Brian,

Apparently, this was the worst showing by U-M in the draft since 1994 when Derrick Alexander was the program's only player selected that year.  People are using this as further evidence that the cupboard was bare when Rich Rod arrived on campus (as if anyone paying attention needed more evidence).  But one draft doesn’t tell you much about the talent level of a particular team.  For example, that 1993/94 team still finished 7-4 and 23^rd in the Coaches’/AP.  Why?  Well, because that team also had three players who would be selected in the first round of the 1995 draft, and five players overall.  If we want to know how bare the cupboard was when Rich Rod arrived, we also have to look at the 2010 draft.  So, of the current players eligible for the draft next year, who other than Graham is likely to get drafted?  What’s the fewest number of players drafted from a major program over a two-year period?  Does this tell us anything about Rich’s cupboard that we didn’t already know?  Obviously, it was bare but was it far worse than people realize compared with other major programs?

--Matt

U-M/1995

Yikes. A quick combing of Michigan's roster comes up with the following potential 2010 draftees outside of Graham:

• Greg Mathews: maybe a late pick? He doesn't have the speed to go very high.
• Minor/Brown: it's too early to tell with either but both have the raw physical ability to be drafted somewhere decent. One seems like a first day pick with the other going later.
• Ortmann/Moosman: probably not drafted.
• Stevie Brown: Lions first rounder.
• Zoltan The Inconceivable: likely to be the first punter off the board, whenever that happens.

So, yeah, it's Brandon Graham, a couple running backs, and the space punter. I don't know what the fewest number of players drafted from a power program over a two-year period is, but that's probably not the right question. The right question is "how many teams with like one high NFL draft pick and three or four mid-round picks are any good?" and the answer is "none, but there are plenty that didn't go 3-9." 

This following one concerns variance, as discussed in the earlier post on Gladwell and basketball and Carr and the non-scoring offense. It's long, so I've chosen to respond after each paragraph. Though this looks fisk-y, it's not intended to be confrontational.

Dear Brian,

Your recent blog entry, detailing variance, risk versus reward, defense, offense and modern versus older systems, beginning with a basketball analogy, seems correct, but I have some issues. Your presumption seems to be that solid defense allows for a brute strength, low variance offensive strategy, in the style of Bo, and likewise with Carr. At the same time, however, you insinuate that a slow, grinding offense that keeps the other teams’ offenses off the field is of a critical nature towards that end.

I was not entirely clear about my thinking here. I do think that a really, really good defense allows for that sort of offensive strategy, and more specifically makes the run-run-probably-run-punt style of closing up a game make sense. In that sort of situation you're playing towards your strength.

However, when your defense is mediocre and you have a future NFL player at quarterback, shutting up shop and hoping your mediocre defense comes through is playing to your weakness. Carr did this a lot, if we're expansive about the word "mediocre".

As far as what sort of offense you want at the end of the game, yes, the sort of offense that can grind out a first down is nice to have, but if you don't have that offense—and not many do when they opposition is selling out like mad—you're doing yourself a disservice. There are specific situations where grinding it and punting makes sense, but none of them come with more than two minutes on the clock.

This is reasonable, as you can’t rely on a small lead and a low scoring game if you can’t keep the other team from scoring. The problem, however, comes in making the assumption that defense can’t be, or at least wasn’t, considered a weapon. Absolutely, using a prevent defense, clogging the running lanes, and keeping opposing offenses to short, clock eating runs between the tackles works towards that end. But what of the Michigan defenses through the years, especially in the early Carr era, that actually produced more variance, not less? Sacks, fumbles, and interceptions all increase variance in a game. Sudden turn-overs and backward yards are not supposed to happen on an offensive possession. I would say that in as much as a thundering, slow moving, ground based offense is designed to reduce variance, keep games simple and allow dominant talent to win out; the same strategy of good fundamentals (tackling, stripping the ball, pass coverage) has the exact opposite effect, creating lots of variance and unexpected.

Your definition of "good fundamentals" on defense varies from mine. When I think of good fundamentals, I think of a two-deep shell, minimal blitzing, and conservative strategies. Bend but don't break sort of things. A defense heavy on the blitzing and light on deep safeties is more prone to wild swings. And many of the things you cite as good fundamentals are zero-cost activities from a strategic standpoint: tackling, forcing fumbles, etc.

