Strategery round-up – 6/21/2010

Good links all related to football strategy, though we begin with a video of Gus Malzahn’s Auburn O, via Offensive Musings:

Defending the bunch. If a defense plays a lot of man coverage, you can bet that the offense (if they have any sense, anyway) will quickly start using “bunch” or “compressed” formations. Anyone who has ever played backyard football can give the answer: it’s much easier to get open if your defender can get “screened” by congestion of some sort — either your teammate running a “legal screen” (versus, ahem, an illegal pick which no one ever does, right?) or even some cluster of receivers and defenders. Defenses, not to outdone by such offensive wizardry, have responses, summed up well in posts by RUNCODHIT and Blitzology.

Unsurprisingly, discipline is a key factor. Blitzology covers some mechanics, while RUNCODHIT adds some background:

[Y]ou can’t run press-man on both WRs[;] alignment won’t allow you to. Also, to run straight man against reduced splits is suicide. The offense will pick you off and open-up a WR to the inside or outside. Because of this threat, defenses have to stay in pure-zone or combo-man coverage.


versus the run 3-way [coverage] places the [strong safety] in a position to force the ball inside. The corner is assigned play-pass responsibility, and the [free-safety] is a flat-foot read player . . . . Against the pass the . . . [strong safety] has the first man to the flat. If no one attacks it, he sinks under the first WR outside. The corner[back] has the first deep route outside — he is going to [back]pedal on the pass and read the WRs. The FS has the first man deep inside. His technique is essentially the same as the corners’. If a deep receiver does not show in or out, then they play a “zone it” technique and help their partner.

3-way coverage

Bonus: Check out RUNCODHIT on “Pattern Reading vs. Zone Dropping” and Blitzology’s series on attacking BOB or Big on Big pass protection. (I’ve described the principles of this protection here.) Series parts one, two, three, and BOB vs. the 3-4 defense.

Think you have what it takes to be an NFL guy? Check out this Slate article on the work ethic of NFL coaches. The answer — it’s about managing people, as much as it is about strategizing and ideas:

What exactly does a head coach do for 23 hours every day? . . . Imagine telling George Halas that he should have worked 20-hour days. He would have laughed you out of his office, then gone back to inventing the T-formation. No matter how many variations on the spread offense you come up with, it’s still the spread offense, not Fermat’s Last Theorem. . . . The guy with the biggest whistle has a fleet of coordinators and position coaches that handle all the grunt work, from conditioning to game-planning to skill-training. . . . Instead, the coach functions as a sort of CEO, coordinating large-scale strategic planning while ensuring all members of his organization perform competently. Viewed through that lens, this endemic insomnia shouldn’t come as a surprise. After all, CEOs fetishize waking up early just as much as football coaches. . . .

Screenery strategery. If you’re going to spread the ball out or throw the ball at all, day one is usually spent working on the basics of passing: timing, quarterback drops, rhythm, catching, and the basic routes. Day two, however, goes to screens, those little gadget plays that, particularly at the lower levels, make being a pass first team really worth it. These impressive little suckers manage a quite impressive trifecta: (1) they are easy to complete (and maybe should be thought of as runs rather than passes), which can build your quarterback’s confidence and allow you to get the ball to your playmakers in space; (2) they are often your best weapon against aggressive, blitzing defenses, which can otherwise overwhelm young players just learning how to throw the ball efficiently; and (3) unlike a lot of passing-related concepts, these make heavy use of misdirection, that great equalizer between teams of greater and lesser talent.

In that vein, two great primers out there are Mike Emendorfer’s UW-Platteville screen presentation and this recent post from Brophy’s blog.

Football and math, oh my. Good post on the basics of “football math” — i.e. who and where do you attack. Here’s a test: Where would you attack in these two situations?


Bonus: See clinic notes here and here from DACOACHMO.

No surprise there: Bill Walsh gets it. Brian from Advanced NFL Stats quotes one of Bill Walsh’s best nuggets on playcalling:

We know that if they don’t blitz one down, they’re going to blitz the next down. Automatically. When you get down in there, every other play. They’ll seldom blitz twice in a row, but they’ll blitz every other down. If we go a series where there haven’t been blitzes on the first two downs, here comes the safety blitz on third down.

Brian has more on the theory behind Walsh’s practical wisdom, and I discussed the subject of randomizing playcalls a few years back:

In Edgar Allan Poe’s story The Purloined Letter, a character recounts a story of a young man who excels at game called “odds and evens,” known somewhat more popularly now as “matching pennies.” The game is a two-strategy version of rock-paper-scissors: Each player secretly turns their coin to heads or tails and then both reveal their choices simultaneously. If the pennies match (both heads or both tails) then one player gets a dollar; if they do not then the other gets the dollar. As told in the story, the young man quickly sizes up his opponents, gains a psychological advantage, and amasses a fortune by outguessing his opponents.

I suppose all playcallers think themselves like the young man, but most are probably more similar to the suckers. But here’s the rub: The suckers could nullify the young man’s psychological advantage. How?

By choosing randomly. If the suckers put no thought into whether they chose heads or tails, they would do better than if they tried their best to out-think him. They would break even — a fantastic result against the world’s greatest matching pennies player, an unnatural genius who, according to the story, would go through lengthy Sherlock Holmsian deductions to determine if his opponent would choose heads or tails, and of course almost always guessed correctly.

This is a breath-taking result: you can nullify anyone’s advantage by picking randomly. But it is also scary — would I be better off picking my plays entirely randomly?

  • Old South

    I posted a response over on EDSBS, but, in general

    1) Randomness cannot succeed in the long run if your playcalling is totally random on every play through all of your available options (obviously).
    2) In certain contexts, randomness can be useful
    3) Randomness, as the Holmes example suggests, benefits the weaker coach
    4) The optimal ratio of random to nonrandom playcalls should increases as the discrepancy between the coaching skill increases
    5) All of this is incredibly contextual

  • Joe H.

