Sloan 2018: Something For Everyone
Ask any five people who attended the Sloan Sports Analytics Conference in 2018 about their experience and you’re likely to receive five different answers. Not because the quality has suffered – this edition featured some of the best content, across the board, in recent years – but because Sloan is something different to everyone. Two people who attend the same event in the same year and see the same number of panels and speakers could meet up for a beer afterwards at Bukowski’s Tavern and realize they didn’t see any of the same material.
Down one hallway you’ll find the commissioners of both the NHL and MLB fielding questions in front of a thousand attendees. In another room students will learn how to code in R or present research papers featuring NBA defenses created by artificial intelligence. Next door to that will be a sponsored presentation by Kitman Labs or Fusion Sport offering Sports Science solutions to the teams in attendance, and outside those doors is a trade show full of startup companies. That’s all on one wing of the Boston Convention and Exhibition Center. Head to the other side and there’s a full eSports component.
Oh, and there’s Barack Obama, too. We’re just not allowed to say anything else about him.
Whatever your interests or goals, Sloan is what you make of it. It’s still made up predominantly of white males, but each year it becomes incrementally less so.
While Sloan has always been geared towards hoops, being the brainchild of NBA executives, the basketball track still had more to offer than ever. There’s the annual Basketball Analytics panel that, despite sporting exceptional participants each year, usually manages to be a forgettable affair simply due to a lack of specificity in the questions and discussion. Even then, putting old friends Daryl Morey and Mike Zarren together to riff is hardly the worst way to spend an hour, with Morey offering an insightful kernel in response to a question about the demise of the center position.
“I think it’s the single skill guys [that are gone]. You used to be able to be an absolute rebounder, or an absolute defender.”
(As a quick aside, Morey also made the point about mid-range shots being more helpful late in games when you simply need any points, not the highest value shots, as we discussed regarding Goran Dragić last year. Looking at Houston’s rate on mid-range shots, shots they famously avoid, that point bears out.
Houston Mid-Range FGA Per 100
5 Min Left: 9.8
3 Min Left: 12.8
1 Min Left: 14.7)
Since nearly everyone speaking on a panel works, or worked, for a team or some other corporate entity, there’s only so much you’re going to learn as state secrets are protected. You’ll get some interesting details here and there, like Shane Battier, on a panel with Steve Nash and Morey regarding that advent of “Modern Basketball” noting that Memphis Grizzlies practices back in 2004 used to begin with running over all the different ways they were going to double team the post – when few teams run post-ups much at all these days. But most of what you’re going to get is just enough to get your mind working.
That’s what makes a panel like “NBA 2.0” so enjoyable, because it’s essentially a bunch of league lifers getting their own minds going as they pontificate on possible changes to the game – whether it be draft and/or draft lottery reform, a possible play-in tournament, on-court changes or looking at the playoff format. Thankfully they were able to reasonably address the tired ‘Are there too many threes’ question without invoking any sort of historical referendum, as Evan Wasch, Senior Vice President, Basketball Strategy and Analytics for NBA, did.
“We’re comfortable where we are,” Wasch said. “If we have 75 threes a game we have a problem.”
The underlying current to the conference, as it’s been the past couple of years, is all the work being done behind the scenes with artificial intelligence. Machine Learning has become something of a buzzword at the conference – and just about everywhere else in the world – but it’s for good reason when you read a research paper like ‘Bhostgusters’. What the group behind that project built (we won’t list everyone involved but here’s a link to the paper) is essentially a play tester for coaches, where they can digitally draw up an offensive play. So far, so good, but coaches have been able to do that for years. The next step in ‘Bhostgusters’ is that those prospective plays then actually play out against an A.I. that has been ‘taught’ how to play defense through all the optical tracking data the league has these days.
In other words, coaches are able to test out their ideas against a machine before ever getting to game action.
Now we’re still very early into the advent of this type of technology as far as basketball goes, but you can see where this might be going. Analysts have been doing player and team projections for years based on different types of available data, but what if every move a team is going to make is put to the test by a machine that has every data point in the history of the game available to it.
If we can teach a computer to play chess at a Grandmaster level, can we teach one to coach and manage at an NBA level?
That’s where the mind goes during Sloan – to the possibilities. The conference is something different to everyone who attends, but everyone who attends should come away from it thinking about something new, or thinking about something old in a new way. As an industry event it might not be perfect, but it encourages forward thinking no matter the climate of the game.
As Wasch said, “When things are going well that’s the perfect time to innovate.”