It seems that you’re positing that the more an offense scores, the more variable and therefore less predictable a game becomes. I think that’s the exact opposite of the truth. Offenses are supposed to score. To assume they will do ANYTHING but that is fallacy. I think the variance comes in when they fail to. Therefore, I don’t think that Bo’s and Lloyd’s game plans were low variance at all. I believe they simply tried to keep the variance, the sudden swinging changes, to one side of the ball. After all, if your defense FAILS to produce variance, the worst that will happen is the other team will score. That can be recovered.  If your offense does produce variance, then the worst that will happen is you will lose your chance to score back. You can’t get that back.

This wasn't what I was getting at, but it wasn't the opposite of it either. What I was trying to say was this: all other things being equal, I'd rather Michigan play a game where both teams have sixteen possessions than eight. (Assuming that they don't suck, of course.) Michigan's more likely to come out on top in that situation. The way Michigan played under Lloyd, however, seems like it lent itself to a lot of long drives on both sides of the ball and generally depressed the number of possessions.

Simplified – You’re saying that offenses produce variance by moving quickly, scoring. As talent entropy occurs, this is harder and harder to stop, and so Bo and Lloyd saw their wins weaken, because their goals were to reduce variance. I believe that defenses produce variance by preventing scoring, and scoring on defense. We saw less success against higher level and middling teams in the last few years because talent entropy, and the coinciding spread of more complex, harder to stop offenses, has leveled the playing field, reducing defensive variance.

Different song, same title.

Thank you,

Eric Fischer

Okay, to properly address this we need to bring in variance's buddy: expectation. In layman's terms, expectation is the average of all expected outcomes. When you roll a die the expectation is 3.5. When you kick an extra point the expectation is 0.98. Variance is a measure of the average difference between trials. I could kick up the variance of the dice roll by turning 1 into –101 and 6 into 106 without affecting the expectation. I could kick it down by weighting it so that 3 and 4 came up twice as often as other rolls.

If you expect to win a game, variance is your enemy. I'm going to borrow some graphs from the excellent Advanced NFL Stats to demonstrate:

So here we've got two teams with the same variance in their play, one of which is a touchdown favorite. The underdog has about a 31% chance of winning.

Now the underdog has gone mad, probably going for it on fourth down a lot, inventing and deploying something called HELICOPTER PUNTING, and trying to block every extra point. They get blown out a lot more but also win more: 35% of the time.

This effect is powerful enough to overcome reductions in expectation:

But this time the underdog’s average is reduced from 17 to 16. The increase in variance still results in a slightly better chance of winning despite its overall reduction in average points scored. In this case, it's 33.2% for the underdog.

And it's the same for the favorite and reducing their variance: sometimes it's worth reducing expectation to get it, but only in certain situations and when you're a considerable favorite. In Bo's time, Michigan was a considerable favorite much more often and the game lent itself to low-variance moves: a 40-yard punt is much more valuable in an era when ten points is a potentially game-winning number.

Anyway, to the assertion above: modern offenses have more variance to them* because they score more. Don't lose sight of expectation here: Missouri had a lot of variance in their scores but that was because they averaged 42 points a game. Michigan had far less but they were averaging 20.

Offenses that do this quickly are actually more predictable because they get in more trials. Moving fast without sacrificing expectation is advantageous to the better team, which is why Oklahoma was in zero even halfway close games against the Big 12 rabble. (Texas is not rabble, obviously.)

Defenses reduce variance by, you know, having safeties that can tackle. The very best defenses are low variance because all of the outcomes have the same result for the opposition: shame and humiliation. In that situation, punting your ass off makes sense, because you're a big favorite, you're not giving the opponent much of an opportunity and you're reducing variance in a way that helps your overall chances of winning. The main problem with Michigan's defense over the last few years has been their suckiness, which by the way increases variance as your defense falls to a point where opponents can drive the field on them regularly.

