NBA Analytics Q&A With Seth Partnow of Nylon Calculus

facebooktwitterreddit

HoopsHabit is proud to partner with Nylon Calculus (and the HP Basketball Network) as part of the FanSided Network and recently, HoopsHabit’s Michael Dunlap had a chance to do an email Q&A with Nylon Calculus managing editor and analytics savant Seth Partnow.

We talked about what analytics mean to teams, players and fans, we looked at the future (and the past) of the industry and Seth peeled back the curtain to give us a glimpse at what’s going on at Nylon Calculus right now.

Michael Dunlap (HoopsHabit Editor-in-Chief): Thanks so much for doing this Q&A, Seth. Tell us a bit about yourself.

Seth Partnow (Managing Editor, Nylon Calculus): I’m a lifelong NBA fan and former small college player. I took over as managing editor of the basketball analytics website The Nylon Calculus earlier this summer. At Nylon Calculus, I have developed new metrics covering such topics as defensive rim protection, individual playmaking ability and have also written and reported extensively on trends in the practice of statistical and quantitative analysis in and around the NBA. In addition to Nylon Calculus, my writing can be found at such places as Washington Post’s #FancyStats blog.

MD: For the uninitiated, can you explain what analytics are and what role they play in basketball?

SP: That’s not a question with a simple answer, since the term itself – “analytics” – means many different things to many different people.

My personal preference is to use it as synonymous with the study of information, as I strongly dislike the artificial dichotomy of “the numbers” and “the eye test.” Focusing exclusively on one or the other leads to getting only half of the story. Still, analytics generally means the numbers and stats side of that ledger.

In terms of the role they play, teams, coaches and players have relied on statistics for as long as we’ve been keeping score, the “analytics movement” if you will is mostly about applying new techniques and information gathering systems to the exploration of who has performed best on the floor.

MD: Can analytics enhance the basketball watching experience for fans or is it meant more for the teams and media?

SP: It definitely can, otherwise why would fantasy sports he so popular? The presentation for fans needs to be different than it would be for a coach, scout or GM. However, good use of statistics, even more quote-unquote “advanced metrics” can increase fan engagement and enjoyment by better illuminating both what happened and why it happened.

Again, statistics have been used to detail and describe performances as long as there has been an NBA, and there’s no reason that shouldn’t continue. The challenge is to make newer and less familiar numbers meaningful to and digestible by the audience.

MD: You mention the presentation for fans. Is it that the casual fan is intimidated by new numbers, that they’re resistant to change or something else? What is Nylon Calculus doing to make analytics more attractive to those folks?

SP: That’s a big question. I think the “casual” fan is smart enough to meet you halfway. When I talked to Jim Petersen (color analyst for the Wolves and the best in the business for my money, as well as an assistant coach with the WNBA’s Minnesota Lynx) last season, he talked about how he tries to introduce these new concepts.

Effective field goal percentage – that’s just shooting percentage accounting for threes being worth more. “Oh, I get that!” Points per possession is a better way to look at stuff than just the scoreboard numbers. “Sure makes sense.”

Now, those are easy examples, but it’s like getting baseball fans to adapt to on-base percentage – if you’re a fan, you’re already familiar with the units of measurement so it’s easy to go from “traditional” to “advanced” – I hate the “advanced stats” nomenclature by the way, but doubt it’s going away any time soon.

Unfortunately, for the people that “do analytics” stuff like eFG% and PPP is so ten years ago, so a lot of the stuff that’s more cutting edge now is further removed from the experience and knowledge base of the curious but lay fan. Those of us who produce these measures need to do a better job putting these figures in terms people can understand.

At Nylon Calculus this can be pretty difficult, as we have several different readerships. A number of our authors have, are currently or would like to work for teams in some capacity, and the degree of precision and rigor you need to show that audience is very different from what a fan who wants to know just how good Rudy Gobert is at protecting the paint. Squaring that circle is really about tying an analytical finding back to basketball on the floor so the reader can picture a play in their mind’s eye which demonstrates or illustrates the point one is trying to make.

The other thing that needs to happen is “analytics” needs to stop being used as a cudgel. Every day on Twitter you can see people debating who is better between this guy and that guy. Almost invariably “the numbers say…” will be used as an argument. The numbers don’t say YOU say, by deciding which numbers you think are important. Too often the rationale behind the stats chosen is “because they happen to support my point this time” but that’s a separate thing.

Getting back to TNC, I hope we essentially say “look how much more we know than you” far less frequently than “look at this neat thing I found out.” It’s important to have that humility, because things everyone, even the smartest people, are certain about can become “the earth is flat” very quickly as new information appears.

