Hacker News Reader

> saying that the "tails" of market changes are going to be larger than the current models

The claim is that people using the models are well aware that the gaussian distribution is only an approximation, and that substituting fatter tailed distributions into the models don't actually affect the outputs very much.

View on HN

Yes, I know that's been claimed.

Do you think that, perhaps, the massive collapse of the financial system in the past year might at least cast a bit of doubt on whether that "awareness" translates effectively to behavior?

This is a not a very solid claim. The difference in calculations using a distribution with a finite variance versus one with an infinite variance could not be more different.

The approximation is not done for closeness, because they are not close at all, but for convenience and simplicity. Probability theory is amazingly difficult when you are dealing with infinite variance.

Reply on news.ycombinator.com