's is 6th (we're 17—Mattison all the points). Here's how that stacks up next to last year's (ordered from toughest to easiest)
|Team||S&P+ Rk||Team||S&P+ Rk|
I put that parenthetical exclamation next to UMass because that's 124 out of 124. Ultimately this is pretty even, if more dispersed. Note too that the remainder of the schedule is comprised of opponents who tend to be stiff against the run. MSU and the three losses are the only teams better than those coming up, though that's a big gap between Iowa and ND. Anyway there has been ample opportunity to face non-scary run defenses and make up the damage they've wrought—everybody left on the schedule should be more like Nebraska than Purdue.
2. Maybe you're in the wrong major?
This has happened to friends of mine in the past. You meet a nice new coach who points at you, and he and his staff seem uber competent at running a football program, and well gosh just look what they've done with the defense, and all of a sudden you're signing up for all these West Coast Offense and MANBALL classes because that's how they got there. But you gotta be doing what's best for you and your skills.
Yesterday was the fist anniversary of the 5th of November Manball Treason and Plot against Iowa. We still remember. We still keep a detailed database. Let's see if I-form ISOs have been sneaking in when we're not looking (just the called runs):
|O-Form (s)||2008||2009||2010||2011||2012||4-yr AVG|
Nothing untoward here. I mentioned earlier in the season that Borges has come to Denard more than the other way around, and you can see some of that in the completely Rodriguezian formation selection we choose to run out of.
And we can check the yardage while we're at it:
|O-Form (s)||2008||2009||2010||2011||2012||4-Yr AVG|
It's climbing down. Notably the Ace seems to have lost its novelty. Denard Jet was replaced by Gallon Jet out of the shotgun. When Michigan's gone I-form, it's often in short-yardage situations, or as a "conservative" measure, e.g. the first half of Minnesota, and when Bellomy first came in.
The part that's probably more on Borges is running into stacked fronts, which has been a topic ad nauseum 'round here. Here's Al Borge's inaugural RPS chart.
2011: +65/-89/-24, = +1 on 14% of plays and a –1 on 19% of plays.
2012: +14/-24/-10 = +1 on 5% of plays and –1 on 8% of plays.
I don't know if that's Brian giving out RPS for fewer plays this year, or if this is evidence that Al's not kidding about "playing conservative" at several points this season. Running into a stacked box 24 times in 9 games feels like a lot to have lived through. What I think we're seeing is he's just been less willing, as the defense has proven it doesn't need as many points to win, to not gamble as much. Again, the UFR numbers are interesting but not canon—the greater evidence is his own comments.
3. Is it just that you miss Molk?
Last year Michigan had the best center in the country. They even gave him a shiny Rimington Award to recognize that fact. The drop-in replacement for the Rimington winner is Elliott Mealer, the guy whom I thought (as of the pre-season) had a good chance of getting replaced by a former walk-on (Burzynski) or a freshman sometime this season. Hoke recently mentioned they're considering this. Mealer's 2012 in UFR Charting:
|Alabama||1||4.5||-3.5||Now a lot more worried about Molk transition after flip|
|Air Force||4.5||2||2.5||Seems like a decent player.|
|Umass||9||1||8||Mobility in space a pleasant surprise.|
|Notre Dame||8||6||2||Big time struggles early, did better, also bad snap.|
|Purdue||5.5||5.5||0||Can't seal like Molk against big time competition.|
|Illinois||9||6||3||Got blown up pretty good a couple times, otherwise okay.|
|MSU||8||6.5||1.5||Around even usually, which is meh.|
|Nebraska||5||5.5||-0.5||Also also crappy pass blocking.|
That's a +50/-37/+13. In an average 2012 game, Mealer will do five really good things and four bad things. Mealer is also consistent—he fared okay against everybody except Bama, be they Louis Nix or Kawann Short or Akeem Spence or Baker Steinkuhler, or Double-A Gap-blitzin' MSU (I know he'll come in for at least one big minus from Hageman this week, but it seemed he started faring better in the 2nd half). As for Molk's 2011, I'll save you a chart and tell you he totaled a +130.5/-38.5/+92. On a per-game basis that's 11 good things and three bad, a major difference-maker. These are subjective and not at all scientific. But they do kind of wink at a story, wherein Mealer isn't any kind of problem so much as Molk was an extraordinary human being who cannot be replaced.
He was also a four-year starter, and if there's a Molk effect going on, it's possible, though hard to judge (Heiko maybe you ask Borges at presser?) without direct knowledge of huddle goings-on what kind of difference it makes that a new starter and new center is in charge of calling out blocking assignments pre-snap as opposed to a guy in his fourth year doing it. I don't remember the coaches saying anything to this regard, but the football players I lean on for this kind of dope are mostly offensive linemen, and one thing they all seem to value more than John Q. Fan is the wisdom of the center.
This is going to be a gross misuse of UFR data but here's the average CHART data for run blocking this year versus last:
The Mealer thing yes, but there's regression across the board. And while Schofield is within the margin of error for treading water with Huyge, the boost that Koger lent to the running game is gone too. As for the returning guys and their apparent regression, maybe that's not having Molk around? No way I can say from this. But it's as good an explanation as any. Feel better?
That depends, how's Denard's arm? Because:
|2012 Michigan Rushing YPC when...|
|Denard is QB||Someone else is|
When yer talkin' you ain't drinkin'. Booze will help him forget. Tell him that.
I just threw up in my mouth a little
I'd say Schofield looked much better as a guard last year than as a tackle this year (similiar to Steve Schilling), and he was far better than Barnum has been at left guard. The biggest problem I see with the run game is the three interior linemen just aren't that good. Omameh is probably average in this system, Mealer maybe slightly below,, but Barnum has looked pretty bad. If we had any depth at tackle I'd love to see Schofield replace Barnum at guard, but this is another case of poor OLine recruiting biting us in the ass. No one to play tackle. At least this staff has taken the mantra "OLine depth" to heart.
If lewan stays next year if we could see a lewan-schofield-miller-joeyb-magnusson line.
to see Kalis in there at a guard spot...so if Lewan's back and Miller adds enough weight to start at C, i think it might be a battle between JoeyB and Magnusson (and Bryant? seems doubtful) for the last spot, with Schofield starting at either G or T, depeneding on if they like JoeyB or Magnusson more.
oversight on my part. I was mostly considering a line where Schofield moves back inside to make room for Magnusson
Polynomial curve fits are completely meaningless in this case. It looks good and makes you feel smart, until someone remembers it is just a few mouse clicks in excel.
There are many ways to analyze the data set that would bring more meaning conclusions.
First thing I thought too - a Weibull or generalized Beta would be much better.
TWIS had stanford schade and it sounds a lot like you two smart guys debating data cloud and regression analysis. that is so dope.
Which? I didn't use it to feel "smart" but because it seemed to show the curves. Straight-up trend lines? I'm a history major trying to do my best in a world governed by math.
I put the data in a Googledoc: https://docs.google.com/spreadsheet/ccc?key=0AlZp0ugnJIGddEtWZjg5MlloQ2F....
No curve fitting is really going to get the job done here. If you want to show that there are less 'long runs' in 2012, you are looking at the tail of the popultion with very few points.
I would just look at a data set of only 'long runs' and exclude about 90% of the data to start. Then you wouldn't be looking at a relatively huge number of ho-hum runs and trying to make a conclustion about a long tail.
Stand by, I don't have much love for google docs, but I'll get back to you in a few...
So, here I've plotted sorted run lengths of runs greater than or equal to 8 yards. Why 8? I just wanted a tail of the population and it seemed like a decent cutoff point. Since the 2012 season is not yet over, this isn't a fair one-to-one comparison, HOWEVER, in terms of 20+ yard runs, we are already just about equal to 2010. To try to make up for this and extrapolate to the end of the season, I've randomly sampled all 8+ yard runs from 2012 to make a new data set with a length that is the average number of 8+ yard plays based on the 2010 and 2011 season. (Yes, this is very very crude, I know).