MD: The direction of analytics — what’s the next big thing in the industry? Anything big going on at TNC that you can share with us?

SP: There are several things going on at once, as is often the case. From the outside, it’s hard to tell which will make the biggest difference first, and it’s probably different for each team, but here are the three main things:

1. Improved and deeper use of SportVU data. We’re really only scratching the surface of what this data can tell us. Part of it is the complexity of the data – it took the team of literal rocket scientists at Second Spectrum about 2 seasons to come up with a “vocabulary” allowing them to categorize and catalog all sorts of on and off-ball screens that occur in the NBA. And that’s just the rough equivalent of collecting box scores – how to apply that information is certainly proceeding with some speed, but almost all of it beyond closed doors. Also, the full potential of the SportVU data requires something of a change in thought process. We’re so used to looking at stats either from the individual player level or at the team level that really breaking down how interactions between players working in cooperation hasn’t been explored all that much. With spatial tracking data, it’s possible to start looking into those questions, but it gets really complex really quickly, so it might be slow going at first.

2. Improved “performance” analytics. Biometric tracking, injury prevention, fatigue modeling, training science. This stuff is all getting much better, and can present teams which master it more quickly with a huge competitive edge. In a salary cap-driven league if I can have 100% of my cap money on the floor, but 20% of yours is in suits on the sidelines, that’s advantage: me. Zach Lowe had a nice piece on the topic last week.

3. This isn’t analytics per se, but it is vital for the new information to become more widespread and actually used in the NBA is teams appear to be thinking much more about the process of how the data goes from quant to coach or quant to GM and ultimately down to the players. One of TNC’s writers, the eponymous Positive Residual, discussed how some new hires and organizational restructuring from the Kings and Lakers this summer reflect this change in progress. If you bring up the topic of “communication” to an analytics practitioner, you’ll probably get a massive eyeroll in return, because pick a year at Sloan and there have been multiple panel discussions on the topic of how to get the geeks to talk to the jocks and vice versa. But it’s an issue that’s neither going away, nor confined to the sporting realm at all.

You can have the best technical analysts in the world, but if they aren’t communicating their discoveries to the ultimate decision-makers (coaches and GMs) well enough or if those coaches and GMs aren’t really open to hearing about it, it won’t ever really matter on the court,

As far as what we have coming at TNC, we just have more. We’ve started doing some technical tutorials on some of the tools we and others use, as one of the best ways to demystify this stuff is to illustrate how much of it is driven by basketball logic rather than mathematical, statistical or computer programming technique. With the season coming up, the writers who have projection systems are getting geared up to make some predictions. We’ve (finally, I hope) come up with a better way to display some of the metrics we’ve come up with to allow the site to be used as a resource. To that end, we’re also continuing the Nylon Calculus 101 series as a sort of permanent glossary.We’re starting to learn how to best use the public SportVU movement data and hope to come up with some really interesting stuff from that. More college stuff, more WNBA stuff. Basically whatever flights of imaginations our writers are moved to go off on!

Ooh, I forgot one thing we’re doing which I’m actually kind of excited about. A lot of people do pretty interesting basketball analytics-related projects, papers and studies as senior theses or comprehensive exercises for their economics, mathematics or computer science degrees. We’ve actually had some people submit some ideas to us that way and we’re hoping to help people publicize those projects a little more so they can contribute to the conversation and not just be a necessity for graduation!

MD: Ultimately, what’s your goal for the site and is there anything else you’d like to add?

SP: As far as the overall goal for the site itself, I don’t think I’m being too ambitious by saying we want to become something like the Fangraphs of basketball. We have a long way to go to reach that level, but it’s doable, and there is a hunger for more and better statistical information about the NBA and basketball in general.

On a personal note, the response to the site has been pretty amazing. It’s equal parts exhilarating, intimidating and humbling to hear feedback from teams themselves about stuff we’ve published. That is definitely a good problem to have of course, and it very much inspires us to strive for great precision in our work. If people who do this for a living are going to be reading and considering what we publish, we owe it to both them and ourselves to make as sure as possible that we are correct or at least not making silly errors!

I also have to thank Ian Levy, the founding editor of TNC (and now EIC of Hardwood Paroxysm and the HP Basketball Network) for building such a strong platform for the rest of us to work from, and as the person who has helped me personally the most in terms of my own writing and analysis.

Next: Check Out Seth And All The Nylon Calculus Writers

More from Hoops Habit