Can't disagree with any of what you write, nate, though it doesn't affect my take that I wouldn't swap Bryant for Turner straight up at this point. (& of course, especially, I agree that I would have paid a premium to lock him up for 4 years rather than 3)
One thing worth mentioning about defense -- or, rather, about the relationship of defense & offense in the NBA: there are "offensive" stats which are better understood as having a defensive aspect as well.
Presumably, overall, team defense is judged by the % of the opponent's possessions which turn into points (w/, obviously, an adjustment for possessions that become 3 - or sometimes even 4 - points). But, this leaves out 2 elements which, to me, seem quite relevant. The first is that effective rebounding limits the
number of possessions your opponent gets -- yet, how are rebounds calculated into effective "defense." Even a rebound on offense has the effect of preventing the opponent from initiating
their offense.
The second (& related) point is that the more effective your team is on offense, the more occasions on which the opponent initiates
their offense by taking the ball out under the basket. W/o checking, I postulate that teams score with much lower efficiency in those situations than via a defensive rebound or a steal -- & that lower efficiency improves your team's defensive rating. Yet, it was created via efficient offense.
And, obviously, a guy w/ a 68.7% TS% is doing an outstanding job of creating that situation for the opponent. Not to mention that your team's offensive efficiency may force your opponent into a different offense from the one they prefer, because they are playing from behind.
I have no numbers to support any of the above, but it does seem to make sense. Or... does it?
One more point: since basketball is a zero sum game, defensive effectiveness is only really measurable with offense in mind. The ultimate measure is how many points are you up or down to the league on the season. If a team that measures well using metrics that isolate defense from offense is nonetheless ineffective in winning games, then there's room to doubt the significance of those defensive metrics.
Moving that notion from "team" to "individual" would allow someone to argue for Bryant over Turner based on box score numbers -- since those numbers translate 100% into wins & losses.
The intellectual effort to build any/all of the above into a model w/ predictive power, using "real" (i.e. quantitative/statistical) tools, is beyond my capacity/training/interest (despite what people here may think, mostly I just like to watch the game!!
), but I bet it's been worked on.