Comment on To Paul GrahamparentComments−aswanson16yThey probably wouldn't change much if the commenter had a long history to be averaged in. You'd want to look at the delta of the delta (2nd order stats) to see if any noticeable effect was occurring.−ErrantX16yIIRC averages are calculated over a set of your most recent posts - so it can go up or down quite quickly depending on your current comment quality.−aswanson16yI wasn't aware it was a sliding window. In that case the method Paul mentioned is probably a good enough solution.
Comments
They probably wouldn't change much if the commenter had a long history to be averaged in. You'd want to look at the delta of the delta (2nd order stats) to see if any noticeable effect was occurring.
IIRC averages are calculated over a set of your most recent posts - so it can go up or down quite quickly depending on your current comment quality.
I wasn't aware it was a sliding window. In that case the method Paul mentioned is probably a good enough solution.