Sometimes I get requests from high school students to answer questions. For example, Noah C. sent this list:

I’m writing with some questions for my final economics project. We pick a career aspiration, write about the job market and experience necessary to work in the industry, and conduct an interview with someone who has done relevant work before. I chose sports analytics as my field of choice, and I know you’ve done some statistics work with a professional sports team in the past.

Can you provide a general overview of the work you did with the team?
What was the workload like? How did it compare to the normal amount of work you need to do as a professor?
How did you begin working with the team in the first place? Did they contact you, or was it from your end?
What was the culture like from the organization’s side? Were they excessively demanding?
What advice do you have for someone interested in studying statistics?
What facet of experience is most valuable for someone looking for a job in sports analytics?
If someone wanted to work in the behind-the-scenes aspect of sports, is it best to start from within the world of the particular sport, or come in from outside?
How does the application of statistics in a social science/political context differ from a sporting context?

And some less serious questions:

Python or R, which is more useful to know?
What is your favorite metric for evaluating players?
Favorite athletes currently?
(My teacher is a big mets fan) Why are the Mets so bad historically? (be nasty please)

Before answering the above questions, let me emphasize that I only know about some small corner of sports analytics. That said, here are my responses:

1. I helped the team fit multilevel Bayesian models for various aspects of game play and player evaluation.

2. A couple of us met weekly for an hour and a half with some people from the team’s analytics group. They were the ones who did almost all the work. I think we were helpful because they could bounce ideas off us, and sometimes we had suggestions, and also we helped them build, test, and debug their code.

3. People from the team contacted me. They had read one of my books and found the methods there to be useful, and they wanted to go further.

4. We agreed ahead of time on a certain number of hours per week. On occasion we’d do a bunch of work outside the scheduled meetings, but that didn’t take up too much time, and, in any case, it was fun.

5. There are lots of ways to learn statistics. Ideally you can do it in the context of working on some application that is of interest to you.

6. I’m not sure what is the most valuable experience if you want to go into sports analytics. If I had to guess, I’d say programming with data, being able to manipulate data, make graphs, extract insights.

7. Some of sports analytics people I’ve met have a strong sports background; others have strong backgrounds in quantitative analysis and are interested in the sport they are working on. I think it would be hard to do this work if you had little to no interest in the sport.

8. There are some similarities between social-science statistics and sports statistics, also some difference. With sports we typically have a lot more data on individual perfomers. See this post, Minor-league Stats Predict Major-league Performance, Sarah Palin, and Some Differences Between Baseball and Politics.

9. I use R, which is popular in academic statistics, economics, and political science. In the business world, it’s my impression that Python is more popular. I fit my models in Stan, which can be called from R or Python.

10. I don’t have a favorite metric for evaluating players. If you fit an item-response model (see here, for example), then player abilities are parameters in the model and are estimated from data. So in you don’t have a metric (in the sense of some summary that is a combination of an individual player’s stats), you have a model that allows you to simultaneously estimate the relative abilities of all the players. It also makes sense to model players multidimensionality: in the general sense you can break skills into offense and defense, or subdivide them further (different sorts of rushing, receiving, or blocking skills in football; strikeouts, walks, home runs, and performance for balls in play in baseball; different sorts of matchups in basketball; etc.).

11. I haven’t been following sports too closely recently, so my favorite athlete depends on what I’ve been watching lately. Shohei is amazing, Mbappé and Messi gave us quite a show last summer, Simone Biles can do incredible things, ya gotta love Patrick Mahomes, . . . we could go on and on.

12. Hey, the Mets just won today. You gotta believe! Regarding their history, I recommend Jimmy Breslin’s classic book, Can’t Anybody Here Play This Game? Breslin also wrote a beautiful biography of Damon Runyon—a great read for anyone who’s ever lived in this city.