NHL Machine Learning Draft

A statistical approach to the NHL draft.

• Home
• Where Am I?
• Twitter

Info:

• This is output from a shot-based expected-goals (xG) model. The model is similar to the model explained here. Our model is simplified relative to the linked work, and some input data is estimated.
• The plot below shows the increase (or decrease) in odds of a player's team scoring a goal when they're on the ice (x axis) against the increase (or decrease) in odds of a player's team giving up a goal when they're on this ice (y axis).
• This is an even-strength model.