I'm not sure this is the right question. Even if torture worked 100% of the time, and by "worked" I mean:
* provided the intelligence quickly
* was 100% accurate
* and significant
Even then, it's not fucking worth it. There's simply no possible intelligence that can outweigh the disadvantages. We stop some roadside IED maybe... and what do we pay for that? We lose the moral high ground, we teach our enemies (not even sure that word is right) that they have nothing to lose by being suicidal in their attacks, that there's nothing to lose in torturing Americans that they themselves capture, and we create 10 more "terrorists" in all his cousins and brothers that arguably have an excuse for what they do.
That's not a sustainable model of terrorism reduction.
And that's supposing that the intelligence is good. So there's no point in asking whether it works or not.
* provided the intelligence quickly
* was 100% accurate
* and significant
Even then, it's not fucking worth it. There's simply no possible intelligence that can outweigh the disadvantages. We stop some roadside IED maybe... and what do we pay for that? We lose the moral high ground, we teach our enemies (not even sure that word is right) that they have nothing to lose by being suicidal in their attacks, that there's nothing to lose in torturing Americans that they themselves capture, and we create 10 more "terrorists" in all his cousins and brothers that arguably have an excuse for what they do.
That's not a sustainable model of terrorism reduction.
And that's supposing that the intelligence is good. So there's no point in asking whether it works or not.