In his book, "Thinking, Fast and Slow," cognitive psychologist Daniel Kahneman tells a story that launched him into 30 years of research that would eventually earn him the Nobel Prize. The story represents a very common flaw in our thinking; it goes something like this:
Kahneman taught flight instructors in the Israeli Air Force. He was telling them that punishment for mistakes was not helpful. An instructor adamantly disagreed and said that in his vast experience yelling at the pilots for mistakes did lead to improvement. So what was going on?
The instructor was not lying—following some verbal bashing and good old fashioned public humiliation, the pilots did do better, but it wasn't the punishment that changed the behavior; in fact, the behavior didn't really even change—it just regressed to the mean.
Regression to the mean describes the phenomenon where relatively stable systems over time tend to have temporary peaks and valleys from the stable norm (the mean). In the story above, the pilots all had countless hours of training but had off days (valleys). These off days were inevitably followed by the return to their solid flying performance, or regression to the mean. Likewise, the pilots who were praised for exceptional performance (peaks) usually did worse the follow day—not because of the praise, but because of the regression to the mean.
Think about how this applies to your own life. We all have good and bad days, and eventually things get back to normal—with or without the lucky rabbit's foot. We all get minor illnesses and eventually get better—with or without the special "all natural" medicine. It eventually will rain again—with or without the human sacrifice to the gods. We are generally stable creatures that regress to the mean—with or without attributing any cause to the regression.
The next time you regress to your mean don't be so quick to issue blame or credit—unless you are blaming or giving credit to statistical inevitability.