## TL;DR
AI is transforming sports -- tracking every player and the ball at 25Hz across all NBA arenas, predicting injury risk from movement patterns, and using reinforcement learning to optimize tactics. From expected goals (xG) in soccer to Statcast in baseball, AI-generated metrics are changing how games are played, coached, and analyzed.

## Core Explanation
Sports AI stack: (1) Tracking -- computer vision (pose estimation + multi-object tracking) processes multi-camera arena feeds. Output: player positions, velocities, accelerations at 25Hz. Ball tracking: tennis (Hawk-Eye), soccer (goal-line technology), cricket; (2) Event detection -- from tracking data, ML classifies events: pass, shot, tackle, dribble. Automates what was previously manual coding; (3) Performance metrics -- xG (shot quality), xA (expected assists), pressing intensity, passing networks; (4) Predictive models -- injury risk (movement pattern analysis), player development (trajectory forecasting), game outcome; (5) Tactical AI -- RL learns optimal strategies from game data. Counter-tactic recommendation.

## Detailed Analysis
xG (Expected Goals): ML model trained on thousands of shots with features (distance to goal, angle, defender positions, shot type, assist type). Output: probability of goal. Used to evaluate finishing skill (actual goals - xG). Player tracking: NBA Second Spectrum (2023-present) uses optical tracking across all 30 arenas, providing 25Hz 3D player + ball tracking with skeletal pose data. Metrics: speed, acceleration, distance covered, defensive impact. Injury prediction: Kitman Labs, Zone7 analyze player load (training + match) and detect movement pattern changes preceding injury. Can predict non-contact injuries 1-3 weeks in advance with 70-80% recall. Tactical RL: Liverpool/DeepMind partnership trained models on thousands of matches to learn optimal pressing triggers and defensive positioning. Key challenge: the "open system" problem -- sports have too many variables for pure data-driven optimization. The best systems combine AI insights with coach expertise.