4 Comments
User's avatar
Andy's avatar

Does this only account for the games this seasons games or does it take into account previous seasons as well?

Expand full comment
T.J. Breshears's avatar

Great question! I should have pointed this out, but the data set starts in 2020 so it is cumulative. That being said, there is some built in rating regression/deflation adjustments that are made between seasons.

Expand full comment
Zephan Rodriguez's avatar

Please go over the math! I’m a huge math nerd lol and when I was in highschool I was unimpressed with the rankings put out by our high school league. (I was in Kansas city where they just started a league), I actually tried making my own elo based ranking system but I feel it was lacking.

Expand full comment
T.J. Breshears's avatar

Love meeting other math nerds! Here are some quick tidbits. I am using K=60 (after a lot of testing and some great suggestions from another reader). The home court advantage is currently +50 points (just for that match) to the home team. Travel is -5 points for every 250 miles travelled with a cap of -25 points per match. In between seasons, teams revert towards 1500 (the aimed for mean) by one tenth. There is some other stuff in there but those popped into my mind right away.

Expand full comment