Yeah, except that's not what they're talking about at all.
If you read the article, it's about predicting that it's likely that a crime will be committed in such a place in such a time window, and putting a couple of cops there to deter it. That's really no different from having police line the route of a street demo or a carnival.
Of course there are obvious problems with measuring the performance of this model - a bit like Lisa Simpson's tiger-repelling stone - but actually, it's easier than you'd think. All you have to do is deploy it in some areas of the city for a controlled period, and measure the overall reported crime rates in areas where it is used versus those where it isn't. If there's a statistically significant improvement, then great. If not - oh well, it was only a million bucks, pretty small potatoes really.