The point values were based on the commonality of the letters in common use, the idea being that it's harder to come up with appropriate words with letters like Z, X, Q in them. The letters' commonalities may be different for the word set of serious play-- the Scrabble dictionary-- in which case the scores become strategic features only loosely connected to historic frequencies.
I'm surprised that J is so uncommon, though. I guess it seems much more common than X and Z because there are a lot of words beginning in it, and it's also at the front of a very large number of common first names.