BASKETBALL AI
Welcome to this edition of the Basketball AI newsletter.
Basketball analytics are being transformed by AI. Machines are learning the Triangle Offense. AI has become the ‘Ghost in the Game’ uncovering new insights for all facets of a basketball organization.
In this issue:
How Top College Basketball Programs Structure Their Analytics Departments
Four Core AI Terms In Action
4 Key Analytics to Winning
Basketball Innovation at CES26
Most Undervalued NBA Player
Let’s dive in.
LATEST DEVELOPMENTS
COLLEGE ANALYTICS
How Top College Basketball Programs Structure Their Analytics Departments
Image source: Manus / Basketball AI
Overview: Top college basketball programs have evolved from having a single "video guy" to operating sophisticated multi-layered data departments. The structure usually involves a blend of dedicated coaching staff, student run labs and university wide AI research partnerships.
Details: Here are a few top college program analytics department structures.
The "Bamalytics" Model (University of Alabama) - Alabama is arguably the most public facing example of a data first culture led by head coach Nate Oats.
Structure: Director of Scouting and Analytics (currently Adam Bauman) oversees a specialized student cohort called "Bamalytics."
The Workflow: Unlike many teams that only analyze game film, Alabama tracks "micro-data" during every single practice.
AI Use Case: They use AI enhanced tracking to label specific bio-mechanical details such as a player's footwork on a three-pointer (e.g., "was the foot in the pocket?"). Data is fed into real time dashboards so coaches can adjust practice intensity or shot selection immediately.
The Academic-Athletic Hybrid (UConn & Florida) - UConn and Florida have created formal bridges between their elite basketball programs and their University’s Data Science departments.
Structure: UConn’s Sports Statistics Experiential Learning Program pairs graduate and undergraduate data science majors directly with teams.
AI Use Case: Interns use computer vision to track player positioning and movement through game footage automatically tagging "penalty corners" or "ball screen coverage" to identify inefficiencies.
The "HiPerGator" Advantage: The University of Florida uses its HiPerGator supercomputer to run massive AI simulations predicting how different practice "workloads" (tracked via wearable sensors) will impact a player’s performance in the final minutes of a high stakes game.
The "Director of Scouting & Analytics" Role (Duke & Kansas) - These programs have standardized a high level coaching position that acts as a "translator" between data and the bench.
Structure: Duke’s Director of Scouting and Analytics (Zach Marcus) manages the flow of information from the Duke Sports Analytics Club to the coaching staff.
AI Use Case: During the season they use machine learning models (like those from EvanMiya) to calculate "Bayesian Performance Ratings." This allows them to see the true defensive impact of a player that doesn't show up in a box score such as how many points per 100 possessions an opponent's efficiency drops just by a specific player being on the floor.
Why it Matters: The evolution of top basketball programs has accelerated with the integration of AI, shifting from descriptive (what happened?) to predictive (what will happen?) and prescriptive (what should we do?). Top programs leverage AI for real time adjustments during games, optimizing player load management through biometrics and predicting transfer portal success. Strategic decision making and player development has transformed from reactive to proactive.
BASKETBALL AI TERMS
Four Core AI Terms In Action
Image Source: Manus / Basketball AI
Overview: Basketball AI uses advanced algorithms and computer systems to analyze vast datasets ranging from player body movements to historical performance trends. By automating the collection and interpretation of data, AI provides a level of precision that traditional coaching and manual stat tracking simply cannot match.
The Details: The Four Core AI Terms
Machine Learning (ML) is the "brain" behind predictive analytics where ML models are fed years of historical game data to identify patterns.
Usage: Teams use ML to predict game outcomes and shot success based on variables like defender distance and clock pressure. It also powers "DraftGPT" style models that evaluate prospects by comparing their college metrics to historical NBA success stories.
Computer Vision technology allows computers to "see" and interpret video footage.
Usage: Systems like Second Spectrum use cameras in the arena rafters to track the $X, Y, Z$ coordinates of the ball and all 10 players at 25 frames per second. This generates heat maps and gravity scores showing how much space a shooter like Stephen Curry creates just by standing on the perimeter.
Neural Networks & Deep Learning are complex algorithms modeled after the human brain capable of recognizing nuanced "off-ball" actions.
Usage: Deep learning is used to classify specific play types, such as a pick and roll or a pin down screen, without human tagging. It also powers bio-mechanical analysis, identifying slight hitches in a player’s shooting form that could lead to injury or decreased accuracy.
