How Basketball is Using AI in Cool Ways

AI seems like one of those ethereal concepts that have always been around, but it’s still kind of hard to explain just what it does.

Movies make it seem like robots are the most terrifying thing ever. Something like Ex Machina made me scared to ever go to an evil genius’s private island. Meanwhile, Upgrade had me wondering if I’d ever get a computer chip planted in my brain.

But AI and machine learning aren’t totally scary. In sports, they’re downright useful. Here are three cool examples of AI and basketball in action.

Stats Perform is Making Broadcasts More Entertaining

I’ve written about sports for more than a decade, and the vast amount of data out there is simply bananas.

While I can rattle off certain stats with no problem — like, say, the final score of Northwestern State’s upset victory over Iowa during the 2006 NCAA Tournament (it was 64-63, in case you’re curious) — it typically requires digging through pages of information to find exactly what I need.

It sure would be handy if AI could help with that.

Oh, and it can! Thanks to Stats Perform’s PressBox platform, sports journalists, broadcasters, social media teams, and other content creators can use AI to find data-driven stories, delivering better broadcasts and content for fans.

The platform includes PressBox Live for editorial coverage, PressBox Video for instant reactions from players (with an editing interface that Instagram could borrow a page from, JUST SAYING), and PressBox Graphics to create stellar designs that utilize more than 40 years of data.

For example, the University of Memphis men’s basketball team uses PressBox Graphics to improve its workflow and social media presence. The digital team drops an opponent into one of their templates and the stats will update in real-time.

No more digging around for how many points a player or team scores per game, what their shooting percentage is, or even what other teams are in your conference.

Hey, with all the conference realignment lately, it’s tough to keep up! I would have never thought UCLA and USC, two teams that are notably located in Southern California, would ever join the Big 10 conference, which consists primarily of schools east of the Mississippi.

I’m all about making life easier, and Stats Perform is checking that box on both ends: for journalists and broadcasters, and for fans wanting to learn more about their teams.

Watch the full virtual launch of the latest and greatest with PressBox here, with basketball goodies kicking off around 7:15.

Vanderbilt University is Helping Players Improve Their Shot

Shooting a basketball on its own isn’t necessarily hard. But shooting a basketball in rhythm during a game is a different beast entirely.

I know I’ve had moments when I’m playing where every shot feels great. I’m like Prince in the classic Chappelle’s Show sketch, yelling out “GOOD!” before the ball swishes through the net.

Then I’ve had other days where everything feels off. The ball slips out of my hand like I’m throwing a wiffleball, and it’s just a mess.

So if I get some consistency in my shooting game, that would be GREAT.

Some researchers at Vanderbilt are aiming to do just that. Well, they’re aiming to help their own players more than me, but still, I could probably benefit in the long run.

Vanderbilt’s team is working with NOAH basketball, which has delivered more than 50,000 hours of footage.

NOAH’s software IDs players using facial recognition and computer vision. The software offers interactive shot charts and data, featuring things such as arc, depth, shooting percentages, and consistency.

Using a type of AI software known as a temporal relational network, the researchers could determine which players were taking what types of shots, with a classification accuracy of 96.8%.

In the course of a game, it’s important to know if a player is catching a pass from the left or right, if they’re shooting off the dribble, if they’re playing hopscotch with a friend before taking a jumper, and much more.

Plus, in the video below, you get to see one of the researchers demonstrating the different shot types in slow motion while wearing jeans and loafers. That’s very fun.

Toyota’s CUE Robot Hits Free Throws With Nearly 100% Accuracy

In the movie Pleasantville, basketball players are very good. The fictional land is so…ahem, pleasant, that it’s impossible to actually miss a shot.

Then, a young Tobey McGuire causes a young Paul Walker to become so distressed that he chucks the ball at the hoop. And after spinning around the rim several times, the shot misses. It’s pandemonium.

Pleasantville may not have perfect free throws anymore, but there’s a robot that’s picking up the slack.

CUE is a basketball-shooting robot developed by Toyota volunteers in their free time, and it’s become an impressive display of shooting touch, a beautiful portrait of AI and basketball.

CUE uses sensors in its torso to calculate the exact distance from the hoop to make shots. Armed with that information, it determines the best angle and power, and ends up making nearly 100% of its shots.

However, as the video below shows, CUE takes a LONG time to get a shot off. By the time it’s fully calculated the right trajectory and releases the ball, it’s been 10-15 seconds — more than enough time for a defender to run to the bench, drench themselves in Gatorade, and skip back onto the court.

Still, it’s a super cool application of the potential. We saw CUE wowing crowds at the Olympics. As it gets smarter, it might even be able to help people with their jump shots, even if CUE can’t jump itself.

https://www.youtube.com/watch?v=cG8SvzSDB08

Did you get this article shared with you? Subscribe to Crisp Bounce Pass here!