zlacker

I'm curious about why this post appears to be on it way to being flagged off of the home page. It's valuable information and is likely very relevant to the HN community. Thanks for the heads up @winston_smith.

>>throwb+(OP)
It is off the homepage again.

>>throwb+(OP)
Since I can't delete a post I've removed the original content after seeing dang's message. Appreciate your reply dang :)

>>throwb+(OP)
It set off the flamewar detector. We review submissions that are affected by that software, and turn it off for threads that aren't flamewars. I've done so for this one. Other than that, moderators didn't touch this post or (as far as I know) even see it.

>>PudgeP+W4
We don't moderate HN that way. It's just the opposite when a YC company is involved:

https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

>>dang+V5
Hey that's neat, I didn't know HN had a flamewar detector. How does it work? Content matching and post frequency?

>>zkim+G7
Might be depth of comment branches and ratio of upvoting and downvoting to comments?

>>zkim+G7
I know at some point there was a penalty attached to articles that had more comments than upvotes.

This struck me as incredibly unnatural, since I frequently comment and very rarely upvote an article. I don't really see what the one metric has to do with the other.

But apparently everyone else has a different model of HN in mind.

>>thauma+wg
I think there's a lot of variance in the way people use HN (I don't upvote stories much either, though plenty must, given the vote counts we see), but dang has had plenty of time (i.e., nearly 8 years in the job) to observe what combination of upvotes, comments and other behaviour may indicate a possible flamewar.

>>thauma+wg
I’m sure some heuristic that combines comments per hour, average thread length, and ratio of comments to upvotes could do a pretty good job detecting flame wars in a community the size of HN.