Generative AI is the newest player on the court creating new content and simulations.
Usage: Coaches use GenAI to run "what-if" simulations generating thousands of potential outcomes for a specific late game play. It also powers personalized fan experiences, such as AI generated highlights tailored to your favorite player.
Why it Matters: AI bridges the gap between raw data and actionable strategy. For players, it means personalized training that reduces injury risk. For coaches, it provides real time tactical adjustments during a game. Ultimately, AI doesn't replace the human element of the game but illuminates the hidden details that can make a difference between wins losses.
GAME STRATEGY
4 Key Analytics to Winning
Image Source: Gemini Nano Banana / Basketball AI
Overview: In May of 2020 I interviewed Coach Butch Carter on my Sports Analytics Podcast where he discussed an analytic model for winning basketball games that He and Mike Ellis created. The model worked so well in the 1999-2000 season that as head coach of the Toronto Raptors Coach Carter broke a 54 year old NBA record taking the team from less than 20 wins in 18 months to the playoffs. That record still stands today.
The Details:
The Carter-Ellis model broke down winning basketball games down into a framework of 4 key analytics:
#1 – Win the 1st and 3rd quarter of games. NBA teams that win the 1st and 3rd quarter win 80% of regular season games and 85% of playoff games. Including 1st and 3rd quarters strategies in practice sessions is important.
#2 – Stop scoring runs and/or create scoring runs.
#3 – Get back on defense and match up or officials will let you lose the game. Official are a key component of the model and must be factored in.
#4 – After timeout (ATO) execution on both offense and defense.
I was curious if this model was still valid today in the world of AI simulations and machine learning studies.
As mentioned, the Carter-Ellis model simplifies the game into four "must-win" segments. AI simulations from firms like Sportradar and academic studies (e.g., Frontiers in Psychology) show that teams focusing on these high probability predictors achieve win rates of 75–85%, far exceeding the 50–60% accuracy of traditional models.
The Quarter Factor: Research in 2025 confirms that the 1st and 3rd quarters remain the ultimate "alarm states." Teams winning the 1st quarter see a significant correlation to final victory, while AI models identify the 3rd quarter as the most critical "momentum pivot" of the game.
Momentum & ATOs: Recent machine learning studies (using XGBoost and SHAP algorithms) rank After Timeout (ATO) efficiency and Stopping Runs as top predictors of success. AI now quantifies "Momentum Shifts" as a rate of score change per unit of time proving that a successful ATO does more than add points. It resets the team’s psychological "decline."
The Transition "Official" Link: Modern tracking data shows that "unmatched" transition defense correlates directly with high foul rates. AI assisted officiating tools penalize late positioning more strictly, validating Carter’s focus on matching up to "stay on the right side of the whistle."
Why it Matters: In an era of high volume three point shooting and fast paced play, the Carter-Ellis model provides a scaffold for focus. By focusing on these four analytics, coaches can use them as a framework to win more games.
NEW PRODUCTS
Basketball Innovation at CES26
Image Source: ces.tech
Overview: CES26 (January 5–8, 2026). The most powerful tech event in the world showcased some advanced basketball technology continuing the move from simple stat tracking to Physical AI systems that act as active participants on the court. The following basketball related products, keynotes, and innovations debuted at this year's show:
Details:
Flagship Basketball Product: LUMISTAR "CARRY" AI Training System. This was the standout basketball hardware of CES 2026. LUMISTAR moved beyond traditional "ball launchers" to create a robotic training partner.
Active Defense: Uses a 4K dual camera system and computer vision to track player movement in real-time. It doesn't just pass; it uses gesture recognition to "read" the player and adjust its positioning to simulate defensive pressure.
Real Time Form Analysis: Analyzes shooting mechanics and arc during the workout, providing instant audio feedback.
Kickstarter/Launch: The "CARRY" system is scheduled for a Kickstarter launch in April 2026.
Keynotes & Industry Panels:
George Hanna (CTO, LA Clippers): Discussed the integration of AI into the Intuit Dome (Clippers' new arena) focusing on how AI powered cameras are automating instant replays and fan engagement metrics.
Matt Fleckenstein (CPTO, Genius Sports): Detailed the shift toward GeniusIQ, an AI layer that converts live video of basketball games into 3D data in milliseconds allowing for "augmented" broadcasts where fans can see player speed and shot probability in real-time.
