Comment by alterego

I think this article isn't talking about random vs. profiled (which is what the mit paper is about), but random vs. 100% (where there 100% has been analyzed carefully to be sufficient to detect large bombs, etc.).