I think were missing some key pieces of data to evaluate the claims made in this article; and I think the authors of this article know that and are actively trying to avoid looking at that data.

A fairly obvious example, imo - if police know their local prosecutor wont prosecute certain crimes (theft, property damage, drunk and disorderly), police tend to stop trying to arrest people for those crimes, and the overall rate for those crimes will go down (in the data at least). The victims on the other hand, tend to get little comfort from the nice looking excel charts.

Ive lived in enough west coast cities over the last decade to have a pretty solid understanding of how this particular game gets played, but I'd love to be proven wrong; Lets just see all the data eh?