A Refreshing Q&A With Seth Partnow

shooting the summer breeze with the author of The Midrange Theory

Chances are if you’re subscribed to this newsletter then you’re already familiar with Seth Partnow and no introduction is needed.

But just in case you need a refresher, Partnow is the former Director of Basketball Research with the Milwaukee Bucks, author of The Midrange Theory, and now the Manager of NBA Data Science at Penn Interactive.

In many ways, this newsletter is just an imitation of what Partnow has been doing for years in Nylon Calculus, Vice, The Athletic, and elsewhere. Few people have shaped the way I think about the sport — and more specifically, writing about the sport — more than Partnow. His annual Player Tiers Project was appointment reading and I regularly revisit his analysis on subjects such as The Midrange, Shot Quality, and Three-Point Defense.

If Partnow is the Coca-Cola of analytics writing then The F5 is the Olipop — a watered down version with millennial branding.

These days, you can find Partnow on Bluesky sharing his thoughts about the NBA and airing his grievances about Manchester United. He’s also in the process of putting the finishing touches on a second book about basketball analytics for Triumph Books, the same publisher as his first book.

I reached out to Partnow to talk about the silliness of the NBA’s schedule release week, coining the term Heliocentric Basketball, and his expectations for Manchester United this season.

This Q&A has been lightly edited for length and clarity.

---

F5: I can generally count on you to be upset or annoyed about something. So I begin this Q&A by asking, what's something NBA-related that's been getting under your skin?

SP: I’ve got one new, small thing and one longstanding larger thing.

The small thing is the silliness of “schedule release week.”

First of all, aside from a few key dates (Christmas, MLK) the schedule isn’t really as much of A Thing in the NBA as it is in the NFL. 82 vs 17 will do that, especially if the NBA Cup groups have already been released.

Second, a week? Really? I understand and fully support the fact that we’re making a TV show here, but what’s the plausible theory that dragging it out this way in early August makes a lick of difference in terms of people actually watching the games in 10 weeks?

Second, and larger, can people stop trying to trade Giannis? If there is actual news, a round of reporting and discussion is fine. Transactional stuff makes the world go around, but since at least Christmas 2019, the level of thirst for Giannis slop has been silly.

Maybe I’m making up an enemy in my mind, but I could have sworn there was “what does this mean for Giannis’ future in Milwaukee?” talk during the 2021 Conference and NBA Finals. Maybe focus on the actual games that are happening instead of playing solitaire or whatever

You recently wrote something on Bluesky about how the phrase "impact on winning" has become as vague and undefinable as terms like "culture" or "alignment" — what does impact on winning mean to you?

Something that has become increasingly clear to me as I spend more time working with deeper and more granular NBA data is the importance of keeping the main thing the main thing. There are so many side quests you could go on, but it’s really necessary to keep as narrow a focus as possible for a given line of inquiry.

If we’re talking about winning basketball games, how do you do that? End the game with more points than the opponent.

In that light, almost by definition, impacting a team’s ability to win games means causing the scoreboard to move in the right direction. That could mean helping your team score more efficiently or inducing the opponent to struggle in their attempt to do so or a combination of both. Raw plus/minus isn’t a great measure of impact because of all the associated causal and variance issues, but it’s the right frame. The main thing is really "how much does this player help us outscore the opponent?”

With that in mind, if what a player does on the floor doesn’t show up the scoreboard, is it really impacting things? This isn’t to say we have a perfect grasp of scoreboard impact or that there are some less tangible things that a player might do that aren’t necessarily associated with them being on the floor or not that can enhance or detract from team performance on the whole. For example, being a locker room energy vampire might not show up in a player’s on/off splits or RAPM scores, but it can’t possibly help the overall squad. But I think players who materially impact results via vibes either positively or negatively are pretty rare, so for the most part, moving the scoreboard is the thing.

This recently came up when Kevin Durant was arguing on socials with someone - as he does - and the other guy questioned KD's ability to "impact winning at the highest levels" and what does that even mean? Since his first All-Star season in 2009-10 his teams have been 8.0/100 better with him on the floor than off. His 29-year xRAPM is +6.1/100, in the 97th percentile of all players over the PBP era.

