Analytics is overrated in a recent article on Sport Techie Mark Cuban discusses the future waves of technology and what is the new drivers for sports teams.
‘What’s the biggest overrated and underrated technology in sports right now?’
“Overrated is analytics. Underrated is artificial intelligence and its derivative.”
Why does Cuban think the way he does?
“Because it’s become efficient now and because AI is going to kick (analytics’) ass all over the town. It’s going to be not gone because analytics becomes your AI. Your analytics department will have to learn AI, machine learning, deep learning, neural networking, etcetera, and be good at that. But historically, when analytics, and this is going back to the day I bought the Mavericks when we started using advanced plus minus, you have to make choices. And you have to be able to say, ‘OK, what are important to me? How am I going to weigh those variables and what are the conclusions I’ll draw from those variables?’ At the beginning with the Mavs at least, we had huge advantages. We would force teams to play their bad lineups that they didn’t know were bad. We forced teams to play bad combinations that they didn’t know were bad combinations, and it helped us win games and playoff series.
“Over time, teams became a lot more aware, and the market became a lot more efficient. And then it didn’t become so much about the analytics as it became, ‘OK, could you find the players you needed to play the style that you wanted to play?’ And that to this day remains the biggest challenge. You can have the greatest analytics department in the world, it’s going to be very difficult now to get an advantage over the other team. Maybe there’s a couple teams that are behind in analytics, but you’re not going to get an advantage on smart teams.”
With not many NBA teams that are behind in analytics, Cuban called it an “efficient market” now.
“Going forward, I think what data you incorporate into your models, how you clean your data, how you train your data, whether it’s structured, unstructured, how you label it, all those things will be critically important,” Cuban continued. “The hope, at least from my perspective, is we’ll learn things and see things that we never imagined, that we couldn’t perceive in a traditional analytics environment.