I emailed moneypuck in 2016 to ask for more details on how his playoff odds model works. At least at that time, it was a linear regression of the published variables, trained on decade of past games. Without getting into the technical reasons why, I can summarize by saying that's not a good way to predict playoff performance (source: myself, a data scientist). I have not been back to the site since.
edit: removed a typo
|