    The problem with random playcalling is that it seeks to isolate plays as independent events- of which there IS NO SUCH THING. If behavioral economics have shown anything, it’s that it’s very hard, in fact impossible, to prove independence of events where human activity is involved (and at some point, human activity mucks with an event).
    You could cluster plays based on an effectiveness quotient (like all your screens on the same table or cluster, for instance- you could, and probably would, incorporate down and distance into the probability of selecting certain plays) and then let a random event take over- a system clock or even a dice roll choice base. You can even create different tables in response to scouting different teams- 3-4, 4-3, and personnel. That way, you’ve removed a predictability bias from your call, and you might still get your 3.3 yards. There is the problem of execution, and the fact that your team executes some plays better than others- the system couldn’t be too complex with too many plays.
    There’s a double integral equation in soccer (proprietary, so I don’t have direct access to it) that teams use to calculate effectiveness of overall strategy- it calculates the likelihood of a win in a probability cone, and it would be fun to adapt it and run a series of randomized playcalling instances through it- the problem would be creating enough real cases to where the results produced (and the equation used to describe them) are valid (one tailed, right?). The question is, in the end- does it even out the chance to win (great if you’re not nearly as talented) or give you a distinct advantage (great for everybody, and extremely intriguing for high end teams).

  • Joe H.

    FIgure two Davidson Grads to post on the impossibility of randomness

  • Rob

    Ive thought about randomness and play calls for a while. One place I think it can be used is the pre-scripted plays. For an example, lets say you come into a game with 10 plays you are going to run and are going to use them on the first ten plays that are 1st and 2nd downs between the 20s (3rd down and red zone on either end you may want to run plays in a context). Pregame, randomize the order of those 10 plays.

    I think that works better than scripting an order to those plays.

  • I’m not sure randomness would give an advantage to, say, Nick Saban who can call a great game. In the example, the incredibly good odds and evens player wouldn’t benefit from going random. But his opponent — someone who was outmatched — can eliminate the advantage by going random, which is a powerful tool.

    Also, OldSouth, I don’t think anyone literally would just call everything “random” per play. You’d go random based on the situation — i.e. first and ten you have a selection of base runs, a selection of deep play action, and a selection of short/intermediate passes and you’d randomly call one of those; another situation (say 3rd and 1) would have a similarly random menu of plays that fit that scenario but otherwise make sense.

    I think the edge you’re looking to avoid is more along the lines of individual play risk — i.e. that you call the sweep and they destroy it, or you call the bootleg and they were on top of you — rather than the idea that it’s 3rd and 1 and you might line up in five wides and throw a bomb or put two tight-ends in, etc.

    Indeed, your personnel grouping should match the situation and your menu is going to be limited based on that. Keep in mind too that a lot of plays are called at the line by the quarterback these days, which is something that a defense can game — i.e. make it look like a defender is not in the “box,” sucker a run call, then come up and blow it up, etc. Intelligent randomness could be used to bail out your quarterback’s checks.

    Another point is that it should be random within a coherent system. No one is advocating “randomness” based on some Madden-esque total universe of all plays (run and shoot followed by wishbone followed by wing-T followed by I formation). If your offense works as a system then picking the right play at the right moment shouldn’t matter as much. A lot of the Airraid is built on this, among other offenses.

    In any event, Orson basically had it right on EDSBS: if your playcalling isn’t working, think about going a bit more random.

  • Joe H.

    I think you can minimize the risk by “leveraging” the random factor by excluding plays on your play table that are very low probability for the necessary gain (slant on your 1 yard line, f’r instance).
    I think randomness helps those who have much less talented teams- any chance you get to make the probability of a win as close to 50% and negating another teams advantages, be they athleticism, playcalling, or something else, you need to do it.
    I’m pretty sure the down, distance, and personnel factor into the computer player’s choice in something like NCAA- It’s probably not dynamic or anything like that (processing power is limited), but dynamic randomization of your play selection could parallel it in some ways.
    Just roll the dice, and trust the probability tables to select the right play- it could work. The trick is constructing effective measures of success.

  • TD

    Execution is the key.

  • DrB

    All good articles there

  • Old South’s first point seemed to apply to LSU and Gary Crowton at times last year. A lot of randomness with no sense of unity to his attack style.

  • Mr.Murder

    Auburn’s most dangerous QB was able to run better. One side of the read for their main passer you always play pass or pitch man to the QB run side. Slow as those plays develop you might as well cross blitz inside and by the time those players clear the man asssigned to block them in their gap the play is pretty much developed enough to track. Tackle every fake man, the Qb will stop acting like he might have the ball if you float an end out on him for a few free hits.

    Spoke the safeties on lateral fakes.

  • Mr.Murder

    The link you give on this thread on Walsh study contains a TCU playbook. That link describes a way for force defenders to key run support in language consistent with teaching pattern read coverage as per Saban.

    6) Read the relationship between the blocker and the ball carrier. If the runner is “in phase” with the blocker, attack the blocker low and hard.
    7) If the ball carrier is “out of phase” with the blocker, then we bypass the blocker(escape inside or out) and attack the ball carrier.

    Interesting that a physical key on spatial play development matches guidelines for approaching contact and footwork. It also reinforces the man coverage principle of getting in phase by getting on the pass catcher’s hip to play the ball.

  • White End Tables

    I find it damn weird that demonstrants are against the plans of the mosque, as the planned mosque isn’t only a religious building. In fact, it welcomes everyone, and even a basketball court is planned to be built inside the mosque. A community center is a better word, and way less scary.