Statistics, in what concerns data processing, is the realm of inductive reasoning. I use to compare the situation with the sec. XIX medicine one when diagnose tests was almost unknown. In contrast the deductive reasoning we has to be prompt to the scene: Throughout the window I see a lot the people who carries unfold umbrella, even that I cannot see the rain drops falling I deduce that is raining, but, on contrary, if umbrellas was folded I conclude that it was raining, not at this instant. This kind to get conclusions is very different than those provided by deduction, and on contrary to these, they risk being wrong. Basing only at *symptoms* the diagnosis could fail. Deductive reasoning never leads to wrong conclusions.