In short, we are basically on pace (if my crude assumptions aren't totally wrong) to match our long run play total from 2011. Our ability to run long distances in one play hasn't changed.
The only thing here is yeah, we are on pace for as many long runs as before, but what I was trying to show is there's a difference in that chunk of 6- to 12-yard runs that Michigan got with relative regularity in 2010 and 2011, but isn't getting this year. The long runs against bad run defenses is what's keeping the YPC high -- the distribution of the runs less than 8 yards per play is where the action is.
So, as an explanation, I'm a nearly done science phd student and 80% of my time is spend on data analysis, writing about that analysis, and then getting it picking apart by my advisor.. and then hopefully improving. Reading back over this thread, I hope you haven't mis-read my tone.
In the OP you stated "The 2012 offense is ripping off big runs more often, but not getting those 5- to 12-yard runs with the same regularity" which is two different statement. When I'm writing manuscripts, I start with making graphs and seeing what the data is like and then draw conclusions from that. When I'm trying to communicate those conclusions to others, I almost always end up creating a entirely new set of figures that have been honed in to just the conclusion I'm pointing out... and not the entire data set. My main issue was a trendline over an entire large data set (from negative yard plays to 70+ yard runs) and making conclusions on only part of that data set.
If you want to make the claim about 5 to 12 yard runs (note, I agree with your conclusion), I would suggest working backward when it is time to make support your claim. In this case (as gbdub makes a post about below) I would suggest looking at the percent of 5 to 12 yard plays as a part of all runs from a given year.
Sorry for the possible overkill on your work. I hope any of this has helped and I would be more than welcome to give advice or run data analysis in the future.
Dork Rule 1556: There's no such thing as "Overkill," just "Enough kill."
As you may note from where I respond, I take criticism, especially correct, well thought out, constructive criticism, way better than acclaim. I very much appreciate any help you can provide.
I totally get what you mean about recreating the graph. I have a tendency with these to want to show my work, but there is always a process very much like what you described. I'm just fumbling my way through all the time.
I made a LOG of the trend lines as a separate thing but then didn't post. Here's that:
click embiggens. Note it cuts off most of the top stuff. I guess this would have worked better. The idea I got from a guy who does the same to stack WAR values of Hall of Fame candidates against each other. Different career lengths and different peaks make for different trend lines but you can clearly see a difference between a guy like Trammell and how he was, by WAR anyhow, way more valuable than any shortstop of his era except for Cal Ripken.
I still can't understand why the run game hasn't shifted to take more advantage of Lewan being Jake Long Jr. It seems the run actively avoids going behind him for some reason. Maybe he isn't as good as he was hyped to be, but you would think he is better than the other 4 guys on the line.
It was especially interesting to see that in the goal line situtation in the last game when they had 7 linemen in that they ran to the right instead of the left. That just blew my mind.
If the RB is getting swallowed behind the guard, it's hard for him to get out behind the tackle.
Also, Lewan's run blocking has never been spectacular.
Finally, it also seems pretty clear that Lewan is hurt. He's limped off the field every game since Alabama.
This is just a guess, but could it be that we are running primarily to the right side because of Denard's injury? His right side is facing traffic when we run left.
I think the bigger problem is probably that Omameh has never been a good pulling guard and Barnum just isn't very good at the point of attack.
And for the record, Lewan is no Jake Long. Frankly, he's not even a Tony Pape. He's more of a Backus/Stenavich. Good, occasionally great, but not a dominator.
So, respectfully....you'd rank Tony Pape ahead of Backus?
