Tuesday, January 22, 2019

The Best Defense is . . .

The game of football has changed in the couple decades in a fundamental way. Nowhere is this more acutely witnessed than in the college game. And perhaps no team more demonstrates the odd juxtaposition than the Oklahoma Sooners.

Last year Oklahoma had a very poor defense looking at the traditional statistics. While these do not outright lie, they also do not tell the full picture. Oklahoma played in a league with very good and innovative offenses. And Oklahoma itself had one of the all-time great offenses last year. Their yards per play and points per play, to name just two meaningful statistics, were at near record levels. Keep in mind that historically most teams who have touched the type of numbers Oklahoma put up last year did so playing lower-tier opponents. In contrast Oklahoma played in one of the tougher conferences (those who find this to be a controversial statement ignore results and are caught up in old-world expectations of what football "should" look like), and they competed in the college football playoff against an exceptionally good Alabama team. The Oklahoma offense was Platonic while the Oklahoma defense was non-Platonic beyond all recognition.

While the Sooners are the extreme case, this combination of great offense and bad defense (albeit not necessarily on the same team) is a growing phenomenon throughout the sport. Teams with "old-world" offenses fail to reach the pinnacle within their leagues cough...Michigan... and teams with "old-world" defenses such as, sadly for me, Oklahoma look like two entirely different teams between when they possess the ball and when they do not. This resulted in Oklahoma parting ways with its defensive coordinator, Mike Stoops, after a disappointing loss to Texas mid season (in case you don't know, firing a coordinator mid season in what is otherwise a very promising season is a very rare event). After scrambling to fill the gaps to finish the year, Oklahoma has cleaned house bringing in a promising and perhaps innovative group to run the defense.

So far this is just me discussing my rather unremarkable and unoriginal observations. While I absolutely do not claim to have the answers, I think I do have an advanced framework for discovering the solution. Here we go...

To restate the premise: Defensive strategies and tactics desperately need to evolve since offenses have evolved at a much more rapid pace and in a more dynamic way for some time now.

Defense has a chance to catch up before offenses figure out that they don’t have to punt every time on 4th down. If offenses figure that out and continue to innovate in other ways, defenses are screwed. Case in point, offensive production is up remarkably over the past 19 seasons (2000 vs 2018). Across all FBS-division college football teams the average production is up meaningfully in the categories of yards per play (11%), points per play (14%), and points per game (12%). Maybe it is too late for defense?

Assuming this is not a lost cause, let's focus on a solution. The key is to be disruptive. Yes, that has always been the key. It’s just that the strategy of how to achieve that has changed over time. Back when running the football was the dominant offensive strategy, defenses needed to try to limit offenses to about 3 yards per play. Achieving that play in and play out is much more difficult today. Limit an opponent to a gain of 2 yards twice in a row (creating 3rd and 6), and you just don't feel like down and distance is on your side. A crude but probably telling stat: In the 2000 season just 10 teams went the season averaging gaining over 6 yards per play. In 2018 46 teams did.

Defensive strategy seems to have fallen into a losers games where defenses are trying to be very safe and limiting every play relying on the offense eventually making a mistake. "Just don't get beat deep..." goes the saying. The problem is offenses don’t take as many iterations (plays) to score as they used to, and they are more explosive. Hence, they don’t need to be as steadily productive as they did in the past.

Defensive strategy has to shift the probability back in its favor. The key is to increase volatility which increases disruption. If you will more likely eventually give up a score (and this is increasingly the case for all teams in nearly all situations), you might as well give it up right away having taken a big chance to stop it. Simply put: Just get the ball back--create failure or fail fast. Second and 11 is not as difficult for an offense to overcome as it was in the past. Likewise, defenses need to realize that second and 2 is not as bad a position to be in as it once was. Defenses need strategies and tactics that increase the probability of turnovers and tackles for large losses at the cost of increasing the probability of giving up meaningful offensive yardage gains and scores. The defensive mindset needs to begin with the assumption that everything is four-down territory--don't assume an offense will punt except in the most dire of positions.

A critical next step is to think terms of making an offense have as difficult a time scoring as possible regardless of what they do in trying to score. Defenses need to push offenses into making high-risk, low-reward decisions. Lowering an offenses chances of success is not entirely enough. Defenses must make offenses choose lower payout expected values. Think about an analogy with basketball. If defenses had their way, they would make offenses shoot every shot within a zone that is within about 0-5 feet inside the three point line. What is the football equivalent?

The most disruptive defense one could imagine from an NFL fantasy lineup would perhaps be the five best pass-rushing defensive linemen and six players among the strongest defensive backs or fastest linebackers. Notice that I am not saying the best run-stopping defensive linemen or the fastest defensive backs (cornerbacks or safeties) or the biggest linebackers. This group is going to blitz (send a disproportionate number of players rushing toward the quarterback) nearly every play. The remaining players are going to attempt to shut down the quarterback's options. What would it look like in result? Lots of small yardage given up, painful amounts of big plays given up, but all at the benefit of relatively many great defensive outcomes (turnovers, big yardage losses, etc.). Of course, that is a fantasy team. But perhaps we can get close to that by selecting for players in those molds and for plays in that fashion. Consider an analogy from baseball. There is a reason The Shift has become very popular--it shifts leverage in favor of the defense in a number of ways.

