Postdiction

According to critics of paranormal beliefs, postdiction (or post-shadowing, retroactive clairvoyance, or prediction after the fact) is an effect of hindsight bias that explains claimed predictions of significant events, such as plane crashes and natural disasters. In religious contexts it is frequently referred to by the Latin term vaticinium ex eventu, or foretelling after the event. Through this term, critics postulate that many biblical prophecies (and similar prophecies in other religions) appearing to have come true may have been written after the events supposedly predicted, or that their text or interpretation may have been modified after the event to fit the facts as they occurred.

Skeptics of premonition use these terms in response to claims made by psychics, astrologers and other paranormalists to have predicted an event, when the original prediction was vague, catch-all, or otherwise non-obvious.

Most predictions from such figures as Nostradamus and James Van Praagh are written with such seemingly deliberate vagueness and ambiguity as to make interpretation nearly impossible before the event, rendering them useless as predictive tools. After the event has occurred, however, details are shoehorned into the prediction by the psychics or their supporters using selective thinking — emphasize the "hits", ignore the "misses" — in order to lend credence to the prophecy and give the impression of an accurate "prediction". Inaccurate predictions are omitted.

Supporters sometimes contend that the problem lies not with the wording of the prediction, but with the interpretation — an argument sometimes used by supporters of religious texts. This argument may lead to the question: "What is the point of a prediction that cannot be interpreted correctly before the event?" However, the argument is not that the prediction could not have been interpreted correctly prior to the event, but simply that it was not in the case in question, thus the question is working from a false premise. Of course, any "prediction" that is so vague as to not be correctly interpreted before the event it allegedly "predicted" is functionally equivalent to no prediction at all.