It depends on how I'm feeling that day. I think Backus was made to look better than he was because of who he was playing next to. For 4 years, he lined up with the greatest lineman Michigan has ever had and one of the greatest players to ever play his position, Steve Hutchinson. But Backus and Pape were similiar: almost no MAs, almost no MMs, far fewer penalties than Lewan. I think Pape was the better pass protector, while they were pretty even in the run game. And I have a lot of respect for Pape because, as his lack of an NFL career shows, he was not a great athlete, just a great technician and a smart player. But, long story short, it could go either way.
until he cleans up the penalties... taylor is not jake long jr... remember how jake gave up like 2 sacks and 1 penalty in his career? yeah... lewans got time though.
When I look at 2011's run performance compared to 2010's, then 2012's compared to 2011, it's really, really, really hard for me to conclude that Borges is the problem here.
I respect your opinion on these kinds of things - how would you diagnose the issues? Is Borges completely "clean" and it falls on other aspects of the game? Or would you say that Borges's conservative calls have contributed to the reduction in productivity (whether that play-calling be purposeful like vs. MSU or otherwise)? I suspect its like everything else in life . . . people want to place blame on one thing but like Seth describes, it is always multiple things that end up being the cause.
I don't know a thing about scheme.
But I'd always side with "a variety of things".
That being said, I doubt that Borges suddenly forgot how to call this offense. Like, this is the guy that buzz-sawed Illinois, Nebraska, and OSU to close the regular season last year, all of whom were top 30 defenses.
My theory was sparked by Brian's comment that suddenly, there are a ton of unblocked LB's. My thought is:
1. No professional Offensive Coordinator draws up plays where MLB's are routinely unblocked. That just doesn't happen.
2. We replaced a 4-year starter and Rimington winner at C with a guy that became a C 2 days before our first game (if you recall, Barnum was listed as the starting Center on the last depth chart, and Mealer at guard. Furthermore, that guy was nearly beat out by a younger walk-on.
3. Lewan has been visibly close to being a cripple this year. His leg is not, at all, OK.
So, I think Mealer, who has been a C for about 9 weeks, likely doesn't call assignments at the line as well as Molk, a 4-year starter, accounting, to a degree, for unblocked LBs. I also think that the efficacy of Mealer means fewer successful combo blocks (I seem to recall Molk being very effective at doubling with a team-mate, then releasing to the second level). If Mealer and Barnum or Mealer and Omameh can't effectively kill the DL with a double, nobody can leak out to the 2nd level. I think Schofield appears to be a better G than a T, and was definitely better than Barnum. I think Lewan 2012 is not Lewan 2011.
The whole thing, honestly, makes me appreciate Denard more. Our OL can't block, and somehow he was on pace for about 1500 yards.
This post comparing years, while not strong on objective data, is a decent trend on "are things similar". Brian is likely pretty consistent with his subjective UFR judgement the past three years, and if anything would have been biased toward 2010 running.
Another aspect is what you can't know about is the execution of plays in practice. But the coaches definitely know how well the plays are being run this year compared to last year, and if there are plays that the OL just can't execute, then they won't get called, even though schematically they are "the perfect play" for what the defense shows you.
With Borges, and more importantly Hoke stating that they are simplifying and being conservative this year, you really have to ask your self, "is that because they like to be that way or the have to be that way?" Considering how many people are now 20/20 hindsighting the "Devin as WR" decision as clearly wrong, and Hoke stating in his press conference that he essentially would make the same decision again because it was the best decision at the time, I don't think you can really state Hoke is conservative because he's risk averse like Debordian-Lloyd. That leaves, the answer to be, "he's gotta be".
For me the two factors are that the OL is not as strong/smart as it was before, (also remember that Lewan was pretty clear in the press conference that Mealer makes the first call, and Lewan's is based on Mealer's call) and Denard is just not a good decision maker, especially under pressure.