You’re going to need to show me some pretty extraordinary evidence that someone with that profile isn’t an extremely winning player no matter how big a pain in the ass he might be. The argument certainly seems to be more about disapproval of KD’s career choices than about his on-floor impact. Which is a perfectly legitimate opinion if we’re talking about overall legacy or whatever, but that’s a vibes-based discussion, whereas “impact” as measured via the scoreboard is much more amenable to statistical analysis. To put it another way, defining “impact on winning at the highest levels” in a way that doesn’t explicitly tie back to the scoreboard is the sort of non-falsifiable hand-waving that has already consumed terms like “culture” and “alignment” which in a dictionary sense describe real things that we know are important but in practice are now just codewords used by executives or commentators to give the appearance of Doing Something. They’re the NBA news breaker equivalent of verbiage in a corporate press release announcing the great success of something or other.

You've long been beating the drum for teams to play hardball with their own free agents instead of just giving them whatever they want as soon as they want in contract negotiations. I think we saw a little more of that this offseason than in previous years. What do you attribute that to? Is it just team's trying to dance around the 2nd apron or is there something else happening?

I think I would restate your interpretation of my position slightly. I don’t think of it in terms of playing hardball against players so much as trying to maximize the value/ability/versatility of the whole roster.

Let me back up for a second and ask a rhetorical question. Why is the current CBA and especially the Second Apron described as being anti-player? The players as a whole are getting the same money no matter what. The systematic changes probably affect the distribution of salaries across different levels and types of players in ways we will continue to see shake out, but it hasn’t changed the size of the slice of overall pie given to players in any way.

That clarification is important because I’ve never advocated for teams to be cheap in negotiations. Rather I want them to better recognize that every dollar they give player X reduces the amount they can give to players A-N on their roster. The CBA provides tools through which teams sometimes have leverage, especially in Restricted Free Agency.

Obviously, you have to consider the human element and not unduly piss guys off, but how much should that be worth? Give a guy an extra 500k Average Annual Value (AAV) to get the deal done and preserve team harmony or whatever? Sure, why not? However, the general pattern has been for teams to take on way more risk than they should especially with respect to rookie extensions signed in players’ 4th years. Many of the deals I end up criticizing heavily are those that would be more or less fine if the player gets through their 4th year well without injury and on the expected developmental path.

The question I always come back to is why give the player that deal a year early and take on all the risk on the team?

A hill I’ll die on is the Timberwolves extending Jaden McDaniels a year early to high 20s AAV was a bad gamble.

Let me preface this by saying I REALLY like McDaniels and think he’s an excellent complimentary player for just about any construction of contending team. But what kind of season would he have had to have in his 4th year to get an offer sheet well north of 30 AAV? How likely would he be to have that kind of season on a team with Anthony Edwards and (at the time) Karl-Anthony Towns as huge usage sinks? Especially given that the RFA offer sheet has effectively vanished from the NBA landscape? How much of the justification for the deal would be one season of near 40% three point shooting for a guy who profiled much closer to 35% in terms of prior seasons and free throw percentage? While the McDaniels deal has largely worked out, that doesn’t mean it was a good risk to take on. And to do this well, you have to think of the transactions as bets. Just as terrible shots sometimes bank in or a player might blow a tire and miss a wide open layup for time to time, there are no sure things, and sometimes shit happens to derail the best-laid plans.

So back to the McDaniels example, from the viewpoint of heading into the 2023-24 season, would you have looked at the season ahead and thought it would be more likely that at the end of it, someone would be looking to give him $35m a year or $20m a year? Given the hammer of RFA and the fact that teams don’t usually line up to give huge deals to average efficiency scorers on 16 percent Usage, there sure seemed like a whole lot more room below than above in terms of the contract actually signed. It’s a great deal for McDaniels, locking in perhaps 85% of his likely best case scenario a year early, but what did the Wolves get for taking on that risk? By the way, is this a good spot to note that McDaniels has hit exactly one third of his 3s over the last 2 years, making that one season of 39.8% look distinctly outliers?

Now to the second part of your question, I think the Second Apron has had an effect insofar as the kinds of deals which were already bad, subtly but significantly harming team-building for many franchises are now so obviously bad through the restrictions created by the second apron that teams have come around closer to my point of view. Many years ago, a friend who works in hockey analysis told me that NHL teams shouldn’t give real money on long term deals to 3rd or 4th line players because if a first or second line declined a little or you simply misevaluated on the high side, they could still be serviceable and useful 3rd or 4th liners, but if a deeper rotation player declines or was over evaluated, they are no longer NHL-caliber players. It’s the difference between “bad” money and “dead” money, and if you look at the history of teams that have risen up before ultimately falling just short of inner-circle contention status, you’re going to find a lot of instances of 8th-10th men making 10+ million for 4+ years while wondering why they always seem to run out of dudes in the post season. An extra 3 million here an extra 1.5 million there doesn’t take long to add up to a rotation player or two your aren’t able to acquire.

