Ask HN: Let's do an A/B Test on HN | Hacker News Reader

Ask HN: Let's do an A/B Test on HN

4 pointskashif6 comments

How about if HN allowed down votes for a week. Then we could compare the quality of stories, instead of just speculating.

View on HN

As jacquesm says,

+ http://news.ycombinator.com/item?id=1986986

+ http://jacquesmattheij.com/HackerNews%2C+now+with+downvotes

... flags are already counted as down votes. He does also say this (in his item)

  > The downside of this being an obscure feature is that
  > knowledge of it could be abused by the less scrupulous
  > to push undesired articles off the homepage using a
  > series of flags.
  >
  > Another way to abuse this is by downvoting everything
  > except for the articles that you upvote

PG has stated elsewhere that he has an "indiscriminant flagging" detector, and that when it gets triggered, subsequent flags from that person don't get counted.

Hn already allows downvotes.

On stories?

Yep.

http://jacquesmattheij.com/HackerNews%2C+now+with+downvotes

So why do comments have actual downvotes and not just flags?

I suspect this is historical. As I understand it, the original intent was that submissions would be ranked by their upvotes, and that only items that were inappropriate should be flagged. Then sufficient flagging resulted in deletion.

Comments, on the other hand, couldn't be deleted by flagging, but simply got ranked according to their votes.

PG has decided to count flags as downvotes. He continues to tinker to try to find a good balance of these limited controls. As the site grows so its focus is getting lost, and these changes are an attempt to keep things "on track."

That's my understanding. It's probably wrong in detail, but it feels right in general.

Reply on news.ycombinator.com