Hoke explicitly stated he took away Denard's ability to audible. Borges said that the plan for MSU was to avoid getting into a "chess match". Put those two together and the coaches decided that the defenses were masking coverages too well and confusing Denard, and baiting him into bad checkdowns. Yes, once you take that away, you take away the ability to constraing the defense, and now you can quote everything "smart football" has transcribed down from all the coaches he has interviewed (he is a sports fan who trained as an attorney, so even though I find his web site informative, I don't think you can consider him a schematic expert, he's likely a well informed amatuer, but he's never had to coach players and watch them mis-execute schemes. He also tracks the NFL quite a bit, where there is no limit on practice time cause they pay those players lots and lots of money)
Yet despite making Denard completely unsurprisingly one dimensional, as long as he stays in the game he is 75% of the offense.
Gardner got some good experience against a less talented team, and if you look at all the young guys who got playing time, you'll notice that the OL stayed as the startes the entire game. I think it's because Hoke knows they need the practice. And now he's thinking of replacing Mealer.
Everything starts with the Center and QB. I hope Jack Miller is the Molk-in-waiting we've hoped he could be.
Very well said. I'd only add that expecting Miller to be Molk is begging for disappointment. I would put Molk out there as the best center Michigan's had since Gerald Ford, certainly the best in my time watching, and that time includes a lot of dudes with long NFL careers and Superbowl rings.
And remember Molk was a hell of a player but also hellishly undersized in 2008--when a guy like Ogbu got leverage, Molk was deposited 9 yards into the backfield. Even if Miller is Molk 2.0, he wouldn't be strong enough yet to go against the hellish DTs all over Michigan's schedule, and in this offense the mitigating ability to get perfect seals on reach blocks wouldn't matter as much since the center is barely ever asked to do that and I don't think Borges or Hoke really think that way.
And though he may not look it now...
This guy might have an argument. He didn't win the award, because it didn't exist then, but he was the best center I've ever seen for Michigan. And Molk was pretty damn great.
I watched every game since '91 but yeah, Everitt (or "Brady Hoke's Pet Viking" as I'm proud to have christened him last year) was something I got to witness before I really knew what I was witnessing. That was several years before Michigan Replay became a staple of my life. I think 1995, when I was a sophomore in high school, is when I started knowing things like "our center is Rod Payne," but even then my knowledge was at like a Talk Radio level--more "Biakabutuka is better than Wheatley!" or Griese-Dreisbach fights than "Jansen, Runyan and Payne are awesome but the guards need help. Dammit in re: Denson!" I've had to re-watch a lot of games that Wolverine Historian puts up to recalibrate.
Anyway there's no way I can really compare Everitt to Molk, whose career I watched in slow-mo while--at least for the last two years--writing professionally about its minutiae. I can say Ford because he seems to be the gold standard. Interesting note: none of these guys was named an All American, weirdly (though Payne was!).
Because (wiki caveat) he's listed as AP and concensus All-American.
Everitt didn't win because also a senior that year was everyone's favorite Lion and 12 yeaer professional Mike Compton. Steve spent a good 7 years in the League, so he got more than a cup of coffee though.
There's way more unknown (is it technique, an injury to Lewan, to Toussaint...?) than known when it comes to a problem that seems to be originating among the Daves.* I hope people noticed that most of the "is it Borges?" section debunked going theories of Borges-blame. The part he does seem culpable for is the conservatism. Since Brian started giving them out, an RPS +1 usually leads to--on average--a 10-yard gain, while an RPS -1 usually led to an average loss of 1. You'll note last year there were bunches of 10-yard gains and bunches of losses of 1. It's possible that was the effect of the "chess match" stuff Borges has said he's not doing this year.
The other thing I hope people noticed here: 5.77 ypc on called rushes and 23rd in the country in S&P+ isn't awful. It is good. Part of the effect is the extremism of the the schedule: we have faced the best rushing defense in college football, AND the worst rushing defense, AND the 2nd best rushing defense, AND the 15th worst, AND the 6th best AND the 28th worst... After 9 games the only defense that you might define as any kind of "norm" was the Blackshirts, and for half of that game we were running into blitzes so that something other than getting the freshman pulverized could happen once in awhile.