I always go back here to the end of the 2005 Ohio State game: Michigan has a two point lead and drives down to the Ohio State 40. Facing third and ten, they run a wide receiver screen for six yards, and then punt on fourth and four from the 34, gaining 15 yards. Ohio State promptly drives the field for the winning touchdown. This came after a Henne-demanded fourth-and-short conversion on Michigan's 40 that led to an apparently-clinching field goal, and was interpreted by yrs truly as a panicked reversion to base instincts from another time.

*(The variance of something that's always zero is zero and it's not much higher for something that's almost always zero. As offenses move towards 50/50 efficiency the variance increases, but in a world like the 54-51 game against Northwestern the variance is low because everyone's always scoring touchdowns. An even distribution of probabilities is always more unpredictable than a set where most of the events are drawn to one or two outcomes.)

Sort punchy posts today, I guess, but a reminder: softball takes on Alabama in their WCWS opener tonight at seven EST on ESPN. Varsity Blue has a thorough preview.

Cupcake specifics:

"We appreciate the efforts made by both UMass and Bowling Green to accept one-year contracts to play at The Big House," Athletic Director Bill Martin said in a statement issued by U-M. "The scheduling landscape is becoming more difficult and we discussed the open dates with a number of institutions and want to thank them for their time and effort during this process."

The Wolverines will welcome UMass to Michigan Stadium Sept. 18, a week after traveling to Notre Dame. U-M will then host Bowling Green Sept. 25 before opening Big Ten play Oct. 2 at Indiana. The Maize and Blue still have yet to fill the final open date (Sept. 4) on their schedule.

As you might remember, UMass was rumored to be the newly-renovated Michigan Stadium opener, but this will not happen. Martin on that date:

"We are working hard to line up an opponent for the opening game in renovated Michigan Stadium," Martin said. "We hope to have an announcement in the near future."

The implication behind Michigan's strenuous denials that UMass would be the opener was that they'd try to find an opponent that was appealing, or at least moderately interesting. We'll see if that comes to pass.

John Calipari, folks:

UK announced Tuesday that three scholarship players will not be back on the team next season.

The three were A.J. Stewart, Donald Williams and Jared Carter.

Each of them has been told to play nice or else and has quotes thanking everyone for the opportunities, but privately they must be seething. By the numbers this is vastly worse than the Alabama stuff, as Calipari signed the class in the full knowledge he'd have to boot almost a quarter of his team to do it, without the luxury of medical redshirts. And he's not even done: if Jodie Meeks comes back and there are no academic issues, two more guys will have to get show the door. It's indefensible. Kentucky should be ashamed they allowed it to happen.

Meanwhile, a walk-on was taking about a scholarship with Gillespie and then got the cold shoulder. This doesn't come close to the level of the departed above since the player didn't come to UK under the impression he could spend four years there and end up with a degree, but the manner in which it was handled is revealing. The JCCW on that:

Time for tweeting? Check. Time to give a good kid, a lifetime Kentucky fan and Kentucky native, the common courtesy of telling him he's not needed in person? Or even over the phone? No dice.

These guys got cut so Calipari could cram his five-member recruiting class, which will no doubt feature a number of one-and-dones, on campus, and the idea of the "student-athlete" dies a little more. Calipari's now two for two on abandoning schools just as they get nailed with major sanctions for activities that—like Steve Fisher—the headman didn't know about because he didn't want to. Add in his record as an assistant at Kansas and Pittsburgh and Calipari has been at four schools, all of which have been hit with major infractions stemming from his time there. (Here's the NCAA database for these things; unfortunately it's impervious to links.)

Yeah, John Calipari had no knowledge of (probably) Derrick Rose's fraudulent test score, but that's sort of the point: he didn't know. And he didn't know Marcus Camby and the agent blah blah blah. He's not an idiot and neither are the people at Kentucky. And neither is the public. We're all terribly cynical now.

I find this stuff hugely depressing. Calipari can't take his two recruits and coach the guys he's got and wait a single year to graduate some kids, he's got to boot upstanding players off the team now so he can win now because that's just what he does, and the Kentucky administration just watches. All Kentucky has to do is wait and they'll have their full compliment of NBA-focused players who regard school as a nuisance and Kentucky as a marginally preferable alternative to Europe. Not even that's good enough.

