Careful down this path. What you're saying is that law enforcement re: terrorism should not have to respect the same principles and standards we place on other law enforcement action.

Which is the precise argument that the other side is making. When we special-case terrorism we're no longer on solid ground - whether it's in favor of looser or stricter treatment.

So, here's the thing: the level of false positives, when calculated in aggregate can suggest unreasonable enforcement. If 95% of all searches yield nothing, there is a legitimate question of whether or not your targeting is reasonable. This is the primary legal argument being used, for example, in the ACLU's case against NYC's stop and frisk policy.

But a single case of a false positive, with no further information, cannot be taken as evidence either way. The fact that the Wall Street plot may have been a false positive sheds no light on whether or not the suspicion was reasonable in the first place.

Supposedly there have been 10 "plots" in the last 12 years that have been disrupted by US government surveillance of its citizens. That's much higher than 95% yielding nothing. This is a fact, something we already know.

I think you're making a category with respect to traffic stops, though. It's not "law enforcement" at the point the searches are being done in the context of surveillance, and I'd argue that meatspace comparisons here are inapt on the whole. The level of inquiry through surveillance would be comparable to a LEO having reasonable suspicion of something, then downloading your GPS history and searching your car after you leave it at your destination (imperfect, but close enough for internet debate).