The last thing I hope folks notice the difference between interior DL among our opponents vs their DEs. This was something that became apparant while doing that draft-o-snark this summer: that other than State's Marcus Rush, OSU's John Simon, and Illinois's Michael Buchanan there's not a lot of elite edge rushers. On the other hand there's been some monsters lining up directly across from Mealer: Jesse Williams, Hageman, Short, Spence, Nix, even Steinkuhler will probably get drafted. And you know Hankins is still coming up. The best way to clog up a running game is to have a guy who gets a 2-for-1 consistently--that's really hard if you're a linebacker or end who has to pop a fullback or tackle into the pulling guard, but automatic for a DT who demands a double-team every play.
* Dave Baas, Dave Pearson, Dave Petruziello. It has been my goal since 2002 to replace the bulky phrase "interior offensive line" with "The Daves." Join me!
More Kwiatkowski blocking!
I blame all four for our running game's maddening lack of consistency. It seems to be pretty much a bad play or a great play, with very few of the 4-6 yard runs in there.
I am a Borges fan, but you can't let him off the hook. No matter how big the challenge, a coach must find a way to maximize the strengths of his team. Of coures, there's no way of knowing if the guys we have can play better than they have, because, well, they haven't.
Denard's limited skill-set is, by far, the biggest villain. Don't get me wrong, DENARD is not a villain, and I love him as much I love bacon, but he never really made the leap as a passer, and has proven that when he throws the ball, there is a very good chance that he will throw it to the other team. Which is not good strategy. The fact that we know he is an amazing runner who will break a big run at some point in every game combined with the fact that he is a bad passer who will do something dumb in every game has hand-cuffed the offense. And our great defense gives us another reason to be conservative.
The O-Line has not been good. In fact, their inconsistency is probably the biggest reason for our inconsistent output. Yes, they are facing stacked boxes, but you still have to block, and you still have to hold your block. They seem to get their assignments (most of the time) but can't seem to finish blocks...that points to talent more than coaching.
And finally, the Universe. We aren't causing as many turnovers and we lost our starting QB to a freak non-contact injury. Fitz seemed to lose some of his magic wiggle over the summer. The throwback screen has been figured-out by our opponents. I blame the universe. And science.
I think the run game has suffered because 1) the line is not quite as good as they were last year, and; 2) opposing teams have NO respect for our pass game and sell out to stop the run.
But, I still think the biggest factor is Fitz' vision. It looks like he has blinders on sometimes and he just runs into the pile. Granted, he does get swallowed up in the backfield sometimes because of poor blocking, but there are other times when there is a hole or a cutback lane and he completely misses it. Norfleet does the same thing on kickoff returns...
But he had better vision last year? I'm not sure that makes sense to me.
at bat. Sometimes they claim they're "just not seeing the ball", and so they don't hit it. Same thing in basketball. Call them "slumps". In football nomenclature they say "he's not seeing the field well". Its not meant to be taken literally; its more of a timing issue. I'm saying Fitz doesn't appear to see the play/blocking develop in front of him; or, he's being impatient and not letting the play/blocking develop, hence the slamming into the backs of his linemen; or, he's just not seeing the field well. From what I've read, notwithstanding the oline changes, there seems to be a growing consensus that he doesn't look like the same runner as last year. I agree. But, he's not slower; he's not smaller; he's not older... therefore, its likely a timing/vision issue.
Is it just me or did it look in the Minny game like Lewan was calling out some of the blocking assignments?
Lewan has said that Mealer calls assignments, and Lewan coordinates he and Omameh based on Mealer's call.
Between replacing Molk with Mealer, and Lewan clearly not being 100%, it makes sense that the Oline is struggling compared to the last few years. It's still frustrating and concerning that MLBs are filling gaps completely unblocked, and I would love to know if this is a technique or scheme problem.
It's the OL, and since everything starts up front, it's near impossible to get consistent production with a weak OL. Borges could be replaced by Vince Lombardi, and people would be saying Vince has lost his touch. This will be patently obvious next season when Michigan has 3-4 RS Freshman OLinemen starting, and there is a significant improvement in rushing production even without Denard.
