stat wonk

The voting is finished and it's time to present this year's College Football Blogger Awards. Where possible, last year's winner, ineligible to win this year, will be presenting the award to this year's winner. Please check in at Rocky Top Talk and EDSBS for a schedule of all the awards to be presented over the next two days.

The next award presentation - for "Best Community" - will be noonish at Burnt Orange Nation

AWARD PRESENTATION: "BEST POST OF THE YEAR: Analysis"

2007 Nominees

Fiascotime Post-Mortem, by MGoBlog

Playing the Numbers Game: The BCS Computers, by Burnt Orange Nation diarist, Horn Brain

Requiem for a Triple Option, by Sunday Morning Quarterback

Simple Complexity, by Blue Gray Sky

The Divide Route in the Multiple Smash Concept, by Smart Football

2007: Pythagorean Wins Revisited, by Roll Bama Roll

2007 WINNERS

2007 RUNNER UP: "Requiem for a Triple Option," by Sunday Morning Quarterback

This could have been any of a half-dozen different SMQB posts tackling anything from statistics to the origin of the term "single wing"; I'm probably not spoiling much to tell you that SMQB won the blog version of this award in an epic landslide. As it is, there is another in this category...

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

2007 WINNER: "The Divide Route in the Multiple Smash Concept" by Smart Football.

Hey, this is kind of the reason we made jury awards for certain categories. Smart Football is low traffic, posts every month or two, and exists only in the feed readers of wonky football obsessives who would like things explained to them very slowly by trained professionals in an effort to become one with their dork. Smart Football does this with aplomb, and though its posting rate and overall traffic made us pretty sure it wasn't going to win Best Analysis -- these things are usually go with what you know -- each post is a little jewel of clarity.

An excerpt:

I refer here to the "Smash concept" or the "Smash route." Both refer to a two-man combination with the outside receiver on a 6 yard hitch and the inside receiver on a 12 yard corner route. Some coaches and teams go further and actually refer to either the corner route or the hitch route as a "smash" route. Again, "smash" to me is the combination - i.e. the concept - rather than any individual route.

In any event, the quarterback has a progression read: (1) corner, (2) hitch underneath. In his progression read he will "key" the cornerback: If the cornerback sinks back to stop the corner route, throw the hitch; if he comes up for the hitch, throw the corner. The best way to describe this to a QB is that you have a progression read and you "read" your receivers. You simply "progress" from one to two. In doing this though you have to understand what guys you are "keying," as their reactions should determine your progression. A Quarterback must understand defenses and defender reactions, but at the same time there is no telling where those 11 guys on defense will go, and as long as he knows where his receivers are and if the QB and the receivers are all on the same page we can run a successful play. We tell him his general rule is to throw the corner route until they take it away (though by gameplan or defense you can tell him to always throw the hitch until they come up for it).

This is the basic explanation; things get more big-play oriented and complicated as they progress, but Smart Football never wanders off into seriously incomprehensible jargon. By the end you feel like you have a handle on an important facet of beating zone coverage, down to the slight adjustments in each receiver's route, and at no point are you overwhelmed. Clearly explaining a difficult, obscure concept is a terribly hard thing to do, so we give out awards for it. Here is an award.

First in this year's occasional digressions into NBA blogging. Complaints about topic choice will be considered and then dismissed.

When the Bulls decided to spend every last bit of their cap space on an older version of Tyson Chandler, the Pistons were declared dead. And at first blush, the departure of Ben Wallace has hurt the Pistons severely. The Pistons are on pace for 11 fewer wins this year than they had last year. Opponents average 92.2 points per game, up from 90.2 -- a major leap when the Pistons' overall point differential is just +4.1. But how much of the fall is due to the loss of Wallace? And was resigning him worth it?

We must first establish the players receiving Wallace's minutes. (All numbers from 82games: 05-06, 06-07)

Lindsay Hunter and part of Flip Murray have picked up Arroyo's minutes. Delfino and the rest of Murray pick up Evans' minutes. Four post guys receive more time: Webber, Mohammed, Maxiell, and Davis. For convenience's sake we'll apply Davis's minutes to the deficit in Rasheed and McDyess' minutes; Wallace's replacement is the three-headed Webbmomax. Webbmomax plays 74% of the Pistons minutes; Wallace played 73% a year ago.

So what do the Webbmomax Pistons do worse than the Wallace Pistons?

