About/Best Of
Mathletix is a blog about statistics, sports, machine learning, AI and behavioral economics, written by me, Casey Durfee. You can contact me at csdurfee@gmail.com. Unless explicitly mentioned, all content is the product of natural stupidity, not AI.
All code is available at https://github.com/csdurfee or https://www.kaggle.com/code/caseydurfee
Someone else used to have this domain and associated twitter account many years ago. I have no connection to them. Apparently they were a tout, someone who sold wagering tips for money, and they weren't very good at it. I should have done my due diligence, but I'm sticking with the domain for now due to the cosmic irony, since a lot of the writing on here is about why gambling on sports is a bad idea.
Here are some of the blog's highlights:
The hot hand in the NBA
Read the full series: https://mathletix.net/tag/the-hot-hand.html
The "hot hand" doesn't exist for field goal attempts in NBA basketball. In fact, NBA players are significantly less streaky than they should be. This unstreakiness is mostly due to shot selection. For NBA players, their shooting percentage by type of shot doesn't change much based on how well they've been shooting lately -- there is definitely no hot hand in that sense. A player who has made 4 or 5 out of the last 5 shots isn't significantly more likely than average to make the next one, controlling for shot difficulty.
While FG% by type of shot doesn't change, shot selection does -- on average, NBA players take higher percentage shots when they've been shooting poorly recently (they go get a bucket), and lower percentage shots when they're shooting well (they take a heat check). These heat check shots tend to be more inefficient shots as well (lower expected value), so this behavior definitely costs teams points, and possibly wins.
The real world is far streakier than our brains are equipped to deal with. Players believe they are streaky, which causes them to be less streaky than they would be by chance. Even though they're not as streaky as they should be, NBA players still seem streaky to people who watch basketball, myself included. Our brains attach far too much meaning to streaks, and are really bad at intuitively understanding randomness.
Your parlay sucks
Full series: https://mathletix.net/tag/your-parlay-sucks.html
A long investigation into sports gambling and how it works, mathematically and psychologically. So far, I've published pieces about betting spreads, line movements, which teams get bet on the most, and how random walks work.
"Some Educational Value"
Full series: https://mathletix.net/tag/some-educational-value.html
Various pieces about statistics and machine learning including how normal approximations work, pitfalls of ensemble learning, and harmonics in music.