Y
Hacker News
new
|
ask
|
show
|
jobs
by
timy2shoes
720 days ago
Then take a Cauchy or a t-distribution. Basically anything with a longer tail than exp(x^2). The Gaussian summary will be misleading because of the tails.