Hacker News Reader

It's just a blog post. Regular economist articles are better edited, but mainstream economics doesn't make the HN front page.

>25 standard deviations has to be an exaggeration? I fail to see how any model could ever be that far off the mark.

It's easy to be that far off the mark. E.g. assume a normal distribution for a fat-tailed distribution.

It's 25 standard deviations because they only take like a year's data into account. So anything that happens and seems out of the ordinary for that particle date range will show up as 25 standard deviations. If you've read about the LTCM crash, they keep citing absurd standard deviations. Basically they didn't have any historical datasets that they were basing their models off of. It's completey ridiculous that they talk about such absurd numbers.

25 SD is exaggeration only for people who believe everything is normal distribution. I will suggest reading Nassim Nicholas Taleb's work specially The Black Swan and Fooled by Randomness.