Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

But you often want a number that’s more correlated with the average. Imagine a distribution that has 80% very low values and 20% insanely high values. The median won’t capture changes in that 20% region at all. So you could see numbers not changing when in fact they may have very dramatic changes


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: