Breakout Groups: Who's Ready to Take the Leap?
The NBA season is almost upon us. So let's use some basic machine learning to predict which players are next up.
The NBA season is oh-so-close after an offseason of megatrades, Wembanmania, and more. With all of the attention around the big dogs of the league (the Dames, the Jrues, and the Embiids), I feel like we haven’t talked enough about the guys who are ready to step up into Big Dog™ status — but there are actually quite a few candidates to do so, so let’s talk about them.
However, because I don’t want to solely rely on my own opinion here, we’re going to mix a little objectivity in. Here’s how it’s going to work:
We’ll collect data from Basketball Reference on players’ stats over the past 10 seasons, including All-Star selections and advanced metrics
We’ll create a random forest model, which is a common machine learning model that combines multiple outputs to make decisions. I won’t bog you down with the specifics of it, but it usually proves to be relatively accurate in static situations.
We’ll add a couple of different criteria to the model, such as being under 25 years old, never making an All-Star team, minimum usage rate, etc.
The test model is supposedly 86% accurate, which is better than 99% of sports journalists’ prediction rates — if you know a journalist who does better, take them to Vegas. We get quite a few players here, but I’m only going to focus on three. First, let’s begin with a guy who is the youngest triple-double threat since Luka…