Genius Sports "Executive Breakfast": Genius Sports hosted a dedicated track focusing on "The Modern Fan." They showcased how "Data Capture" is moving from manual entry to fully autonomous AI tracking where every dribble and foot plant is recorded without human input.
Notable Innovations & Articles:
Akta AI-First Video Platform: Akta (built on Google Gemini) demonstrated a new feature that allows broadcasters to search within live basketball footage. For example, a producer can prompt the AI to "Find every pick-and-roll involving Player X from the first half," and the AI generates the clip instantly for the broadcast.
Amazfit Nutrition Tracking: While not a "basketball" product per se, Amazfit previewed nutrition tracking concepts designed for high intensity interval athletes (like basketball players) to manage recovery during tournament play.
"Redefining the Business of Sport": Casey Wasserman (LA28 Chairman) discussed how AI will be used to personalize the fan experience for basketball during the upcoming 2028 Olympics stressing that AI must "accelerate communication" between players and fans.
NBA PLAYER VALUE
Most Undervalued NBA Player
Image source: okcthunderwire.usatoday.com / Manus
Overview: The NBA season is half over. Who is the most undervalued player so far this season? While names like Jokić and Gilgeous-Alexander dominate the MVP conversation, Isaiah Hartenstein is quietly putting up historic efficiency numbers for the Oklahoma City Thunder. Despite playing a "traditional" big man role, AI driven impact metrics suggest he is the most undervalued asset in the league relative to his national recognition.
The Details:
According to current 2025-26 advanced analytics, Hartenstein is an "impact unicorn." The data reveals:
LEBRON Metric: Hartenstein currently ranks in the top 5 globally in the "LEBRON" impact stat ($+4.11$), trailing only superstars Jokić and SGA.
Lineup Dominance: OKC’s Net Rating increases by nearly +12.0 when he is on the floor largely due to his elite rim protection and "short-roll" playmaking.
Efficiency: He is recording a career high Defensive Box Plus-Minus (DBPM), making him a statistical twin to Rudy Gobert on defense, but with the passing touch of a secondary playmaker.
Why it Matters: In the modern NBA, "value" is often equated with high scoring. However, AI models prioritize winning probability contribution. Three key Hartenstein value factors.
Contract Value: While his salary is significant, his production mirrors that of a Max-player in terms of "Win Shares."
The "Jokić Lite" Effect: His ability to facilitate from the elbow allows OKC to run elite offenses without a traditional point guard on the floor.
Modern Defense: He is one of the few centers whose Defensive FG% at the rim remains elite without compromising the team's ability to switch on the perimeter.
QUICK HITS
🛠 Tools Spotlight:
Tools mentioned in this edition:
Sportradar - Sports technology.
LUMISTAR “CARRY” AI Training System - Integrated AI training robot with camera system that analyzes shooting form and trajectory in real time while tracking performance trends.
GeniousIQ - Next generation data and artificial intelligence platform. Powerful machine learning and AI large scale data for game analysis, ultra-immersive experiences and ways to reach and engage sports fans.
Akta AI-First Video Platform - Allows video operations to function with greater automation and responsiveness, reducing manual tasks and improving decision-making speed.
AmazFit - Nutrition tracking designed for high intensity interval athletes (like basketball players) to manage recovery.
📰 Everything else in Basketball AI:
Expert Insight: "AI is not going to replace humans, but it's going to help augment humans... the sports analytics job market will not completely disappear due to the implementation of AI."
— McCormick Professor V.S. Subrahmanian, NNCI founding co-director (January 14, 2026)
Teamworks Acquires Sportlogiq (January 19, 2026) The sports tech giant Teamworks announced its acquisition of Sportlogiq, an AI-powered analytics platform.
The Tech: The acquisition integrates Sportlogiq's advanced computer vision and automated player tracking into the Teamworks ecosystem.
Impact: While historically dominant in the NHL, the expansion is targeted at providing basketball teams with a "complete data and analytics platform" for talent evaluation and real-time game strategy.
📆 Upcoming Conferences, Events and Guides:
January 6, 2025 Last issue of Basketball AI
February 22-24, 2026 National Sports Forum (NSF)
March 6-7, 2026 MIT Sloan Sports Analytics Conference (SSAC)
April 10-11, 2026 Connecticut Sports Analytics Symposium (CSAS)
That’s a wrap for this edition. The next newsletter will be published Tuesday February 3, 2026.
Tell us what you thought of today's newsletter to help us improve it for you.
Until next time, Terry