the Mandela Effect is proof of predictions of the past where the average model outputs were highly skewed from the "outlier" model outputs that had better data inputs