One of the problems I see with defensive ratings is that a player is not given credit for fouls.
Not "given credit for fouls" you say. Here is what I mean. According to my research on real statistics, a foul represents about one tenth of a failure that is every 10 fouls leads to 2 points. How can that be you say; at least 1/2 of all fouls lead to FTA so 10 fouls should lead to at least (foul/2*2FTA*.75 FT%) = 7.5 points. This is my point exactly. Clearly a team that fouls 10 times is giving up at least 7.5 points ON THE PLAYS WHERE THE TEAM FOULS. This means that teams that foul a lot give up fewer points on the plays that they do not foul. 5.5 points less, somewhere.
Now, of course, the point of any such analysis is to examine the results at the team level and apply them to the player level, and so you should probably conclude that a player who fouls more is playing more defense in general. To put it another way, if you have a team with a lot of fouls, it should be a team that gives up a lower shooting percentage, in general, because, in general, that is what happens in real life. But, it is still a team that fouls a lot.
Now I know that someone is going to say I've seen (player x) play defense and (player x) fouls a lot and is a crappy defender. Okay. It's about the same as saying that (player y) blocks a lot of shots but (player y) is a crappy defender. In general, however, fouls indicate an ability to lower an opposing team's FG%, at least on a team level, and so it probably makes sense to apply this finding to individual players. Now if a player's defensive rating influences the chance of a player getting a FTA or a FGM, their defensive rating should go down by just a little for each foul, but if a player's defensive rating influences only the chance of a player getting a FGM, and fouls influences only fouls, then a player's defensive rating should actually go up if they foul a lot.