Just to finish your thought: Because of this, it’s possible that “A” and “the opposite of A” actually can both raise your estimate of p(B), even though “A” and “not A” can’t, as BlueSun stated.
Just to finish your thought: Because of this, it’s possible that “A” and “the opposite of A” actually can both raise your estimate of p(B), even though “A” and “not A” can’t, as BlueSun stated.