It is often mistaken in the mind that dependent events are independent. Take life expectancy. Every birthday I congratulate a person for beating the curve and increasing their life expectancy.
That is, p(live longer than life expectancy) != p(live longer than life expectancy | already lived 25 years old).... this is sort of obvious. As you get older, people your age die... they pull down the average that you are trying to fit into your situation.
Similarly, if you are looking at rotations in a given market and have a 3rd standard deviation move. The probability you have another 3rd standard deviation move soon is higher than at a randomly selected time in a market. These events are not independent.
This is often where the saying "the market can be irrational longer than you can stay liquid". If you trade an edge case using the distribution for the norm, you are going to have a bad day.