Ask HN: Why is the 40-comment penalty being applied to Show HNs?

[fivedogit]

First, I'm extremely appreciative of Hacker News, its moderators, and the forum it provides for poor builders like myself to get some eyeballs on our infant products at no cost via Show HN.

However, it seems the 40-comment penalty is being applied to Show HN posts, which seems counter-productive to me. Isn't the whole point of Show HN to generate healthy discussion and answer user questions? Shouldn't they be exempt?

(Yesterday I posted a Show HN that had reached #15 on the front page when the 40th comment came in. Ding! Welcome to page 3. Irony: Several comments were my own, replying to questions.)

Source: http://www.righto.com/2013/11/how-hacker-news-ranking-really-works.html

[md224]

Personally I think "generate healthy discussion" should be the goal of all HN threads, which is why I think comment-count penalties are unproductive. How do you tell the difference between "vigorous debate" and a "flame war"? If the algorithm utilized NLP to analyze the tenor of the discussion, that would be one thing, but upvotes vs. comment count is a troubling metric. I want to read threads where people are passionately discussing a topic with long back-and-forth debates. If that's a "flame war" then I'd like more flame wars, please.

[vbrendel]

According to that article, the 40 comment penalty is applied to all posts. Putting "Show HN" in your post probably didn't change that.

The algorithm for the front page favours newer things and attempts to generate traction for discussions. I'm speculating here but once something hits 40 comments, not only can it be deemed 'controversial' it could also just mean the article may be able to sustain itself better without more eyeballs, and the front page can 'move on' to other newer things.

Ultimately the Hackernews algorithm should prevent anyone using it as a cheap referral source. 40 comments should have given you some valuable feedback, which is what Hackernews is good for, primarily.

[minimaxir → vbrendel]

Note the important secondary criterion: "there are more comments than upvotes."

The OP's article had 22 votes and 55 comments.

[the_eradicator]

From the guidelines, "Please don't post on HN to ask or tell us something (e.g. to ask us questions about Y Combinator, or to ask or complain about moderation). If you want to say something to us, please send it to hn@ycombinator.com."

[joshdance → the_eradicator]

That is a narrow use case related to Y Combinator chatter. You can post to ask or tell something to the HN community.

[gravity13]

What is the 40 comment penalty?

[hkmurakami → gravity13]

It's the anti-flamewar detector, which kicks in when a thread has more than 40 comments, and there are more comments than upvotes.

[asdfologist → gravity13]

"A 'controversial' story gets severely penalized after hitting 40 comments."

http://www.righto.com/2013/11/how-hacker-news-ranking-really...

[ejr]

I've been lurking here for quite a while before I started posting and I strongly suspect it's the ratio of upvotes to comments that trigger this. There may be a "recalculation" at 40 comments for performance or other reasons, but I've seen plenty of other stories with hundreds of comments and an equal or greater number of upvotes - on occasion passing 1000 - that stay on the front page for the better part of a day.

[qzervaas]

I had this happen with my own submission yesterday[1], which wasn't a Show HN. Was sitting at #5 for a few hours, then suddenly dropped to page 2.

It was all healthy discussion, not a flame war. Actually would've been better served by me not responding so quickly.

Currently at 52 comments, 43 points.

[1] https://news.ycombinator.com/item?id=8094684

[fivedogit → qzervaas]

Same here. I wish I'd just not replied.

[keyle]

Interesting, I didn't know about that penalty.

Should it factor the overall down-voting of comments as part of the equation?

[minimaxir → keyle]

Flame wars don't necessarily mean low-quality comments, so tracking downvotes wouldn't be helpful.

Unless perhaps you incorporated a normalized measure of # downvotes vs. # upvotes in a thread.

[aptwebapps]

Just read the part of the linked article relevant to the 40 comment penalty. My question is: assuming that this algorithm works in general, why would it need to be different for Show HN? What makes them different with respect to vote to comment ratios?

[minimaxir → aptwebapps]

In theory, Show HNs should drive more discussion.

Actually, the average amount of comments for each Show HN submission has always been below 10 comments: http://minimaxir.com/img/show-hn/show-hn-comments.png

[aptwebapps → minimaxir]

Because they're actively soliciting comments, I guess? That makes sense, but for me the main point of HN is the discussions anyway.

[plaxis]

Because you mustn't compete with the incubator itself.

[wheaties → plaxis]

Actually YC companies get hit with this just as often as the rest.

[scovetta → plaxis]

I surely hope this isn't true.