A metallurgical scientist’s approach to predicting NBA team success
Peter Tsai

It is a good read. I like your idea that apply simple feature engineering and simple ML method to deliverer promising prediction results. One thought I came up with is using hierarchical structure by segmenting different shooting zones (left field, right field, 3pt, and paint) instead of using all features in a single vector. Then you may can generate models for different zones and then ensembles them into a final prediction model.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.