Basketball Analytics: Reflections and Reservations

Okay, here’s the short version for those require maximum efficiency (n=”the point”/number of words) in their reading environment:

At the heart of my reservations is my sense that value can be defined in many ways: economically, aesthetically, morally, to name just a few.  And that economic definitions of value in terms of efficiency productivity may be beginning to eclipse and drive out of the public conversation considerations of aesthetic and moral and other values, particularly when these appear to be at odds with economic value.  In short: some very beautiful (aesthetically valuable) things are not economically efficient.  I am worried that once we have allowed certain kinds of things that we value to be eclipsed, we could find ourselves in the position of having lost those kinds of things forever.

The rest of you, who enjoy the backroads of an intellectual journey, please carry on with my sincere thanks.

Among the elements of my post on “the Warriors” that got some pushback were my comments about the rising use of analytics in hoops.  Some of that was silly, but some was thoughtful and came from individuals I respect a great deal, so I wanted to clarify my position on what I see as pretty much indisputably dominant trend in basketball culture over the past decade or so. It can be hard to have a reasonable conversation about complex issues on the internet, but I sincerely hope that these reflections can contribute to that.

What do I mean by “basketball analytics”? James Kerti has defined the phrase very broadly as the process of “using different types of data to solve a problem you have.”  Quantitative metrics, in his version of basketball analytics, is just one of many different types of data. Others include scouting reports, psychological profiles, film study, medical reports, and market evaluations.

That’s fair, and I am open to being persuaded that I should be using a more precise term for what I’m talking about because when I use the term I’m really only talking about the quantitative component of this. I understand that all the other forms of data he enumerates also form the basis for the analysis—in the broadest sense of the term—of the sport. However, I’m interested in the fact that within the culture of basketball over the past decade quantitative data has expanded its share of the overall “pie” used by parties interested in analyzing and discussing and making decisions about basketball.

In that sense, for my purposes, I think it’s reasonable for me to offer this working definition of “basketball analytics”: a set of methods for analyzing basketball play that combine 1) the observation of basketball play, either by human beings or by technological prostheses like those provided by SportVU; 2) the classification of various events into different qualitative categories (for example, the classification of shots according to who took them, where they took them from, under what conditions, and with what outcomes); 3) the conversion or translation of these qualitative observations into quantitative data (distance in feet from the basket, trajectory of the attempt, angle of the attempt as measured from the baseline, distance in feet of the defender from the shooter, ratio of makes to attempts, etc.); 4) the analysis of this data using statistical models; and 5) the rendering of this analysis in tables, charts, and, of course, prose.

I have no objections to any of these practices in and of themselves.  For one thing, as my friend Seth Partnow observed recently, as soon as you keep score, numbers are in the picture.  It makes no more sense to me for people to object that quantitative analysis of basketball is somehow a corruption of the purity of the sport than it does for people to object that dunks are that kind of violation. James Naismith kept score at the first game (1-0) and we have records ofScreen Shot 2016-02-01 at 11.38.23 AM box scores for the Springfield YMCA league, in which Naismith played, in 1894-1895.  Historically, counting, numbers, and the quantitive analysis of performance have always been a part of both formal and informal basketball play.

Nor do I think that what might be called philosophical objections make sense here.  I mean that I think it inevitable that we slice snapshots out of the flux of experience, whether in basketball or in life.  Photographers do this when they snap still photos, statisticians do this when they tally, writers do this when they compose recaps or reflections on the sport, even television crews do this by their choice of what to film and what not to film.  The game moves continuously, life moves continuously, and the best we can do is find different ways to capture and try to shape meanings for ourselves out of that flux of experience. I do have an issue with characterizing these things as “objective,” but that’s another story.

So if I don’t want to argue that analytics presents a historical betrayal of the sport, or a philosophically problematic form of representation.  Nor do I have a problem with those, perhaps more quantitatively minded than I, for whom analytics serves as a window through which to access and enjoy all the things in the sport that I access and enjoy through words or pictures. In fact, I sometime find that a quantitive analysis has helped me to understand, and so enhanced, my own enjoyment of some element of the game

So what is it, then?  James Kerti tells us that analytics is “about helping people make better decisions.”  Kerti is speaking now of the purpose of basketball analytics, and his capacious definition of that purpose makes good sense and seems acceptable to me.  Players surely want to learn to take the best decisions possible on the court. Coaches want to learn to make the best tactical and strategic decisions. Owners and general managers want to learn to make the best personnel decisions. I don’t begrudge any individual the right to use whatever data they can to analyze the life situations that are important to them so as to make better decisions.

