After falling to the eventual NBA champs during the Eastern finals, the Toronto Raptors are hungry for a championship title. Thursday’s draft will be crucial in crafting a winning lineup, and when it comes to deciding who makes the team, the Raptors will be able to consult their newest recruit: IBM’s Watson.
The partnership dates back to February, when the Raptors’ parent company Maple Leaf Sports and Entertainment announced they were partnering with IBM, making the Raptors the first NBA team to use the Watson supercomputer to analyze players.
Videos by VICE
“Watson doesn’t answer questions of who the best trade pick would be—rather it compares them on different dimensions,” explained Jon Lenchner, the scientist who led the IBM Sports Insights Central project.
IBMers collaborate through interactive touch screens as part of IBM Sports Insights Central. Image: IBM
For example, if the Raptors were measuring college basketball prospects, Watson could quickly crunch the numbers and display a comparison of their stats on shooting, assists, and rebounds. Compare that to drafts of past years, in which the Raptors would use whiteboards with player stats printed on magnets, and call up statisticians each time they wanted new information, recalled Lenchner, who visited the Raptors’ headquarters while IBM was developing the software. In the days before Watson, the whole process was much more laborious and time consuming.
“They realized they could do a lot better,” said Lenchner, who recently moved to Nairobi as the Chief Scientist of IBM Research Africa.
Even before Watson came along, the Raptors were tech-savvy. They have access to one of the NBA’s leading analytics teams, and have developed a wide range of tools, including a way to use data from the SportVU camera tracking system to model the best moves a player could make.
Still, they’re not looking to speak publicly about their work ahead of the draft, and you can’t blame them—there’s a lot at stake. The Raptors’ analytics team declined Motherboard’s request for an interview. The team actually hasn’t spoken publicly about using Watson since the initial launch, when general manager Masai Ujiri joked: “When the trades start going the other way I’m going to blame IBM.”
Fans gather at 2015 Toronto Raptors playoff game. Image: Toronto Raptors
The million-dollar-range salaries for picks are some of the lowest costs on an NBA roster. This year, the Raptors have two first-round draft picks and, historically, the rookies they choose will cost between USD $1 million and USD $2.5 million each for their first seasons.
This is still a steal by NBA standards, where the average contract for a player on the open market is about USD $5 million, and tops out at USD$25 million. Expect player costs to balloon even more in 2016-17, when the salary cap will reportedly go from $70 million to $94 million per team.
The real question is whether the player performs, fits with the team and—if he’s talented—sticks around, and Watson can’t measure these qualities, at least not yet.
The software used for NBA picks is based on how IBM does corporate acquisitions, which may seem like a leap, but it’s actually surprisingly similar.
Corporations want to acquire companies that excel in their specific field. Otherwise they have to reinvest, said Lenchner. Translated to the NBA, that means getting rid of players who don’t perform, and drafting ones who fit with your existing team or strengthen an area you’re going to focus on.
Critics sometimes complain that analytics only look at the numbers, and don’t take a player’s personality into account. But Watson’s insights can actually help bring more human elements into the draft pick, Lenchner said—for example, by using linguistics analytics to understand a player’s personality based of how they post on social media.
John Lenchner tests a new interactive solution powered by IBM Watson at the Toronto Raptors BioSteel Centre. Image: IBM
He’s become a huge basketball fan since working on the project. He praised the Raptors as a “very pragmatic and well-paced” team with tremendous point guards—and the rare problem of too many talented young players. This has left Ujiri with a dilemma.
“I don’t know if it’s a draft where we can go and get somebody who will impact our team right away,” he said at a Toronto press conference on Tuesday. “Is it a player that’s three, four years down the road and that player has a high upside? We have to look out for the organization that way.”
Don’t expect a major shakeup with the first Watson-era draft pick. The software measures players based on what the team’s management is looking for. “In the end, this system doesn’t make decisions,” Lenchner said. “We want the system to be a synergy between humans and machine.”
Basketball is a rich field for AI. Lenchner can imagine Watson developing the ability to analyze players’ health records and, eventually, to scout videos.
But he said the most important applications for Watson’s scientific approach to team-building might be off the court. It could branch out beyond the NBA to medical teams, he believes, and even to astronaut selection.
“[Watson could be used] for the mission to Mars,” he said. After all, that small crew will be crammed together on a spaceship for a few years at least, and getting along will be essential. “You can’t make changes once they’re up there.”
