With a few minor tweaks, the annual NBA GM poll could be so much more useful

Photo by Matteo Marchi/Getty Images

The annual NBA GM poll is great at generating debate, but less useful at providing actual information.

The annual GM poll on NBA.com should be a window into how the league’s top executives view its top players and coaches. Except it’s not. It’s actually an episode of Glee from 2010 called “Grilled Cheesus.”

This episode deals with “motivated perception,” a fancy psychological term that just means if you stare at something for long enough, you’ll eventually see what you want to see. Finn, an emotionally confused high school football player and Glee Club co-captain, makes a grilled cheese sandwich and is convinced he sees the face of Jesus emblazoned on the bread.

The NBA.com GM poll is slightly more useful than a grilled cheese sandwich, but it still fails to offer any actual conclusions or even point to useful trends that I didn’t already know. It just dumps as much data as possible on the floor like a babysitter who isn’t exactly sure which toys a child wants to play with, opting instead to pour out the whole bin so they can say they found the Hot Wheels Corvette they wanted.

It’s just giving me everything, editing little and explaining nothing, searing the grilled cheese with so many different pans that I can see what I want: my team’s savior, evidence that everyone is out to get us, reasons for optimism or justification for panic. Today, the grilled cheese showed me that my Boston Celtics will three-peat; perhaps in February, if the Celtics are 3 games below .500, it will show me why we were doomed.

So to fix that, we’re going to have to cook something else. The NBA fan base is not a toddler looking for Hot Wheels, and thus does not need 18 different responses for “Which player is most likely to have a breakout season in 2024-25?” complete with a three-way tie for 5th and an 11-person honorable mention section. Within that sit both Anthony Edwards (one of the best players in the league) and Andrew Nembhard, whose “breakout season” might be getting his points-per-game above 10 for the first time in his career.

There are plenty of useful categories; seeing that 40% of GMs believe Victor Wembanyama is already the best defensive player in the league is cool, given his age and team’s overall poor defense last year. His individual effect was just that noticeable. The fact that 97% of GMs picked the Boston Celtics to win the East says something about their perceived dominance.

But the data isn’t the issue. Rather, the exercise lacks the features and formatting to provide readers with anything beyond confirmation-biased gut checks. The entire inclusion of the “also receiving votes” category is unnecessary, merely letting fans of that team or player latch onto the fact that one GM decided they deserved a nod in a given category. If you look at this sandwich a certain way, Jaylen Brown could win MVP this year, for example.

These kinds of honorable mention categories need to be explained or elaborated on. In the actually-good version of this exercise done on NFL QBs every year by The Athletic’s Mike Sando, a mountain of anonymous quotes are included for voters to explain why they made their decisions. If we’re going to throw DeMar DeRozan in the “Which player would you want taking a shot with the game on the line?” category, I want to hear from whoever put him over the 40 or so players in the league better than him, even if they have to speak anonymously.

I, a Celtics fan, could go a million directions from this article since it refuses to make any statements of its own. Jayson Tatum is the best small forward? Yay! But why is Luka Doncic second when I just saw that he was the best point guard? And didn’t Tatum technically play power forward all of last season? Wait, Joe Mazzulla is nowhere to be found in the “best coach” part? But is all over the categories about what coaches actually do like offensive and schemes? Is there a war on Joe Mazz?!

Probably none of my inner Celtics fan freakout would actually be valid if the article bothered to explain what the GMs were thinking or elaborated on the data a bit more, but it refuses to. I am left to my own devices, relegated to a toddler amongst an overwhelming sea of Tonka Trucks and broken Legos. Eventually, I’ll start crying and need to be paid off with a popsicle, or at least some Payton Pritchard preseason minutes

In any interview-based study, it’s important to give your respondents room to interpret the question… but not this much room. I am not familiar with the exact methodology at work here, but Andrew Nembhard, Anthony Edwards and Ja Morant are three players at completely different phases of their careers and could only be sharing 5/18ths of the “Which player is most likely to have a breakout season in 2024-25?” votes if the question was given to GMs without even one more word of context, like a fill in the blank question on my sixth-grade geography test. “What is the capital of Bolivia? ________” has that same energy.

I would certainly love to critique the methodology further, but the article gives us almost no information about it! The only reference to how the study was conducted is this maddening line “Percentages are based on the pool of respondents to that particular question, rather than all 30 GMs.” What on earth does that mean? Did GMs really just get to pick which questions they wanted to answer? Did some not answer at all? Did each GM get offered a random selection of questions? If so, why? This isn’t a clinical trial with 35,000 subjects; it’s 30 people. Surely we don’t need to do separated sample sizes with a group that small, right?

It’s a shame, because this poll is a totally unique and objectively cool concept. I want to know what NBA GMs think about these questions. Sure, I could tell you what I think the answer is to all of them, but I’m not holding the keys to Victor Wembanyama’s locker. NBA GMs are the only people alive with the power to act on these types of basketball superlatives, and I want to know what they actually think about them in a coherent and well-produced manner, rather than this barnyard explosion of meaningless data.

Quotes are a must, perhaps as simple as asking GMs to explain their answer rather than just writing a name in a blank, a staple of most tests in school after like… the fourth grade. They could write as little or as much as they’d like, but I’d at least like some footnoted theories about why on earth Tyron Lue is seen as making “the best in-game adjustments.” I don’t know what that means, but NBA GMs actually do (or at least they’re paid to)! I want them to tell me!

I would also appreciate more robust explanations of methodology, information on poll groups, when these questions were asked, and reminders throughout. Picking out some trends, explaining them, positing why and showing that an actual effort was made to interpret the data set would help non-data scientist readers like myself feel more connected to the article. Make it a piece of writing with understandable underpinnings, not just a presentation of a raw data left up to reckless interpretation by crazed fans like yours truly.

This is an indisputably cool idea, and needs to be treated as such. And if the NBA could just crank up the give-a-crap meter a few notches, the annual GM poll will be more than a slew of personalized grilled cheeses for all of us to argue over the meaning of their golden crispness.