Math and Data on the NBA Court

By Math Values Editorial Board

Steven Angel, a Senior Vice President in the National Basketball Association

Data plays an ever-increasing role in our world, which is seen specifically in sports. College coaches increasingly use data, in some cases leaning on undergraduate and graduate school programs and clubs for help with the analysis. Data analysis is an ever-present, ever-growing application of mathematics. To learn more about how math and data are used in the National Basketball Association (NBA), Math Values posed questions to Steven Angel, a Senior Vice President in the league.   

Question 1: Can you start by telling us about your job as Senior Vice President, Head of Game Analytics and Strategy at the National Basketball Association and the role mathematics plays? 

My group’s primary area of focus is the monitoring of referee performance, game integrity, and how the game rules affect the game and are being enforced. Our team collects rating data on every call and over 300 non-calls each game, which is compiled to analyze patterns and trends in our areas. We use these analyses with various departments, including Referee Operations, Basketball Operations, and others, to provide meaningful and timely information to drive decision making and make changes where needed. 

Question 2: Given you are working with data and officiating in the NBA, what can and possibly can’t be quantified with data? How do these advantages and limitations of data and mathematics play into your work?  

On one hand, we are able to quantify a significant amount of referee performance and game play information to support decision making. For example, we recently developed algorithms that use player tracking data to test our referee mechanics – that is, the positions referees take and rotate through on the court depending on the location of the ball and play action. Using this, we are able to provide the referee staff with tactical information concerning where they should stand or when they need to move given the situation to improve accuracy. On the other hand, while we can measure certain aspects of performance, a basketball game is a complex system, and there are a lot of factors that influence the perception of performance. These factors include obvious things like call and no-call accuracy, especially late-game accuracy, but also referee poise, game management and team interactions, and harder to define factors, such as trust, among others. 

Question 3: Data plays an ever-increasing role in our world. What advice do you have for students studying mathematics? For those interested in working in data? And for those who may not plan to work in data? 

It is my experience that those studying mathematics and looking for a job involving data are best served focusing on the skills and later on, learning the content. The skills of a data scientist involve data prep and management, statistics, visualization, machine learning, and so on. Becoming fluent in these areas, regardless of industry, is important because while data sets may be industry-specific, the tools and methodologies we use to analyze them are agnostic. For those not working directly in data, you will be working indirectly in data. There are no industries and virtually no jobs where knowing basic statistical concepts won’t be an asset. Everyone should have a fundamental understanding of data and be a critical thinker, able to understand what is being discussed or presented, identify flawed logic or methods and how to ask meaningful questions.

Question 4: Suppose a student has a dream of a job in data, which, for some, could be working in the NBA as a data analyst. What advice do you have as they look for that first job and as they are in college preparing for their first job? 

I can only speak from my own experience, which is where I developed my thinking in Question 3. Jobs in sports are really tough to land. Don’t be discouraged. Your first job is likely to be either an entry level job, where you might feel underutilized, or something not in sports. Stay focused and build your reputation. If you take a lower-level job in sports, use it to build your skills, but also to network. If you take a job outside sports, don’t sweat it at all – build your skills. The scientists I’ve hired or considered hiring have come from diverse backgrounds and have wowed me with highly complex problem solving abilities in other industries. We have interviewed everyone from insurance analysts, with experience detecting subtle anomalous patterns in data, to astrophysicists, who are experts in spatial analyses of star movement (no pun intended) because of our player tracking tools.

Question 5: Are there any pitfalls in data science?

It’s subtle, but there are some data scientists who can do the work, but I question their true understanding of it. This sometimes manifests in fundamental things, such as really knowing what a method is supposed to tell you, including what it can’t tell you. Sometimes it comes up in more philosophical ways (especially when moving from academia to industry), such as getting so lost in an analysis that the practical implications of it aren’t useful. Statistical significance doesn’t always translate to tactical information.

Question 6: Do you believe data analytics can make watching an NBA game more fun? If so, how? If not, why?

Ha. If you want to take the fun out of something for someone, pay them to do it.

Question 7: Any final words for professors and students of mathematics?

Getting good at math is about practice. Find nice big data sets to play with, in sports or not. Set up a problem to solve or question to answer. You’ll waste a ton of time digging into dead ends, but you’ll become a better practitioner in the long run.