. . "On a question such as feedback, until it is proven that feedback is positive (or negative) the models cannot make any claim of accuracy if they assume it is positive." . .