Heavy Tails - The Wider View
Nature, while often exhibiting normal behavior, also shows us many other cases where extreme values are persistent, meaningful and may dominate the probability calculation. Until very recently, the computation of probabilities for non-normal distributions has been difficult or impossible. This difficulty has been recently overcome and the study of outliers or conditions under which outliers play a significant role is now within reach.
The presence of outliers creates a pdf with "Heavy Tails" or "Fat Tails". This means that, because there are observations far from the mean and/or those observations are perhaps individually large, the tails DO NOT rapidly descend to the x-axis, rather they extend some distance from the expectation before approaching the x-axis. In the example below this condition occurs on the right side of the distribution. Note the longer, heavy right tail.