Rebounding? No. Despite losing a man widely regarded as the league's premiere rebounder, Detroit still gathers 69% of its opponents' misses and 30% of its own misses.

Blocks? No. Last year Detroit blocked 7.6% of its opponent's shots. This year it's 7.5%.

Causing turnovers? Last year opponents turned it over on 15% of possessions; this year 16%.

Maybe Wallace had some sort of weird anti-shooting mojo that didn't show up in the blocks? No. Detroit's FG% and eFG% defense percentages are actually better this year (44.3%, 47.6%) than they were last year (45.2%, 47.8%). It should be noted that approximately 1% of opponent's field goals have migrated from jumpers to post moves and dunks, though strangely enough opponents are doing worse at them.

The only thing that jumps out as any different is opposition free throw attempts. This year opponents get 25.3 per game; last year they got only 20.5. That huge discrepancy, plus that extra possession per game, is the only reason the Pistons' D is any worse this year than it was a year ago.

The next question: is that because of Wallace's departure? Survey says yes. Last year Wallace played 73% of the time and picked up 164 fouls, only two per game. The three headed monster that replaces Ben's production is on pace to finish this year with 354 fouls between them. 96% of the excess fouls this year's Pistons are piling up are from Webbmomax.

How many points is this worth per game? If you accept the standard 0.44 multiplier when converting FTAs to FGAs -- and-ones, techs, and the like make it a bit lower than the 0.5 you might expect -- this year Detroit is converting 2.1 FGAs per game into free throws. The free throws are worth 3.55 points; the hypothetical FGAs they replace would be worth 2.01. Wallace's marginal contribution appears to be 1.54 points per game, though it does appear that the more aggressive post defense has created a few more missed shots and a few more turnovers.

Perhaps the revelation that losing Ben Wallace is a detriment to the Piston defense is not Nobel-prize worthy, but what is interesting is where the dropoff is. Rebounding: useless. Blocks: useless (with the stipulation that this is the 05-06 Wallace we're talking about. Previous editions averaged over 3 blocks per game). Contribution to FG defense and opponent turnovers? Negligible or negative. Ben Wallace's main marketable skill is the ability to keep opposing shooters off the foul line.

(A side note: this should be further evidence against David Berri's preposterous "Wages of Wins," which you may remember from an earlier Hollinger-Berri smackdown. As Jonathan Weiler explains:

Let's go back to Rodman. A key question that, as far as I can tell, is not answered by the WOW authors is this: what is the difference between the number of rebounds the Bulls would have secured with a replacement for Rodman compared to the number they did secure with Rodman? The answer, I am sure is "fewer." But, how many fewer? If a Rodman replacement snared seven rebounds a game, instead of 15, can we really say that the switch cost the team eight rebounds a game. I don't think so, and neither, I suspect, do the authors. ...

I agree that rebounds contribute to winning, but to know how much Rodman contributed to winning, we need to know more than how many rebounds he got and we need to know more than how many rebounds he got relative to players on other teams.. What we really need to know is how many rebounds his own team got that it would not otherwise have gotten if he weren't there. Do I have a clue how to figure that out? Absolutely not. But, I am convinced that that is the key question and that it has not been answered by WOW (and, in fairness, other similar systems).

So how many rebounds did Ben Wallace get that a mediocre version of himself, an undersized version of himself, and a crippled 33 year old power forward could not? This analysis implies, shockingly, that the answer is "none at all." And yet WOW treats each rebound like it's a diamond. As Dan Rosenbaum points out, the whole thing would be wildly off if not for an enormous team defense fudge factor.)

So... is Ben Wallace worth it?

Absolutely not. His yearly numbers are in steady decline and he's in the first year of a four year contract as a 32 year old. And his offensive deficiencies are severe enough that his net contribution is minimal. Webbmomax has hit 178 of 289 free throws this year; Ben Wallace and his 0.407 FT% would have hit 55 fewer and the Pistons would be scoring 0.9 fewer points a game. That's most of his measurable defensive contribution without even considering the rest of his limited offensive game.

At this point in his career, Wallace is an average-at-best starting center who is paid like Kevin Garnett.

So why are the Pistons losing so much more than they did a year ago?

