Expectation, variance, and Lloyd Carr's fanatical devotion to grinding death football against lower-level opponents have been frequent topics of conversation on this blog since its inception. Thanks to Malcolm Gladwell's latest piece, you can now find conversations about these concepts anywhere you look, most of them criticizing Gladwell for misunderstandings, misrepresentations, and misanthropy, or at least failing to recognize misanthropy.
The piece is about underdog strategy and cites the full-court press as an underutilized strategy that, like going for it on fourth down, is underutilized because of mass idiocy amongst coaches. Of course, there are plenty of reasons an underdog team might decide against pressing:
- Pressing is energy-intensive and could tire out your moderately useful players to the point where you have to bring in the complete gits you stash on the bench.
- Pressing brings more fouls and could force the moderately useful players to the bench, at which point out come the gits.
- A competent press requires practice time that could otherwise be spent teaching the gits to catch balls with something other than their faces.
And so on and so forth. Gladwell ignores all these drawbacks in favor of the hypothesis that everyone who's ever coached basketball as an underdog and hasn't pressed is an idiot. This is not the Romer paper, which restricted itself to the first quarter to simplify its argument and marshaled indisputable mathematics to make its point. Romer built a case; Gladwell offers up a couple anecdotes, one about some 12-year-old girls coached by an asshat, and spins it into a castle of cotton candy—airy, impressive, and ready to fall over if someone looks at it sideways.
So it's way too cute to be useful. But at least it's spurred a ton of interesting discussion. Smart Football on underdog and… uh… overdog strategies:
More importantly though, Gladwell is actually right in a sense: the press (in basketball at least), is a pretty decent example of an underdog strategy. He fails to recognize that what makes it as a good underdog strategy is also what likely makes it inappropriate for Goliaths -- it is a high risk, high reward, high variance strategy. One reason it works for underdogs may have little to do with how good it is on absolute terms; the fact that there is increased variance by itself has value for underdogs because it might give the underdog a chance of actually winning. On the flipside, however, while a full-time press strategy might increase a Goliath's chance of blowing out an underdog, it also might result in them losing a game they shouldn't.
From my distant perch it appears Lloyd Carr hated variance almost as much as press conferences. 10-7 against Utah. All those grinding games against supposedly inferior foes that ended up too close for comfort. Fourth and short punts or field goals, or fake field goals that are punts that everyone sees coming. This makes perfect sense given Michigan's background and the philosophical environment Carr was brought up in.
That environment? Well, I just edited an article Dan Feldman wrote for Hail To The Victors 2009 that cited one of the more famous passages in John Bacon's Bo's Lasting Lessons, so it is fresh in my mind. In the aftermath of Michigan's 6-6 1984 season, Bo checks out a coaching clinic where a whiz-bang kid is detailing a sophisticated system and, in a moment of weakness, thinks maybe he's got to change:
“Now I have to admit—since I’m being as honest as I can be here—there was a time when I doubted if fundamentals were still enough to produce top-notch football teams,” Schembechler says in the book. “I even wondered if the game had passed me by.
“I’m thinking, Maybe you’ve got to do all those things to win these days. Maybe our approach at Michigan is just too simple to succeed in the modern era. Boy, that was an awful feeling. But after this guy finishes his slide show, someone in the audience raises his hand and asks, “If your defensive schemes are so great, then why did you team give up 400 yards a game last season?”
That question caught Schembechler’s attention. The high school coach’s answer? “We were just a poor tackling team.” That made everything very easy for Schembechler. “I walked out of that auditorium, and I knew what we were going to do: Get back to the basics! Get back to Michigan football! And I was determined that we were going to do it better than anyone else.”
This may have been brilliant in 1985, and brilliant against the poor, huddled masses that comprised Michigan's opponents at the time, but it's fundamentally a variance-hating strategy that presumes better talent. In it are the seeds of Michigan's time-honored failure against Rose Bowl foes, and its recent struggles to put away inferior competition.
When you choose to reduce variance you are usually giving up expectation, especially in football. See the Romer paper. Coaches choose to punt or kick field goals because they're "less risky"—i.e., have low variance—despite the numbers clearly showing they are also "stupid"—i.e., have low expectation.
But there's a catch. In football, actions that have low variance on the micro level can aggregate into a whole that has high variance. Take Michigan's oft-stated priority to control the clock and "keep the other team's offense off the field." You do this by engaging in a lot of long drives consisting of inside runs and short passes. When you run for three yards on two of three plays, your variance is very low. When you pass, three things can happen and two of them are bad: your variance is high.
HOWEVA, running a lot of clock and having long grinding drives reduces the number of possessions in a game, which jacks the variance up. Smart Football again:
Going extreme hurry-up to get as many plays as possible -- other than endurance, I suppose -- is a Goliath strategy: it decreases variance by increasing the number of trials. The chance of getting only heads and no tails in five coin flips is much higher than it is in a hundred -- i.e. the impact of the law of large numbers or regression to the mean. If Oklahoma has significantly more talent, better schemes, and everything else than the underdog, then the more plays it run the more likely it is to exhibit its raw dominance over the underdog; the underdog is less likely to "steal" a few good plays and get the heck out of dodge.
Every second that ticks off the clock between plays is aid and comfort to the underdog. This is where the Gladwell article truly breaks down. Pressing is a dumb strategy for underdogs because it—though high variance in the micro—is not necessarily high variance in the macro. Pressing can increase the number of possessions and thereby give the dominant team more of an opportunity to show that dominance via thunderous open-court dunks. Yes, this makes Gladwell referencing Rick Pitino's loaded, national-title-winning Kentucky team as an exemplar of an underdog wisely using a press ironic in the extreme.
When Bo was coaching the relative skill level of most opponents and the nature of the game at that point—low-sophistication passing, low-scoring—made his strategy a good one. Leading 12-6 in the fourth is an entirely different level of comfort than leading 30-24. Back in the day a turnover that sets up an easy opponent score was way more dangerous than punting one too many times. Bo's tactics were sound for his day, and for certain later days. In 1997 Michigan had a world-crushing defense that turned the tactical clock back to 1972, and Bo's philosophy worked just fine.
In other instances, it did not. Michigan fans were intimately familiar with Carr's late game strategy, which strove for low variance above all when Michigan was ahead: run, run, probably run, punt, play D. This, from an article by Vijay of IBFC in the 2007 edition of HTTV, is how well that worked:
Michigan entered 18 games over that period of time [2000-2005] with a lead smaller than ten points and went 8-10 in those games. They were under .500 when entering the fourth with a small lead! When tied or facing a similarly small deficit, Michigan was 6-1. In all games in which Michigan trailed by any margin they were 8-8. Michigan had a better chance of winning if trailing by any margin entering the fourth than they did if leading slightly.
The way Michigan approached its Goliath strategy was fundamentally broken as the Carr era waned. In the new era of modern football the Goliath strategy is clear: score, score again, score again, and score some more.
Uh… I guess. A couple other points:
- Longtime readers will find this familiar. Last year I wrote something along the same lines except with more references to Coach and this chestnut:
Lloyd Carr thought deception and trickery had their place in football, and that place was Northwestern.
The piece holds up pretty well, so if you're confused as to what I'm getting at or want more context I recommend it.
Though Dantonio and Bielema have the most boring offenses in the Big Ten, they'll be fascinating tactically over the next few years. Their clock-killing, suddenly-unusual offenses are excellent underdog systems now, but are terrible ideas for favorites who are unlikely to have killer defenses. I foresee perpetually decent but unthreatening teams for both unless Bielema is just as bad a coach as last year suggests.
Woo for 1600 words at 4PM Friday!