Quick aside here, using the same pre-NBA data I used to build my three point shooting model I ran a model on free throw shooting. The response variable was the player's free throw shooting percentage over the first four years of their career as long as they had at least 80 attempts, which gave me just over 300 players to test.
The model converged to a .529 R2 with two straight forward variables, the player's free throw percentage and three point shooting rate (3pt att/ total field goal att). Once again we see the relationship between free throw shooting and shooting from distance; shooters shoot.
Here's the formula:
NBA Free Throw Percent = .292 + .613 * PreNBA Free Throw Pct + .067 * PreNBA Three Pt Shooting Rate
And here's the formula applied to some of the notable incoming rookies:
Nothing surprising there, the line doesn't move from college to the pros and the free throw "defense" is just the same.
The image above also has the latest iteration of my three point shooting projection model, that model uses the pace adjusted three point shots made and attempted in addition to free throw shooting percentage and three point rate.
Lastly, my initial attempts have indicated that true shooting percentage is also a tough number to predict. It also seems to have a stronger aging component than any of the pure shooting numbers, which is interesting and could possibly indicate that shot selection improves more reliably with experience rather than shooting stroke.
Comments