• Injuries to Billups and Rashee
  d. The Billups injury especially, as it ushered in the horrifying Flip Murray Era. The plus-minus numbers for the Pistons' two backup guards are atrocious. It's Billups who's irreplaceable, and it's Billups who Dumars will break the bank for this offseason. If I could put in a request for an MLE point guard, perhaps in exchange for Nazr Mohammed?
• Average luck. I don't have any numbers on this, but when you go 64-18 you're a lucky team. I'd be willing to wager that the Pistons' record in close and OT games is markedly worse than it was a year ago.
• Familiarity. There was disconnect between the Pistons' regular-season performance and their playoff struggles, but the return to earth started before that. Excluding the four year-end games after the Pistons secured the best record in the league, the team started 37-5 and finished 26-10. Then came the playoffs... ugly. What happened? No one can really be sure, but my belief is the Pistons blitzkrieg caught the league off guard. When Flip Saunders replaced cranky old Larry Brown the Pistons underwent and instant transformation from a pack of ugly grinders into a gorgeous, efficient offensive basketball team. By the time anyone knew what was happening, Chauncey Billups was at Moscow's doorstep. Now that teams have the Pistons scouted, the going is tougher.
• Flip Murray is awful at everything.

For context on the below, see Part I for an extensive discussion of what exactly is going on here (warning: math) and Part II for some examples of why I think this is a useful exercise. Note those posts are from last year. Which is 2005.

(Performance note: generating these graphs is a dynamic process, so they can take several seconds if no one else has looked at the requested data recently. Cached ones should come up immediately.)

First: Third down efficiency. The thick line in the center is the NCAA average (e.g., approximately 68% of third and ones were converted last year). There is a second line that represents an individual team's third down efficiency. Where there is a gap between the lines that gap is filled in with either red or green depending on whether it is "good" or "bad". Being above the line is good for offenses--you convert more often. Being above the line is bad for defenses--you are converted upon more often. You want to see a lot of green in these graphs.

Second: Third down distance distribution. Again, the line in the center is the NCAA average and the thinner line is the individual team's. Green is just "above"; red just "below," since there's no clear distinction on good or bad based solely on what side of the line you're on.

Third: the raw numbers. The following graph shows the underlying data used to construct the first two. Each bar represents one yard line. Blue segments are failed conversions. Red segments resulted in first downs.

(A note on reproducing these graphs: feel free. Right click and "Save As" to get a static copy that won't break if I decide to change the URL... which I might. Please drop a link. Also: if the idea of maize and blue on your site is revolting, you can give me two other colors (specified in hex--ie, #A30924--, please) and get pretty team-color-appropriate ones.)

Zone Gives, Zone Takes

This fall Michigan's run game took a radical departure from the pull-, iso-, and draw-heavy early 2000s. In the place of plays like "argh not another second and long draw," Michigan implemented an Iowa-like zone scheme. At first glance the results were encouraging: Michigan vaulted from 44th to 21st in total rushing. Yards per carry (discounting QB runs and kneeldowns) shot up from 4.3 to 4.9. This despite Michigan featuring an unbalanced offense that ran 56% of the time. So bully for all that.

But there was one major issue that our stretch-play-mad offense developed during 2006. UFR aficionados already know this deep in the marrow of their bones, but now I can put it in attractive graph format.

Here is the Michigan offense's third-down conversion rate by distance in 2005:

(How to read the above: the thick line is the NCAA average. The thinner line is Michigan's performance. You want this to be above the line for an offense; green is good.)

Poor performance on third and long was counterbalanced by a surprising competence on third and short.

Here is the same graph for 2006:

(Disclaimer: this data is sparse and may not breach the threshold of official statistical significance. I don't know. I'm not a statistician. I do know that this is all the data we have and that it can be useful when coupled with previous observation.)

It does not take a jeweler's eye to see that the graphs are near inverses of each other, even down to the weird aberration at third and 11-13. The improved performance on passing downs is easily explained by an improved receiving corps and quarterback. But the startling regression on short yardage is inexplicable given the improvement the zone brought to Michigan's running game.

Unless, that is, you watched Michigan roll out three wideouts and run a stretch play into nine penetrating guys far too many times. Which, uh, I did:

Why do we suck on third and short?

I dunno. We were actually really good last year, but it seems you put us in third and short this year and we run out three wideouts and run a stretch play into nine guys. This doesn't work so good, evidently. What's wrong with lining up in a big set and cramming it down their throats? We have the personnel for that sort of thing.

Hopefully, Michigan will identify this ugly tendency and stop actively trying to turn third and short into punts. In Michigan's transition year perhaps it made sense to focus on the stretch to the exclusion of all else, but now that we've established hadouken, we should try to work on some shoryuken for situations when the objective is closer at hand.

