Heads in the game

2026-06-23 · Source: MIT Technology Review · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Intermediate, long

Summary

The MIT Sports Lab, cofounded in 2015 by Anette "Peko" Hosoi and Christina Chase, is a key player in sports technology, collaborating with organizations like FIFA, the NBA, NFL, and Adidas. During the 2022 Men's World Cup, the lab validated FIFA's Semi-Automated Offside Technology (SAOT), which uses 12 cameras, computer vision, and ball-tracking data to assist referees. SAOT assisted in over 150 offside calls, changing seven game outcomes. Beyond soccer, the lab developed the "Expected Action Value (EAV)" metric for the NBA to analyze mental performance in basketball, trained on 786,208 passes and 1.4 million shots. It also analyzed NFL stadium openings during COVID-19, finding no correlation with case spikes, and collaborated with Adidas to optimize 3D-printed running shoe midsoles using biomechanical models. The lab's work extends to benefiting the MIT community through various educational and athletic projects.

Key takeaway

For sports technology developers and data scientists aiming to impact professional leagues, focus on rigorous data validation and transparent algorithm design. Your solutions, like SAOT, must integrate seamlessly with existing human roles, providing clear, actionable insights rather than full automation. Consider how your models can quantify previously unmeasurable aspects, such as mental performance with EAV, or optimize physical products like Adidas's 3D-printed midsoles, ensuring real-world applicability and trust among stakeholders.

Key insights

The MIT Sports Lab applies data science and engineering to solve complex problems across professional sports, enhancing officiating, player performance, and equipment.

Principles

• Data validation is crucial for sports technology adoption.
• Simple biomechanical models can yield powerful insights.
• Technology should assist human decision-makers, not replace them.

Method

The MIT Sports Lab validates sports technology by analyzing high-volume tracking data, identifying anomalies, and developing protocols to synchronize diverse data streams for real-time assessment and algorithm refinement.

In practice

• Use SAOT-like systems for rapid, objective officiating decisions.
• Implement EAV metrics to quantify player mental performance.
• Optimize footwear design using gait response models.

Topics

• Sports Technology
• Semi-Automated Offside Technology
• MIT Sports Lab
• Data Analytics
• Biomechanics
• Player Performance Metrics
• 3D Printing

Best for: Computer Vision Engineer, AI Scientist, AI Engineer, Data Scientist, Research Scientist

Related on AIssential

• World Cup technology: from ref cams to AI analysts, cutting-edge research is changing the game — AI and advanced sensors are transforming football, enhancing officiating accuracy, fan experience, and team strategy through real-time data.
• 2026 MIT Sloan Sports Analytics Conference shows why data make a difference — Data-driven decision-making provides coaches and teams with confidence and strategic advantages in high-stakes sports.
• SportIQ expands connected sports ball technology with $6.2M raise — Embedded sensors in sports balls provide real-time performance analytics for training and professional applications.
• Former MLB Players Lead Ember Sports Mobile App, Bringing Pro-Level Training to Athletes Everywhere — Ember Sports democratizes professional baseball/softball training via mobile and VR, making advanced analytics accessible.
• The Rise of Sports Intelligence: How the Lakehouse Turns Tracking Data into Competitive Advantage — Unified data platforms are crucial for leveraging high-volume biomechanical data in professional sports.
• Sportway raises €20M as AI reshapes sports broadcasting — AI and automation enable economically viable, professional-quality sports broadcasting at scale for all levels of sport.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by MIT Technology Review.