I live in LA, and I can tell you it is huge. Huge in terms of area and huge in terms of population density. Follow any of the police scanner-reporting or neighb

I live in LA, and I can tell you it is huge. Huge in terms of area and huge in terms of population density. Follow any of the police scanner-reporting or neighborhood Twitter news feeds and you can count the crimes as they happen on a minute by minute basis. It's not hard to find crime in LA, and even after being exposed to these feeds I can tell you where to find something going down. But the most puzzling piece of this is how do you measure a reduction in crime? A criminal act that didn't happen isn't a crime and can't be counted as a "missed opportunity." They mention in the article that they have witnessed "double-digit drops in burglaries and other property crimes." Nine burglaries this month compared to ten last month is a double-digit drop, so can you really attribute one "missed" crime as a success? Or how about a reduction of 2 out of 20? (This is a valid example as property crimes in various police districts here rarely go into triple digits monthly, or even on a six-month basis, look for yourself [1]). Why would you assume that 1. the number of criminal acts is near-constant and 2. that your system was the actual reason for a decline? To me this feels like trying to predict roulette wins by looking at the historical numbers - it's pointless. If you can't measure your success with solid numbers, then it sounds like the only thing this system does is compile and report crime info faster than previous systems. And then someone got the bright idea to send officers to "hot zones" where they have been told crime will happen. Something happens - they were looking extra hard for crime and they found it - win, or nothing happened and they called it a "prevention" - win. [1] http://projects.latimes.com/mapping-la/crime/ http://projects.latimes.com/mapping-la/crime/