It's the weekend, time for something non-techie.
I am convinced that we have a general tendency to over-estimate the value of rational explanations of past data. Consequently we over-estimate the accuracy our predictions based on those explanations. In his talk at Pop!Tech 2005 entitled The Scandal of Prediction, Nassim Nicholas Taleb makes this point very eloquently.
He calls this tendency epistemic arrogance and points to an experiment where he gives the audience a question, say: what is the average age of the U.S. population? He then asks the audience not answer the question, but to give a range where they fill confident at 98% that the correct answer falls into. The goal of the experiment is not to measure what we know, but the difference between what we know and what we think we know. As people can give a range as large as they want, out of 100 people no more than 2 should come up with a range that does not include the correct answer. In reality, on average around 40% give a range that does not include the correct answer.
The title of this entry is a quote from Yogi Berra, the painting, The Fortune Teller, is from Caravaggio, and both are mentioned by Nassim Nicholas Taleb in his talk. Note in the painting that while the girl tells the boy his good fortune, she is also gently removing his ring.