[dekhn]

My "insubstantial" one-line comments are often more valuable to readers of HN than pagelong of pseudointellectual debate (I watch my upvotes carefully. Which reminds me: is there a tool to track votes and comments on my comments?)

[pvg]

Lame things get upvoted all the time so watching your upvotes is not really a measure of 'value'. Adding more lame things to the site just because it already has lame things only makes it lamer, not better.

[dekhn]

I've spent the last several years treating Hacker News as a machine learning source and have learned to understand upvotes (and downvotes).

If you word things in a confrontationally, overly partisan, or ignorant way, you'll get downvoted, even by people who technically agree with you. So, go out of your way to write more neutrally to reduce downvotes on tone.

If you propose an unpopular idea, or criticize one of HN's Holy Cows, you'll get a bunch of downvotes from people who disagree with you. Occasionally, I don't comment at all if I think my opinion is out of the mainstream and not justifiable.

Some folks just downvote people who state the truth because they don't like negative naysayers. Make it clear, if you're stating How the World Is, that you don't necessarily agree with it, but it's a structural problem.

Other folks don't like simple solutionism so avoid saying "It's easy. We can solve world hunger by <blahblahblah> idea" which will never work, because food production isn't the reason people are starving.

Another thing that can get you downvotes is invoking Expert Privilege. For example, HN will downvote you if you don't provide some sort of pseudo-rational sounding argument and instead just say you're an expert. OK, fine, ignore the fact that my PhD taught me how to read science papers and press releases.

Finally, and this is the most interesting thing, votes come in waves. I'll often get -2 on a comment right after I post and then it will trend upward for a whole day. I suppose my comments age well.

Ultimately, maxxing my karma is generally correlated roughly with making good contributions to the site, and I've calibrated well enough to interpret downvotes.

[pvg]

Be that as it may, you should still not post lame one-liners. That's still just obviously lame and you get sensibly scolded for it because it makes the site worse, orthogonal to unified theories of voting you may have developed.

[dekhn]

Disagree. It's not particularly lame. A well written one-liner conveys more information and wastes less time. Maybe HN could have a "vetted user", somebody who generally plays by the rules of the site, so that people can see one-liners and know the person making them has the authority to say it.

Perfect example: https://news.ycombinator.com/item?id=30246830&p=2#30247139

[pvg]

It's totally lame and the example is, indeed, perfectly lame. But there's a trivial way to make these good which, if you look at dang-scoldings of lame one liners, works surprisingly well for a huge number of them, not just yours:

Write the one liner and then imagine you got a moderator scolding. Then write a comment defending your one liner from the charge of 'unsubstantiveness'. Finally, delete the one liner and post the defense.

I bet you'd almost never get scolded for these and the upvotes would be even bigger.

[sillysaurusx]

https://www.hnreplies.com/

Dead simple. Somehow it’s also emailed me minutes before the reply became visible on HN. I later realized that the user probably had a “delay” set in their profile, but HN likely publishes the comment to the API even if it remains hidden on the site for a few minutes. So it was funny to see a secret way to bypass the delay feature.

Upvotes are a terrible proxy for value, by the way. One wouldn’t say that a Reddit comment was valuable just because it was popular. The same standard seems true on HN — particularly for fluff, which tend to be upvote magnets.

[dredmorbius]

Collabortive-filtering tends to have a strong brevity bias.

That's only one problem with collective moderation. Moderators are a trust and quality backstop.

https://news.ycombinator.com/item?id=30376111