Conference Presentation, Lecture
How data transformed the NBA
- The Houston Rockets transformed from mid-table mediocrity to championship contenders over the last decade by leveraging big data to secure a statistical advantage.
- Daryl Morey, the Rockets' computer scientist and analytics leader, identifies winning as a function of two variables: points per possession and the quantity of possessions generated.
- The Rockets pioneered the use of video tracking systems ten years ago, which revealed that three-pointers and dunks offer a superior return on investment compared to long-range two-point shots.
- Analysis showed that making one-third of three-point attempts yields the same point value as making one-half of two-point attempts, rendering the long-range two-point shot statistically inefficient.
- In the 2017–18 season, the Rockets set an NBA record by making more three-pointers than any other team in league history.
- Rockets coach Steve Kerr emphasized that the team's success relies on balancing data insights with human instinct to ensure alignment on every possession.
- Analytics company Second Spectrum, co-founded by Professor Rajiv Mahesharan, now tracks 3D spatial data for all 30 NBA teams at 25 frames per second using machine learning.
- Machine learning models initially recognized 20 basketball terms but now identify over 500, effectively integrating the collective knowledge of elite coaches into algorithms.
- Advanced analytics quantify previously qualitative feelings about shot probability by accounting for player movement, shot type, and defender positioning.
- Recruitment strategies have shifted toward selecting leaner, more agile, and highly skilled players over larger but less skilled athletes to maximize data-driven efficiency.
- The organization claims that data-driven decision-making provides a roughly 5% edge, which translates to a significant competitive advantage in winning margins.
- Real-time data processing is accelerating, with analysis times dropping from hours to minutes, with the industry anticipating fully functional real-time availability for coaches soon.
- Experts predict the sport is in a "transformational stage" where data will soon be used live during games to influence in-game tactics.
- Daryl Morey stated that the league has "not scratched the surface" of the potential applications for data analytics in professional basketball.