I once worked with a doctor of statistics who tracked every pull of the California Lottery since its inception. He was modeling future numbers based on past results. This was his logic: If a number, say 17, was not coming up, that was a number he would play because he said the law of averages dictated that eventually it would come up. I pointed out to this doctor of statistics that there were two plausible scenarios, neither of which supported his logic:
- The pulls are truly random so there's no point to try to predict them based on past results.
- The pulls are not random, and if there was a physical property that made a ball come up more frequently, then past results might indicate that the numbers that came up most frequently would continue to do so, and they are the numbers that should be picked.
He countered by maintaining that, in order to maintain random draws, if the second scenario was true the lottery commission would remove the balls that came up most frequently and replace them with other balls that were less likely to come up. My response was that if anyone was purposely juicing the odds of which balls came up most frequently it would probably result in jail time.
Sometimes experts in a field have no common sense, even in their field of expertise.