"fat tails" - Refers to a bell curve that has more outlying events than a normal distribution.
Ex. Lets look at the probability distribution of the chances of my getting laid this weekend by an extremely hot or extremely ugly girl. Will say a girl that is a "5" is the mean.
first scenario: I am sober. I don't have the balls to talk to hot girls and I definetly not talking to ugly ones. Vey few low or high numbers(ie 2,3 or 8,9) going home with me this night.
=Standard distribution
second scenario: I am drunk as hell and saying just about anything to every girl in the bar. God only knows what I would go home with this night.
This distrubtion would have a lot of outlying events from the mean(1's and 10's)
"fat tails"(pardon the pun)
I think my friends would argue that the second scenario would actually result in more data points on one side of the mean.