“Underdog All-Stars: Constructing Championship Teams with NBA’s Overlooked Talent”
The Results
In order to judge whose team would perform best, I created a model of offensive and defensive NBA TRACR based on historical data since 2012. The model used preseason offensive and defensive DRIP ratings for all players who played for each team in each season, along with their minutes played. Once that model was created, I used it to predict TRACR ratings for everyoneâs selected teams. Everyone had the option of sending me what their ideal rotations would be for a competitive game with a fully healthy roster. I then adjusted those minutes projections to account for rotations in blowouts (so each teamâs 14th/15th guys donât end up with zero minutes played) and player health based on their availability this season. Once I got their projected TRACR ratings, I assumed league average pace for all teams (perhaps questionable given Coach Daleâs four passes before any shot rule), and projected their average points scored and allowed per game. I then used a formula for Pythagorean expected wins to project each teamâs win total.
Post Comment