I wonder about people who don't care about anything past the final score, don't care how that stuff goes down. I'd hate to be the guy behind A Sea of Blue right now, as he's not one of these people:

What he has done is effectively turn UK into an NBA franchise, and while that might be good for wins and losses and national championships, it isn't going to be welcomed everywhere.  Some people are going to be very upset with how this is going down, and they have every right to be.  UK has historically honored its scholarships, and has only rarely (if ever) done what is going on right now -- forcing players to transfer in order to make room under the "scholarship cap."

He excuses this behavior in two ways: blaming the athletic director for letting it happen and citing the massive contract Calipari signed, which "demands immediate results."

Why? It doesn't, of course. It demands eventual results, or at least it would if anyone at Kentucky gave a tenth of a crap about the players currently on the team.

I look at the rest of that guy's post, which is filled with halfhearted defenses of Calipari's long and checkered past and just cringe. I'd hate to wake up and see my basketball team filled with mercenaries and the country's biggest asshole on the sideline, winning the hollow victories of the morally bankrupt. What's the point of pretending Kentucky's basketball team is wing of the university anymore?

All right, back to Paskorz: he's a 6'4", 230 pound LB/DE who will play "spinner" or… no. You know what? I give up. Everyone in the comments calls it deathbacker, so deathbacker it is. This guy's playing deathbacker.

GURU RATINGS

Prepare to hear an inordinate amount about Paskorz and his brother during the 2010 and 2011 Notre Dame games: the elder Paskorz is a fullback at Notre Dame.

It appears ESPN rushed to get an evaluation up as soon as Paskorz committed, because as I was assembling this he went from 40 (unrated) to 78 and picked up an evaluation. The most interesting portion of it as regards Michigan:

At the high school level he plays from a "two" point stance and could be considered as an outside linebacker / defensive end hybrid prospect in the right fit. Overall, we feel he will fit best more as a traditional defensive end. He could be asked to play from a "two" point, but seems best suited to play near the line of scrimmage in an attacking fashion.

This is a Greg Robinson recruit like a 5'8" guy with dreads is a Rodriguez recruit. The ESPN evaluation spends a lot of its time talking about that three-point stance and flaws in his technique that come from not being in it. The implication: Michigan got Paskorz's offer out more quickly than a lot of other school because the others were waiting to see if he could put his hand down full-time; Michigan doesn't care.

The rest of the ESPN evaluation is moderately positive, praising his size and frame but saying stuff like "displays ability to" BLANK "but needs to be more consistent." Solid is deployed frequently, and his style of play is dubbed "workmanlike."

About a year ago, Jim Stefani ran down some rising junior tight end prospects and listed Paskorz #2 nationally. That ranking is old, though, doesn't encompass everyone's junior year, and should be taken lightly.

OFFERS

Minnesota, Virginia, and Pitt were Paskorz' other major offers.

STATS

I couldn't find even a whisper of any.

FAKE 40 TIME

Scout's profile lists a 4.74 combine time, which was a combine time and shouldn't be FAKE.

VIDEO

None available.

PREDICTION BASED ON FLIMSY EVIDENCE

And it's flimsy indeed here, with very little to go on other than ESPN's evaluation and yawns from Scout and Rivals. It's so sparse here that I actually watched Paskorz's highlight reel in an effort to form an opinion—usually a futile task for non-tailbacks—and came up with this: yup, that's a two-point stance.

As the younger brother of a highly recruited player, Paskorz is not a sleeper in any way and we should take the rankings at face value. You might be able to argue that he's a better fit at Michigan because of the deathbacker slot and that an internal ranking at M would have him higher—and evidently did than most major schools. But this is another generic three-star with little upward mobility and eh offers. Paskorz's most direct comparable is a Spytek or one of those workmanlike—there's that word again—defensive linemen from the mid-90s.

UPSHOT FOR THE REST OF THE CLASS

When OH LB Antonio Kinard committed the assumption was he was also destined for deathbacker, but he's listed at 200 pounds, not Paskorz's 230. With Michigan is pursuing a number of other guys—Ken Wilkins, Marcus Rush, Will Gholston (though don't get your hopes up there, I guess)—for the spot Kinard might be regarded as a middle linebacker recruit eventually.

Michigan's picked up a commitment from PA LB/DE Jordan Paskorz. Informative update coming… considerably later. I have to go suck at soccer.