players who play less minutes have anything they do exaggerated beyond meaning in a per48 stat, and sheer luck can result in crazy numbers.
This is just a re-hash of the sample size point. We agree on this, I think.
Also, playing limited minutes will tend to push performance to one end of the spectrum or the other, further inflating or deflating the per48 stat.
Interesting theory. What evidence is there for this?
That is why the NBA on their stats pages have the required minutes played and other limiters to see who really is performing and who is showing inflated numbers due to limited minutes.
Another re-hash of limited sample size.