Hoke & Co. knew the OL was a weakness this season, but it appears that they felt that burning RS's was not worth it, and that they felt it was time for the upperclassmen on the OL to step up. They haven't, and Michigan's rushing has suffered. Would it be any better with Kalis & Co. starting? Doubtful. We're stuck with the OL we have, but it will all soon be a distant memory as the road graders are here, and more are coming.
I wouldn't say they're a weakness. They were supposed to be a strength. Lewan is the best OL in the conference. Schofield was capital-G Good last year as a guard and has the size and recruiting profile you'd expect to translate to tackle. Barnum too had that profile, and until he was hurt last year he seemed to be completely on track to be a plus-plus interior lineman. And Omameh is a 3-1/2 year starter whose chest is littered with commendations for turning Manti Te'o into a blunt object for blocking safeties[insert DENNIS BERGKAMP here].
The only question mark was Mealer and while Mealer is probably the worst of the group, the sum difference between his play and Barnum or Schofield's isn't all that big.
The pre-season expectation for this OL was they were built to dominate at spread stuff--to combo and release and do mean things to the 2nd level and never even bother with backside ends, and if any of the starters went down it's doom doomy doom. Weirdly you'd think pulling is related to this but they never did seem very adept at pulling, and this offense pulls a lot.
OL recruiting is the biggest crapshoot in recruiting. There's certain commonalities that seem to mean "Oh Yeah!" -- like when a guy comes to it late and then picks up the footwork ridiculously fast but needs to get a lot bigger (examples: Long, Lewan, and Kugler and Magnuson among the newer guys, maybe Chris Fox except he's already almost 300) it seems to turn out better than the highly rated guys who are already 300 lbs in high school and get called a "road grader" and "could already play on and FCS line" Examples of such among recent blue chips--5.9 or higher to Rivals--are Steve Schilling, Alex Mitchell, Dann O'Neill, Corey Zirbel, Brett Gallimore, Marcus Slocum, Justin Boren, and while I'm not down on him at all, yes Kyle Kalis and Kyle Bosch both fit this mold. Obvs, some of those guys turned out to be damn good linemen, or washed out for other reasons. OL recruiting success rates say few and certainly not all of these OL recruits are not going to get to Schilling-level by 2016, let alone win a Rimington like Molk or be the first pick in the draft like Jake Long. If 1/4 of the 2012 and 2013 o-line recruits are All Big Ten starters in five years that's a success. If half are that's wild success.
You might say they're a not weakness, but that doesn't agree with the eyeball test. Further, I listened to the Minn. game on WOOD and Branstader was uber critical of the Michigan OL. The Minn. DL was absolutely LIVING in the Michigan backfield, and every team we've played has too. Whatever the issue, these guys cannot block.
S&P+ is not a Fremeau stat.
FEI is the Fremeau stat and is possession based.
S&P+ is a Bill Connelly stat and is play based.
NCAA F/+ is a combined rating.
Beginning with the 2009 Football Outsiders Almanac, Brian Fremeau and Bill Connelly, originators of Outsiders' two statistical approaches -- FEI and S&P+, respectively -- began to create a combined ranking that would serve as Football Outsiders' 'official' college football rankings.
What I'd like to see, but don't currently have the tools/data to show, would be a plot of "% of run plays >/= x yards". This would mitigate the issue of looking at the tail and would normalize for number of run plays. If Seth's diagnosis is correct - we still break long runs because Denard, but can't consistently get medium gains - then you'd expect to see the following:
1) in 2010, the graph would have a gentler slope out to say 4 yards - a very high percentage of runs made 3-5 yards, fewer runs for less than 0 yards.
2) in 2012, there would be a steep dropoff after 2 yards.
3) after you got out to 12-15 yards the two lines would be nearly equal. You could cut off the graph at this point.