Why so aggressive--blitzing and otherwise? Start by thinking about how a defense is going to stop various types of plays used against it. Consider the most impactful offensive play, a deep pass down the field for either a touchdown or a very large gain. The first, best way to stop the play is to prevent the quarterback from throwing it in the first place. The second, best way to do it is to stop him from throwing it well while at the same time disrupting the receiver running the route. The third, best (worst) way is to have a defensive back covering the receiver being thrown to attempt to break up the play at the point of the catch. This holds not just for the deep pass but for all passes down the field. And by extension of the logic it holds for all plays.

Think about a running play. Where does the defense most want to attempt to stop a run? At or before the runner even begins, which would be in the offensive backfield. If the runner gets free for 9 yards as compared to 8 yards, the difference is very minimal. If the runner goes for 20 yards or 15 yards, again the difference is minimal. I am not saying it is not a big deal to give up large rushing plays. I am saying a combination of negative yardage running plays (-1 to -4, say) and positive, larger yardage running plays (9 to 20, say) is preferable to consistently yielding 3 to 8 yards each run. Don't try to make the offense fail small each down. Just one large failure (a turnover or a big loss on 3rd down or a big loss on 1st or 2nd down greatly reducing the offense's play options ) would be enough to allow the defense to win the drive.

The essence of my philosophy is a shift in the risk/reward and, hence, outcome dispersion for each defensive play and set of plays in a drive. Great defense should be torturous to watch for both teams' fans. For the offense and its fans it will look unpredictable and threatening--at any moment it might end the drive. For the defense's fans it will look unpredictable and haphazard--at any moment it might give up a score. Note the classic fan pessimism at play in both cases.

I admittedly do not have all the answers for how difficult it is to coach defense in football today. I just am on the lookout for the next innovators who I think will employ strategies and tactics very different from what has worked before. 

Saturday, January 19, 2019

'Oh, you left out a bunch of stuff' - 2018 New Year's Resolution fulfillment post

What better time to wake up from my no-blogging slumber than with the annual fulfillment of my perpetual New Year's resolution?

I used to strongly believe that “real-world” experience as a substitute for learning through formal study was over rated. There are two significant ways I have changed my mind. I now believe:

  1. Most learning done in school is learning in name alone. For the vast majority of people very little is truly understood and retained much less applied in life.
  2. Because of biases, failure to update/challenge conventional wisdom, poor feedback loops, and long cycles for knowledge updating, there is a chasm between the received wisdom and truth--what we could/should know but basically do not.

Bryan Caplan’s work as summarized in his book brought me around. This one has some irony. I probably shouldn’t be surprised that an esoteric, theoretical academic would be the one to set me free since my bias was built upon a disdain and rejection of those who (I still believe) unduly criticize and dismiss book/school learning and “theory”. I still highly value idealistic university education (at least in theory). I just now understand that experience in the world has much, much more value and applicability than I used to give it credit.

And it is not just that getting one's hands dirty learning by doing should be on equal footing. For most (see point #1 above) it is by far the primary way one should gain knowledge and wisdom and skills.

This change in view was for me a long time developing. As I remember it, the first major salvo came from Charles Murray when I read this piece. Caplan just pushed me from agnostic to full-blown evangelist.

Wednesday, May 23, 2018

Highly Linkable: Pay College Athletes Edition

With two important cases working their way through the courts (Jenkins v. NCAA and Alston v. NCAA), I continue to be optimistic that we are witnessing the beginning of the end for the government-protected, exploitative monopoly.

Just as Patrick Hruby explains in this Deadspin article, I have always found the argumentation along the lines "define specifically and prove explicitly how this change will work" to be shallow and weak. To argue that you lack the imagination to assume the market can devise a way to pay athletes, is no argument at all. As he concludes,
College athletes don’t need a pay-for-play plan, because pay-for-work isn’t a quantum leap. It’s just a small step in the direction of the world the rest of us already inhabit. The NCAA loves to talk about how college sports prepare players for The Game Of Life. There’s an easier and much more just way to do that.
As Ziggy might say, you'd have to have a Swiss cheese mind to not believe solutions will be discovered.

While we are on the topic of the ridiculous, Andy Schwarz takes apart the contention that most colleges couldn't afford to pay a market price for athletes.

But rest assured, Condoleezza Rice's commission fixed it all.

Sunday, May 20, 2018

Highly Linkable: Counter-Conventional Wisdom Edition

Trying to get back into the swing of blogging and just beating the 90-day hiatus limit . . .

Looking through my saved articles for future linking, I notice that just about all of them can be labeled "counter-conventional wisdom". Here are a few that, yes, have some age on them in the world of "that's so yesterday's Twitter", but I think they have value enough to be shared.

