AWS re:Invent 2025-NBA Inside The Game, Powered by AWS - Building the NBA’s New Stats Program-SPF307

Summary of AWS re:Invent 2025 - NBA Inside The Game, Powered by AWS

Overview

• The NBA and AWS have partnered to develop the "Inside the Game" platform, which leverages advanced analytics and machine learning to provide new insights and metrics for fans, broadcasters, and teams.
• The presentation covers the evolution of NBA stats, the technical infrastructure powering the new analytics, the process of developing novel metrics, and a deep dive into the new "Player Gravity" statistic.

The Evolution of NBA Stats

• The NBA has progressed from basic box score stats (points, rebounds, assists) to play-by-play data, and finally to granular player tracking data with 29 data points per player recorded 60 times per second.
• This wealth of data has enabled the development of more advanced analytics to quantify different aspects of the game, moving beyond just scoring and counting stats.

Technical Infrastructure

• The NBA's stats platform runs on AWS, leveraging services like Amazon S3, Amazon EKS, and Apache Flink to ingest, store, process, and deliver the real-time player tracking data.
• This infrastructure allows for low-latency (50ms) processing and analysis of the massive amounts of data generated during games.

The Stats Ideation Process

• The NBA stats team works closely with fans, broadcasters, and teams to identify gaps in existing stats and areas where new metrics could provide valuable insights.
• Key considerations include:
  • Measuring aspects of the game that are intuitively understood but not quantified (e.g., defensive impact, off-ball movement)
  • Leveraging the available data to create meaningful and actionable statistics
  • Ensuring the new stats are both technically robust and easy for fans to understand

Player Gravity

• "Player Gravity" is a new metric that quantifies how much attention a player draws from the defense, even when they don't directly score.
• It is calculated by comparing the actual defensive pressure on a player to the expected defensive pressure based on their positioning and the ball location.
• The metric is powered by a machine learning model that uses a transformer architecture to predict the expected defensive pressure in a given situation.
• Key benefits of Player Gravity:
  • Identifies players who create open shots for teammates by drawing extra defensive attention
  • Provides context around team strategies and how defenses are responding to certain players
  • Can be used to analyze historical data and compare players across eras

Future Developments

• The NBA and AWS plan to release additional advanced stats in the coming years, including:
  • Defensive Box Score: Quantifying a player's defensive impact beyond just steals and blocks
  • Shot Difficulty: Measuring the expected field goal percentage based on factors like defender proximity, shot type, etc.
  • Leverage Score: Evaluating a player's overall contribution to their team's chances of winning
  • Handle Score: Analyzing a player's ball-handling skills and ability to create separation
• The stats will also be expanded to cover the WNBA, with a focus on metrics around team chemistry and off-ball movement.

Business Impact

• The new analytics are designed to enhance the fan experience by providing deeper insights and understanding of the game.
• Broadcasters can leverage the stats in real-time to enhance their commentary and storytelling.
• Teams can use the data to inform strategy, player evaluation, and player development.
• Overall, the partnership between the NBA and AWS aims to push the boundaries of sports analytics and data-driven decision-making.