Whereas before the inability to add those extra couple of useful players was more practical - very few franchises have operated at what are now second apron levels under previous CBAs - the second apron makes it mechanically difficult if not impossible, which gives a bright flashing “do not enter” sign that teams may have ignored before when the sign was just printed but not actually lit.

Has anything happened in the years since writing The Midrange Theory that has made you change your mind about something you wrote at the time?

I wrote the bulk of the book during the 2021 season and can barely remember most of what I said. The biggest changes in my thinking over that time frame have probably been becoming even more of a fundamentalist about the importance of teams with contention aspirations thinking about their roster and rotation explicitly in terms of players who can survive or thrive in the second round of the playoffs or later. Which is as much about having the ability to address different challenges presented by different opponents as it is about overall ability.

I fell like the league has too many players that have gotten too good at shooting unassisted, off the dribble 3s. Those are shots that players can get pretty much whenever they want. I'm not sure we're there yet, but at what point would you start to consider rule changes that either help the defense or do something to make those shots less valuable relative to alternatives.

I think we’re well past the point of needing some rule, or at least interpretation changes. My big hobbyhorse has and continues to be illegal screens. Players are able to get to those shots in large part because it’s hard to stay in front of a ballhandler if the screener can slide or spread out in such a way that the defender needs to take a path two feet further from the basket than the ballhandler. If those pull-up threes require the ability to shake a defender and then rise up on balance, suddenly the number of players who can reliably do so with any sort of efficiency would drop substantially.

Do you have any regrets about calling it "Heliocentric basketball?" It's become part of a group of NBA buzzwords that signal, "I take basketball very seriously." I like it, but I do feel like a dweeb whenever I write it or say it out loud.

It’s always amusing to me that’s my biggest contribution to the discourse because it was a word that was mostly a throwaway trying to come up with a pithy title for an article on the phenomenon that has become known as heliocentrism. I haven’t used the term in earnest in years.

If you were trying to convince someone that defenses have limited control over opponent three-point percentage (a.k.a. Jedi Defense), what's the one piece of evidence you would point to?

The fact that there is virtually zero year-to-year correlation or even 1st half to 2nd half of the season in opponent 3FG%. Aside from the Celtics, which is one of those long standing head scratchers that nobody has come up with a good explanation for. If Boston is doing something special, how come it has persisted across four coaches, two GMs probably five different roster cores yet none of the coaches or players who have left have brought it with them to new spots?

I’ll play devils advocate and point out the following teams and where they ranked in opponent 3P% last season:

• BOS - 3rd

• LAC - 5th: Jeff Van Gundy is their assistant coach, but was in Boston before that

• CHA - 9th: Charles Lee is their head coach, but was in Boston before that

• HOU - 11th: Ime Udoka is their head coach, but was in Boston before that

• UTA - 12th: Will Hardy is their head coach, but was in Boston before that

Five teams with five strong (and recent) ties to Boston all finished in the top half of opponent three-point shooting. I know it's only one year. But that could be something, right?

Four guys with a combined four seasons in Boston (only three on the bench)? Given the amount of coaching movement in the NBA, I think we could probably find someone of the staff of just about every team in the league that has spent time in Boston, so of course some of them will be associated with teams that saw opponents shoot worse than average on 3s.

Let’s look at Hardy’s 3 seasons in Utah. They finished 16th, 30th and then 12th. Udoka’s Rockets finished 1st in his first year before slipping to basically average in 11th last year. Five data points does not a trend make, but the average finish is 14th.

Billy Beane & Luke Bornn recently gave an interview in The Athletic. There's one section where Bornn expressed some skepticism about how much football (soccer) teams are really using the data they have in front of them. Do you find Bornn's comments persuasive and do you think they're applicable to basketball?

Yes and yes, next question.

Are teams too secretive with what they're working on internally? I was reading some of Devin Pleuler's writing and he made the case for why the research arms of teams should consider open sourcing some of what they're working on. Pleuler argues that the best organizations will be in position to take advantage of any innovation that comes out of the public realm so maybe it's in their best interest to pull back the curtain a little. Pleuler was writing about football (soccer), but there seems to be an analogous case in basketball.

Devin is a gentleman and a scholar and I largely agree with him.

This actually relates to the previous question to a large degree insofar as the “state secrets” approach presumes that the *answers* derived via statistical analysis are the value add rather than proper inclusion of data in the overall decision making *process*. In that regard, properly used, analytics is much more about asking the right questions than it is having a statistical almanac on hand.

