as they depend entirely of what the rest of the team does

d-rating got people claiming Carlos Boozer is a good defender (and before any Duke fan goes there, I like Boozer and generally argue in his favor)

what we would really need is access to stats teams actually use, counting positive plays that do not make it to the box score (screens, effective help defense, near blocks, etc...), rather than a bunch of composite stats derived from points scored

*edit, I see you got that covered below. still curious as to why o-rating is not as biased as d-rating in your mind. *