The thing is, if you can record an event, you can formulate a statistic for that event. That is, you can have a statistic that only looks at an assist where the defense was broken down. You could even taylor it to consider how broken down the defense was by looking at how out of position the opposing defense was.
Not worthless at all, merely incomplete. That's like saying on-base percentage is a worthless stat in baseball. No, it's one of the most important stats to excel in, but of course it does not tell the whole story.
How to calculate PER It's not as simple as "tallying up the box score" like some people are implying. There are some calculations included which do things like normalize against league-average performance or attempt to tease out a player's performance vs. his team's performance. That being said, there are some noted weaknesses of PER. It's not "highly correlated" with shot attempts (which implies that you can just jack up shots forever and get a relatively high PER), but a player's PER will go up with the number of shots he attempts to make. So if you're really curious why T-Jones has such a high PER, plug in the numbers and see what you get.