Pontians, a significant part of Hellenism which was heavily persecuted by Turks, are for some unknown reason to me mocked in Greek jokes. The following “joke” is representative: a Pontic Greek scientist (typically named Giorikas, Panikas) was making experiments with a frog. It cuts successively one leg after the other and asks the frog to jump. The frog at the end obviously cannot jump but the conclusion of the Pontian scientist is that after cutting four legs the frog becomes deaf. The reason why I have this “joke” here is because that came to my mind after reading about the famous parapsychologist Joseph Banks Rhine. This post is about the Rhine Paradox. Rhine was performing experiments to test whether there exists the so-called extrasensory perception. Something like the sixth-sense. In one of his experiment he found that a student was performing exceptionally well. The experiment went like this: there were 10 cards of two colors and each subject had to guess the color of the card. The exceptional subject guessed everything correctly. However, Rhine obviously biased towards the conclusion he wanted to draw continued performing experiments with the same exceptional subject who was informed of his special powers but whose performance was declining, coming to normal levels, i.e., random guesses. However, -roughly- Rhine drew the conclusion that the probability of finding correctly the color of all 10 cards is 0 and that the exceptional subject got later distracted, unfocused etc. His final conclusion was the following: you should not tell people they have extrasensory perception. It causes them to lose it. Clearly, this is not the way to do serious research, of any kind. Martin Gardner, a famous math hero, got angry with Rhine as you can read in Wikipedia!
This brings us to Bonferroni’s principle (I was aware of his inequalities, not his principle though :)) which I learned from this book: roughly, consider a finding important if it deviates significantly from what you would expect to see if things were random.
 Mining of Massive Datasets, Rajaraman & Ullman