David Friedman, whose latest book is on my to read list, wrote about attending a Jewish wedding which got him thinking about what I would call modernity-biased myths about the past. I particularly like the Columbus myth.

You can't go very long discussing cryptocurrencies with a skeptic before they bring up the supposed Tulip Mania of 1600's Netherlands. But as I believe I've posted (or intended to) before, this is a myth. Hat tip to Tyler (of course).

I know I've posted before regarding our new puritanical age. I'm in good company with Matt Ridley who makes the case for today's "Millennials" being new Victorians. Yet again the young kids these days are not fitting their own (or the perennial young kids these days) stereotype.

This example of all common sources being wrong by Scott Sumner is a great example of a common view that unfortunately does not get scrutiny or challenge by the watchdogs or fact checkers.

Economics as a discipline itself needs more heresy (and reversals such that the heterodox becomes the orthodox). Arnold Kling, never shy to challenge along these lines, offers four contentions.

Sticking with Kling, he outlines five myths clouding health-care policy in the U.S.

Tuesday, February 20, 2018

Partial List: Twin Peaks - Wax & Wane

Partial list of peaks...

Some I predict we are in (or recently passed):
  • Garage, driver, and long-haul truck driver
  • Oil, et al. price
  • Professional stock picker
  • Bank (traditional) - regulation and innovation are to "blame"
  • Farm (agricultural land use) - see here & here
  • Storage unit

And some I predict we are not:
  • Local truck driver and this
  • Oil, et al. quantity
  • Index investing (true, pure passive even without the growth of factor-based, which is active)
  • Bond price - relatively low rates as far as the eye can see
  • Reality TV
  • Zoning - the Complacent Class isn't done yet "protecting" us from new ideas and FOOL is all about
  • Authenticity - the desire for this is just building and its continued strength is evidenced by the concern so many have that it is going away.

Sunday, February 18, 2018

What's Ahead for Stocks - precise predictions

Seriously?!? You clicked thinking you'd find some nonsense about, say, money about to [do something in regards to] "the sidelines", or perhaps you wanted to know how many technical indicators were crossing arbitrary thresholds. Oh, maybe it was an insider's take on smart money that you sought. But what would make the traders behind it "smart", how would I know what "they" (in unison? all on the same side of each trade?) were doing, and if I did, why would I share it?

Markets recalibrate constantly to new information. They also recalibrate constantly to changes in the weighted-average risk appetite of market participants. Did something change over the past couple of weeks? Of course, there is always something changing. But what?...

John Cochrane offers a great post for that question. Short answer: nobody knows. It cannot be known.

But what if we're in a bubble? Yeah, about that... Scott Sumner has two recent posts on that topic and more. He suggests we not be so sure about labeling past prices bubbles and lower the status of pessimists (I agree). He also suggests we should not offer explanations for events for which we are ignorant (I agree).

The standard advice is still the best advice:

  • Set your asset allocation as appropriate for best achieving your goals and personal constraints.
  • Get broad (very broad) diversification . . . cheaply.*
  • Go for lunch.
  • Check from time to time (not minute to minute) readjusting if needed to more appropriately fit your current goals.


*There are LOTS of investment options out there. The links show just two--albeit, two very good ones for achieving broad, cheap diversification. Also, maybe this.

Highly Linkable - How Are *We* Doing?

This links post is comprised of several items I believe are linked together in theme or subject matter. See what you think...

First Don Boudreaux points to a great website and corresponding TED talk by Anna Rosling Rönnlund. The project is a photographic-based exploration of how people compare. The within-country and among-country comparisons highlight what wealth and poverty look like. Notice the similarities, notice the differences, and notice on what factors these things do and do not seem to correlate.

Steven Pinker makes a strong case that The Enlightenment Is Working--"Don’t listen to the gloom-sayers. The world has improved by every measure of human flourishing over the past two centuries, and the progress continues." Let's suppose you conducted a survey every year for the past two centuries asking people simply, "Are you better off today than last year?" My guess would be the average and very typical response would be hard to distinguish from 'basically no improvement'. YET, the improvement over that time span for all of humanity (not just the average but for EVERY cohort) is dramatic and undeniable (once you look at the evidence). Why might this paradoxical result occur?...

Part II of Russ Robert's The Numbers Game is an examination of economic progress which suggests answers to the prior question above. The subtle yet very dramatic, counter-intuitive lesson, Simpson's Paradox, is awesome. To be sure, Simpson's Paradox would not answer my hypothetical, but it relates to how we misperceive small but compounding change and growth. Also, don't miss the first installment of Russ's video series.

But wait, aren't there too many people (or soon will be) for all this good news to continue? Steven Landsburg explores this issue in this video. He starts where everyone should start but often does not by asking "How would we know?" I believe he makes a very strong case that the answer is 'NO' we don't have and will not have "too many" people.

Tyler Cowen pointed to a couple of posts by Katja Grace who ponders 'Why did everything take so long?' The first and second both cover how and why progress is so difficult.