[savagedata]

This is my favorite statistics factoid! Regression/reversion to the mean is the idea that if you observe an extreme value and remeasure it, it will tend toward the average value on the second observation. For example, students who have a very high or very low score on a test will have a more average score if you retest them.

Sir Francis Galton demonstrated the idea of regression to the mean by inventing linear "regression." He plotted the heights of children vs the heights of their parents and showed that very tall parents have children who are not as tall.

The fact that we call linear regression "regression" has nothing to do with reverting to the mean other than, coincidentally, linear regression was first used to illustrate the concept.