You know, not sensing people's emotions from their faces is seeming less like a handicap and more like I am immune to illusions every time I hear about the biases most people have. Then I realize I don't expect people to be feeling something unless they tell me anyways, and that's potentially just as inaccurate. Then I realize I might not be compensating exactly enough, and worry recursively.

I can't be alone in thinking this; where does one acquire the probability values to make any of this theory useful? Past data? Can it be used on an individual or only organizational level?