Unexpected Origins and the Fermi Paradox

Very few of the metrics that are useful in decision making are so inherently complex that teams gain significant advantage from having a better player value metric. You get to a certain point of accuracy and/or predictiveness and these models will generally produce similar results though each particular flavor will value certain skills more or less and have slightly different holes and blindspots.

Rather than the value coming from the brute force of having more/better/proprietary data, it’s the application to and integration with the decision process that gives those metrics value. This is the part where even if the “smarter” teams were completely transparent in how they were doing so, the organizational will to operate in a process-based manner isn’t a matter of knowing what the best practices are, it’s implementing those practices.

The following famous Micah McCurdy tweet isn’t much less true, across most non-baseball sports, now that it was when he originally posted it.

You've worked in both American football and European football analytics since leaving your job in the NBA. What's one thing you picked up during your time with Statsbomb that made you re-think something about basketball?

A few things. One is that American football made me very thankful for the event sample sizes we get in basketball. Fewer games, less plays, more specificity in situation (down and distance. Personnel packages. Formations) mean that the level of precision needed to properly capture events is higher which in turn makes the analysis much more difficult because it can be very easy to find yourself with conclusions like “in the 5 total times this situation occurred, they have passed 4 times and therefore…”

The second and broader thing was it was extremely useful to get some real firsthand experience with how the data sausage gets made because it reinforced the need to understand with a decent amount of detail the definitions of how various stats, events and data points are tracked. Which in turn is a useful reminder that statistical analysis is only “objective” insofar as the evaluation is consistent rather than being objective in terms of some sort of ground truth. The data itself is the result of many decisions that have been made in in terms of what is being captured and how those events are being described. In most cases, the choices that have been made have been decent ones, but very rarely are they the only possible ways which a certain event or occurrence could be captured, and those decisions can fundamentally alter the results of underlying analysis.

Essentially, knowing how imperfectly the data we collect represents the game itself is a good reminder of the need for some humility in how certain one should be about the correctness of any analytically-based conclusion over and above the normal worries about statistical variance or properly processing data.

What do you think is the biggest misconception about working in sports?

That the jobs are in any way glamorous. That includes players by the way. These guys GRIND, not just to improve their skills but especially as they age to get and stay on the court to the maximum extent possible.

I have heard you say your NBA comp in pickup is Seth Curry. Explain.

First of all, I last played in February 2020. I was already in my 40s and had decided I was going to keep playing regularly until I hurt myself and then be done, but then the pandemic happened and I decided to skip the Achilles rehab portion of the plan and just be done.

So with the caveat this is “was” not “is,” I can/could shoot, with range and off movement. I had a little more off the bounce than you’d expect especially against a closeout, and I had a little more passing and overall playmaking skill than would “just a shooter” I was able to be effective despite being a little small, a little unathletic relative to the level of play. (I mean give me a break, in staff pickup games, I was a 42-year old former D3 benchwarmer sometimes matched up against an ex-D1 film room guy or scout in their mid-20s, but I think I held my own, even if some of those guys would still accuse me of fouling too much. Looking at you Ryan Frazier...)

You're a Manchester United fan. Are you optimistic about this season? I'm a Brighton fan and very happy that United did not back up the Brinks truck for Carlos Baleba.

They played better than I expected in the season opener against Arsenal, but they still lost and somehow managed to both be completely outnumbered in midfield AND never have a target in the Arsenal box when they did get forward.

That said, I am completely lacking in confidence in United doing anything until they start operating as if they understand roster construction and/or team building.

As of now they have had the highest net transfer spend of any team in the Premier League despite adding zero Champions League starter-level players. What are we doing here?

I also know enough second and third-hand about how things work behind the scenes there that I know any good decisions they reach are going to be more about fortune than based in planning or competent execution, which doesn’t really bode well for growth.

What's one thing you can't live without during the NBA season?

League pass!

I’m not sure people realize how lucky we’ve been with being able to watch the current generation of superstars pretty much on demand, regardless of the markets they play in.

I think a perhaps under-estimated portion of Jordan’s legacy is the fact that because of WGN being a “superstation” Bulls games were among the most accessible in the league for the bulk of his career. So being able to watch teams like the Thunder over the past couple years or the Rockets and Pistons that are on the come up isn’t something that should be taken for granted.

The national TV schedule is always going to be tilted towards certain markets for business reasons, but not being completely beholden to that makes being a fan of the game and the league much much MUCH easier than it once was.

---

For more interviews, visit The F5’s Q&A archive. I’ve done interviews with the creators of EPM, DARKO, LEBRON, xRAPM as well as ESPN’s Dean Oliver, Kevin Pelton, and more.