Hacker News Reader

Averages are normalized. Convolutions only are if they integrate to 1.

What are you using non normalized weighted averages for?

I'm not! Averages should normalize the weights. But there are plenty of reasons to define convolutions where neither signal is normalized.

What are they?