AMFB

Remember 2005? (It's okay to say no.) Pat "Moonwalk" Massey and an occasionally interested Gabe Watson were the defensive tackles. David Harris and yuck were the linebackers. Alan Branch toiled out of position at defensive end. The short yardage numbers were ugly:

Yeah... not so much this year. Anyone doubting Alan Branch in even the slightest tiny way should look at this, tremble, and shut up:

You can see 6'6", 330 of angry New Mexican hauling the tail end of that graph down like a black hole in spacetime. That's Alan Branch. 33 percent! On third and one! Six of eighteen! SIX OF EIGHTEEN!

!!!

We are all going to miss last year's defense very much. Except for the secondary, which I fear I won't miss at all.

So... a week or so ago EDSBS and I paged through the Rivals archives and extracted per-class scholarship averages for each BCS school and conference in the country. Quickie conclusions: the SEC is a bit sketchier than the Big Ten, as asserted by Jim Delaney, but not nearly as much as this year's enormous seven-to-eight scholarship disparity implies. Still, an attempt to provide some ethical and statistical context follows.

How big is this gap? Over the last six years, the Big Ten has handed out 22 scholarships a year and the SEC 25. This doesn't sound like a huge difference, but it depends on your perspective. In hockey, there doesn't seem to be much of a gap between a player with an 88% save percentage and a 92% save percentage, but rephrased it as "player A lets in 50% more goals than player B" and the gap is brought into relief. Similarly, 25 and 22 seem close enough but flip it around: if we ballpark the number of redshirts at 50%, a team retaining 100% of its players uses 19 scholarships a year. Every year the average SEC team experiences double the attrition of the average Big Ten team.

Is this scholarship gap necessarily a sign of poor moral fiber? Not necessarily. There are two different arguments getting conflated into one here:

1. SEC classes are overrated on Signing Day and during the media blitz that follows because their increased attrition rate -- something the numbers show is indisputable -- allows them to sign a bunch more players who will never make an on-field contribution.
2. The SEC doesn't care about football people. [/Kanye]

Argument #1, as noted, can be accepted as a given. Argument #2 is murkier and requires us to consider...

What exactly are the ethical obligations schools have here? The conventional wisdom from rabble-rousing sportswriters and tut-tutting moral arbiters is that College Sports Is Corrupt And Evil for even thinking about permitting players who are either dumb or heinously underserved by their schools to breach the local ivy-covered educational edifice. And there is a point in there somewhere: bluntly, most football players are not good at school and very few of them would be in college at all if they weren't huge and fast.

But it's hard to see how anyone's life is improved by strenuously demanding Stanford-level academics of 340-pound maulers from rural Mississippi. Stakeholder by stakeholder:

• 340-pound maulers. What's the alternative for these guys? Most of them will never sniff the NFL but it's a shame to take away their shot at it for an irrelevancy. Even if their education is remedial, that's likely better than they'd have otherwise.
• Normies. A few extra kids in big lectures dragging the curve down doesn't negatively affect the rest of the student body except in the smallest and most incremental way, and even that is offset by the contribution a healthy athletic department makes to the overall life of a collegiate campus.
• The University Ideal. Athletes' altered admissions standards don't necessarily compromise the university's academic purity. There already exist certain segments of the student population for whom the ability to put together a five-page essay or solve a differential equation is irrelevant: music and art students are admitted primarily on their talents in their field of choice, not arbitrary standards for performance on a standardized test. Essentially vocational programs already exist: a journalism major's classes are of secondary importance to his work on the school paper; an art student's GPA is secondary to his portfolio.

  Besides, the idea of a cloistered, ivy-covered thing where you learn all about like Kant and Hegel and Thoreau in intimate seminars went away a long time ago in favor of enormous diploma mills. Large state universities -- where virtually every sporting program big enough to be corrupt lives -- are more expensive vocational schools these days for the vast majority of students. (Private schools, being private, can do whatever the eff they please.)

Where the ethical dodginess comes in is when the 340-pound mauler's education is less remedial and more nonexistent. The latest Sports Illustrated has an article on Greg Oden that details his courseload: a five-credit Sociology 101 course, a five-credit History of Rock And Roll course, and two credits for being a basketball player. It's hard to work up any outrage about next year's top pick in the NBA draft getting shorted in his education, but how many players with far more uncertain futures are getting educations in avoiding education at schools around the country?* The general feeling is lots.

