The predictability ranking is a great angle. Did you check whether it correlates with metrics related to offensive success? Curious if being ‘unpredictable’ actually helps teams score more, or if it’s basically unrelated to how good the offense is.
This would be a good next step for the project, there does not seem to be a correlation between the predictability of my model and teams' points per game. My project did not use data relating to the offensive playcallers and their tendencies, so introducing that data may control better for both predictability and offensive success.
The predictability ranking is a great angle. Did you check whether it correlates with metrics related to offensive success? Curious if being ‘unpredictable’ actually helps teams score more, or if it’s basically unrelated to how good the offense is.
This would be a good next step for the project, there does not seem to be a correlation between the predictability of my model and teams' points per game. My project did not use data relating to the offensive playcallers and their tendencies, so introducing that data may control better for both predictability and offensive success.