My point of contention isn't whether or not we should police high crime neighborhoods more or less, it's whether we should condone the (less-common) explicit or (more-common) implicit police policy of using race as a proxy for the other variables that I mentioned previously. You've danced around my question: is race a valid "indicator" of "problem areas" in your model?
If the statistics agree with it, then yes, of course race can be used as an indicator of problem areas. But as I said in my second paragraph, there are certain things that invalidate that completely.
Of course, the entire discussion we're having is invalid if we work with the premise that more policing = more convictions, by virtue of the crime being everywhere (across boundaries in whatever criteria you wish, e.g. race) regardless, and convictions being a simple byproduct of policing and not of said underlying factor like race, obesity, education, socioeconomic status, or hair-color.