He doesn't play 36 minutes, therefore it is a stat of nothing. Per 36 isn't a stat really it is a projection!
Here is a neat fact on 36 PER. Only 26 guys in the NBA play 36 minutes that is probably about what 10% of the league.
http://espn.go.com/nba/statistics/player/_/stat/minutes/sort/avgMinutes
There are all kinds of reasons this stat is useless besides it doesn't happen with regularity. Guys get tired, guys foul out and guys are not good enough to stay on the floor long enough to get to the 36. It's a projection, not a stat.
It is a ratio. That has value.
No stat ever gives you a full picture. But look at someone like Harden. He was great as a 6th man, but projected out to starter minutes, his numbers might keep pace. Guys who play 40 minutes and are still taking over games (indicating they still have the legs) will have lower stats if they are given less minutes due to a top reserve that the team wants to keep involved, or if the team wants to reduce wear and tear for the playoffs.
Per minute stats are also better than per game when evaluating particular skills that a player may be good at since the reason that a great shot block gets few minutes may be that he is bad at everything else. If he got 30 min per game instead of 10, he might easily get 3x as many blocks, but the team would lose due to playing a bad player who has only 1 skill.
LLikewise, if Sully gets fewer minutes due to defensive lapses (not due to fatigue), per minute rebounds and rebound rates make sense. If he gets fewer minutes because of poor conditioning, than per minute stats could be misleading.
I am not worried about Sully's rebounding stats no matter what they are because it is well know that Sully is a good rebounder. What big ever rebounds well statistically in our system?