But it also raises some questions that mark the areas of concern I do have about the rise of basketball analytics, all of which relate to purpose, use, and, even more so, to effects (intended or not).  First, by what criteria to we define “better decisions”? Better for what purpose? Better for whom? Who decides?  And how are disputes over these questions to be resolved, especially when the parties involved have unequal power?

These questions are neither unique nor new to basketball, and analytics is not the cause of them.  However, given the rising importance of quantitive analytics in basketball culture, I do think they are questions that it would be good for the analytics community, and all basketball lovers, to think about deeply and to talk about honestly and reasonably.

As it currently stands, it strikes me, to put it broad terms, that it is “the market” that is making these decisions, which is to say, that it is those primarily seeking to make a profit from basketball play making this decision (and to varying degrees their employees).  That is, basketball analytics has evolved in the direction of evolving fairly sophisticated measures of on-court value (in terms of individual efficiency, or contributions to team efficiency) that may then be correlated to economic value and used by owners and general managers in making personnel decisions. The idea, I am supposing, is that winning is fun and makes money and that analytics can help you to win and make more money, in part by learning how to spend less for the kinds of basketball plays that most efficiently contribute to winning, especially if you’ve hit on an algorithm your competitors have not.

I know that basketball is a business, a part of the entertainment industry in this country, and owners have a right to want to make a profit. I also know that many athletes embrace these trends since, after all, it can help them to make “better decisions,” to become more competitively effective, more efficient and more productive and therefore more economically valuable and indispensable to their franchises.

For some readers, I imagine, this is only natural and therefore either desirable or at least inevitable.  But it is not natural. It is a social arrangement resulting from countless decisions made by human beings and therefore alterable by human beings. I’m not saying I know that it should be altered, and certainly not how it should be altered, only that we can have and perhaps should keep that conversation alive, in part because while we may consider basketball to be a special kind of workplace “out there”, the very same trends I am describing in basketball are also “in here,” in our workplace.

Where I work, for example, in a major research university, economic efficiency is increasingly the standard to which all other values must be subordinated. In other words, if it is not efficient (say, in terms, of the number of student credit hours generated per faculty position: i..e maximizing enrollments regardless of subject matter or pedagogical considerations), then we simply have to sacrifice other values that we may profess to cherish but that have come to seem at cross purposes with economic efficiency.  Other values become luxuries that the supreme value of economic efficiency dictates we cannot afford.   Numbly and silently assenting to this logic and to its incremental creep has already facilitated the transformation of American higher education (and many other industries I know less about) in complicated ways.

I am not trying to hold the unreasonable position that efficiency or productivity are unimportant or irrelevant because I think they are vital.  Rather, I am trying to hold open a space in which we don’t forget to discuss what we think of the constantly shifting arrangement of different kinds of value in our lives, the practices to which those shifts lead, and the processes by which we those values shift including, not least, the social power differences that are at work shaping those shifts.

At the heart of my reservations is my sense that value can be defined in many ways: economically, aesthetically, morally, to name just a few.  And that economic definitions of value in terms of efficiency productivity may be beginning to eclipse and drive out of the public conversation considerations of aesthetic and moral and other values, particularly when these appear to be at odds with economic value.  In short: some very beautiful (aesthetically valuable) things are not economically efficient.  I am worried that once we have allowed certain kinds of things that we value to be eclipsed, we could find ourselves in the position of having lost those kinds of things forever.

 

 

6 comments

  • Pingback: Let's talk about NBA free agency - Newsity

  • Barnes is more aesthetically pleasing than Green but it doesn’t fucking matter
    Go watch the globetrotters

    • Green is consistently more fun to watch than Barnes but it may have little to do with the perceived aesthetics. Its not the statistics or numbers either. The vigor, effort, sweat, faces, bumps, and general physicality are a huge part of what makes basketball entertaining. So is Barnes fluid, leaping, smooth movement. And, they both have enough flaws (Green’s over exuberance and risky passing, Barnes’ questionable dribble ball handing) that adds to uncertainty and drama. Numbers can analyze and attempt to quantify their play as individuals and as a part of the team but so what? I really don’t think they can explain how they might play in a given circumstance. For me paying too much attention to the quantitative analysis takes away from enjoying the physical kinetic and emotional considerations that give all sports unpredictability.

  • Pingback: Basketball Analytics (Take 2): Winning | Between the Lines

  • Pingback: Towards a History of Hoops Analytics | Between the Lines

Leave a Reply

Your email address will not be published. Required fields are marked *