The sample size I used is his entire 4 year injury riddled career.
It’s literally impossible for me to use a bigger sample size since one doesn’t exist.
Sure, but the third year in your sample is meaningless. The first 2 years are valid for comparison.
But I don't think his shot looks any better.