Basketball Nerd Bragging Rights Were On The Line At The Recent NBA Hackathon

NEW YORK– The NBA is near the forefront of statistical analysis among major American sports leagues, and with the success of the MIT Sloan Sports Conference, co-chaired Rockets general manager Daryl Morey, they’re continually trying to expand on an already incredible breadth of quantitative and qualitative basketball data. In an effort to build on that technological emphasis, specifically as it pertains to future generations, the NBA’s Director of Basketball Analytics, Jason Rosenfeld, organized the NBA’s inaugural Hackathon this past Saturday at Jordan Brand’s cavernous Terminal 23, hidden along 32nd St just west of Herald Square.

We were in attendance early Saturday morning and through blurry eyes, we watched commissioner Adam Silver and then Rosenfeld address a gym filled with some of the brightest college minds in the country.

After 450 undergraduates, Masters and Ph.D students applied, less than half — 210 — were offered an invitation as part of 60 teams competing in the all-day event. Over 50 universities were represented, including traditional behemoths, Harvard, M.I.T, Stanford, Princeton, Carnegie Mellon and Yale. Virginia led the pack with 18 students in the contest, but it was a nice mix of major Ivy institutions and little-known liberal art and technical colleges. Regardless of where they went to school, an invite to the Hackathon is a big win for these future engineers and scientists.

“Analytics is becoming more and more important in sports, especially in basketball,” Rosenfeld told when he wasn’t dashing all over the place as the event tipped off. “And we thought this would be a great way to engage students, get some new ideas and meet a lot of young, smart young people who are interested in basketball, and a lot of whom want to work in basketball.”

Rosenfeld can empathize with those students. He said he’d have “definitely” applied to the event if it had been around when he was in school, which actually wasn’t that long ago. Jason graduated from Harvard in 2011, and he was a part of the very first Sloan Analytics Conference, which he said is when he first met commissioner Adam Silver, and which in turn acted as a bit of a springboard to his current position as the Director of the NBA’s Analytics.

It was clear talking to Rosenfeld that this Hackathon event was more an opening for the students than it was a conduit for the league to cherrypick the best analytical tools before what they develops possibly becomes proprietary information.

“This is such a good opportunity to get exposure,” he said. “It’s a fun for us, and there are obviously a lot of benefits for us, but it’s such an amazing platform for them to show what they can do. The problem (with) getting into basketball analytics (is)…teams get thousands of resumes. It’s often hard for them to be able to tell who is really good and who’s not. This is a way for them to show what they can do.

“This is almost like a 10-hour interview process,” he added.

While that might sound dreary for some NBA fans, the students we talked to seem incredibly excited, and if you’re an enterprising college student who wants to be on the inside track for a sports analytics job in the NBA once you graduate, this is simply another opportunity to get on the Association’s radar — just like Rosenfeld did for over half a decade ago.

But — the question remained. Who would win, and what topic would they broach to claim victory?

Here were the four prompts teams could choose from as they were displayed at Terminal 23:

Most of the student groups we talked to elected to choose the first prompt about defense, and that’s certainly the silver bullet that NBA analytics has so far been unable to answer. But, like all good competitions, there was some gamesmanship at work when selecting a topic to work on that day.

One student, David from South Florida, told me that his group “thought a lot of people would choose one and three, so we thought we were going to do one that was more difficult and was more challenging.” So David and his team choose to do prompt three, building a tool to predict the outcome of shots taken. Except, they didn’t win, or even make the final five.

The teams worked until 5:15 p.m. on Saturday evening to develop solutions using only what they had derived that day: no pre-made code or analytics tools were allowed. (The students were given some hints about the prompts when they were originally selected to compete, but that was it.)

Among the 60 teams who entered, five finalists were selected to present at 6:15 that night, and five judges narrowed them down to the last three winners.

Those judges were Detroit’s Manager of Basketball Analytics, Zach Bradshaw, ESPN scribe Zach Lowe, professor of Business at Columbia University, Mark Broadie, the Associate Director of Basketball Strategy at the NBA, Hao Meng, and the Assistant General Manager and Team Counsel of the Celtics, Mike Zarren.

To be honest, there are simply too many variables to ever adequately answer defensive efficiency, but it was fun imagining a Good Will Hunting character solving such a monumental riddle on a lazy, early-autumn Saturday. Unfortunately for my imagination, defense wasn’t the prompt that won.

The Case Against Hero Ball entry, presented by Alex Wainger (Columbia), Ella Kuzmenko (NYU), William Robbins (Pace) and Zac Robertson (Columbia) was an open topic that won them the first-place prize: a lunch with the league office staff, a pickup basketball opportunity at Terminal 23, free tickets to an NBA game, and $800 of NBA store credit. (Second place got free tickets to an NBA game and $700 of store credit, third place got $600 of store credit).

The winning entry provided data to support a thesis most devoted NBA fans already know: Hero Ball isn’t very effective on a night-in, night-out basis, but it does tend rear its ugly head a lot more when the stakes rise in the playoffs.

The group examined the behavior of teams in the regular season vs. a team’s behavior in the playoffs. By using “Hero Ball” as a symptom of playoff basketball, they explored when and why teams tend to resort to it during the spring season. From that data they extrapolated that teams clearly change their brand of hoops in the playoffs, and end up holding the ball longer, dribbling it more often, passing it less while scoring fewer points per possession. To be honest, Scott Brooks could have used this winning presentation when he was coaching the Thunder.

If there was any question as to whether the NBA was prepared to make the next step in the analytical revolution, the NBA Hackathon provided a pretty compelling answer in the affirmative.