This is because the system has a disconnect. It rewards teams for keeping players eligible, not for educating them. It encourages Harrick-like "how many points for a three-pointer?" classes, academic... er... tutors, and History of Greg Oden majors because the only people judging how educated our mauler is have a deep conflict of interest. The scary idea is not how many kids flunk out but how many "graduate."

It is an article of faith around these parts that the SEC is the worst offender in this department. Anecdotes filter up: former running back Max Martin got in the doghouse because he didn't go to class. When someone in the department asked him why, he replied that he didn't know he had to, since all of his buddies down south didn't. (Later, Martin transferred to Alabama; the coaches at the time reportedly asked if he had been arrested for any felonies, then hung up, thorough background check completed. He lasted a semester.) Varsity athletes in non-revenue sports relate similar tales on recruiting trips. Anecdotes prove nothing, though, proving nothing, and schools all over the place have issues. (Clem Haskins at Minnesota sticks out.)

All we have right now are some numbers that take a look from 10,000 feet up that reveal something indistinct. Even if we drilled further down into these numbers, they would only tell us the lesser half of the story, and the questions about "what about the guys who remain eligible?" would remain unanswered.

*(Referencing Ohio State here is sure to cock an eyebrow or two since I am an avowed Michigan fan. The intent is not to single out OSU as an exemplar of bad behavior; the SI article provided a rare concrete look at the courseload of a star athlete.)

(Named after the (apparently) famous Price Is Right game. Ask Orson.)

Hello! More exploration of the sketchy ethics of college football recruiting. Orson Swindle and I gathered the last six years of classes for every BCS team and found average-sized recruiting classes for each team. For comparison's sake, a team with 100% retention that never redshirts or recruits JUCOs or transfers should average 21.25 scholarships a year. A team with a 100% retention rate that redshirts everyone but still avoids JUCOs and transfers should average 17. If you figure about half of all freshmen redshirt (figure pulled directly from rear), then 100% retention means a class of slightly over 19 kids.

Disclaimers go here: Numbers are from Rivals. Scout has different numbers that are on average a little bit smaller. 100% retention is impossible and there are many completely kosher reasons for kids to leave a college football program before their eligibility expires. JUCOs and transfers were not accounted for in these numbers and obviously have a distorting effect (though transfers don't even show up in the rivals DB, so the net effect of a transfer to your school is to make your school look less like an Indonesian ferryman shoving people off the boat into alligator-infested waters.)

I've got the Big Ten, Pac-10, and Big 12. Check EDSBS around noonish for the ACC, Big East, and (drumroll...) SEC.

Our Indonesian Ferryman for the Big 12: Iowa State. An unsurprising second and third are Kansas State, Texas Tech, and their insatiable lust for JUCO blood.

And the Big 12 Mr. Chips Award For Academic Integrity goes to big, bad Texas -- though their number would be but middling in the Big Ten.

Yea, verily the Pac-10 is a conference of extremes.

Our Indonesian Ferryman for the Pac 10 is Oregon State with a staggering 29 scholarships handed out per class and a whopping 134 recruits over the past four years. Oregon State is the national Indonesian Ferryman and the sketchiest program in all the land.

Meanwhile, Stanford is puttering along at 18 scholarships per year. They win the Mr. Chips Award For Academic Integrity for the Pac-10 and the nation.

Our Indonesian Ferryman for the Big Ten: Michigan State. To be fair to the Spartans, their total would be good for but fourth in the Pac-10 and a piddling sixth in the Big 12.

Raise your hand if you thought Northwestern would win the Mr. Chips Award for the Big Ten. That's all of you, good. Now raise your hand if you thought Ohio State would come in second. Anyone? Bueller? Bueller? Ohio State is increasingly irritating: first they go and win against us all the time, reducing my ability to make fun of them to critically low levels, then they go and stop getting in trouble altogether -- a third string kicker selling weed just doesn't cut it, even if his last name is a popular synonym for semen. What's a blogger to do?

And there's one relevant team floating out there not in any conference...

To its credit, Notre Dame retains like a mother. Either that or Willingham was so bad at recruiting they left large numbers of scholarships open.

EDSBS will have has the ACC, Big East, and SEC up... how do they stack up? Stay tuned! Spay or neuter your dogs and such!