[Almost done, still need to interpret results better] It appears the results aren't very significant in predicting next years performance for the new metric but it is far more predictable ahead of time

Strength of schedule is a fundamentally flawed metric. Losing to a team shouldn't give you an easier strength of schedule. Let's remove those from the calculation and see if this produces better results.

If this is your first tutorial you've used please start with installing miniconda and cloning the repo.
Install miniconda https://docs.conda.io/en/latest/miniconda.html (choose latest python version under your OS. Likely 64 bit)

git clone https://github.com/rogerfitz/tutorials
cd tutorials

The exact python version doesn't matter because with each project I'll have you create a different environment with the proper version of python.

From the tutorials directory

git pull origin master
cd adjusted_strength_of_schedule
conda env create -f environment.yml

From the tutorials directory

cd adjusted_strength_of_schedule
conda activate adjusted_strength_of_schedule
jupyter notebook