MaJiG:
I'm not sure that this thread was very kind or welcoming to the newbie that initiated it. Have we been as hospitable as we could to someone coming here asking for help?Just sayin'...
Well, he never actually asked for any help. Aside from that, I think the information here is of far more actual help to this individual than any technical answers we could have provided to questions he meant to ask. Lie detection is impossible. He needs to know that.
It is possible to convince someone that you can detect a lie, then measure their physiological fear responses to the threat of being caught in that lie. The problem with that concept is that the machine has no idea if the subject is afraid of being caught, afraid of the person administering the test, afraid of saying the wrong thing and being accused of something they didn't do, or just afraid of the big hairy spider they saw crawl across the back wall. The flip side is that if you test someone who is perfectly aware that the "lie detector" can never actually detect a lie, the test will not work because they will not be afraid.