Live data from Hacker News

Design: Voting up or Down is Dead

whattofix.com

71–79 of 79 posts

Re: Design: Voting up or Down is Dead

#71
This is why for a new project I just launched I'm using pageviews as the metric (it's a pic sharing site) instead of voting up/down.

Each IP address can only register one unique "vote" per 24 hour period. This makes the system still vulnerable to proxy servers and botnets, but it's not nearly to the point anyone would want to game it yet. (blacklists will go a long way against proxy server attacks)

Re: Design: Voting up or Down is Dead

#72

Earlier quoted context omitted.

Awesome! Color me skeptical about stopping drift, though. It'll be neat to be proven wrong. I wonder how you'd prevent the system from just emulating the editors, instead of the audience. Now with YC perhaps that's exactly how you'd like it, but for a more general purpose site I'd want more of a user-focus than an external-editor-focus. For the newer people, thank you for the link about what the appropriate topics ar…

It does seem like a "weighted" voting system defeats the purpose of having a voting system at all. What's the point of asking everyone to give their opinions if the only opinions that will matter are those that conform to the opinions of the editors!? Of course, in truth I might be quite happy with this system. I've liked what the editors have let through so far (though last few days less so) and I would probably rat…

Weighted voting is different from merely having editors, because there are two steps. The editors don't pick the stories. They (in effect) pick the people, who pick the stories. Sort of like the president picks a Fed chairman, but does not directly control the money supply.

Re: Design: Voting up or Down is Dead

#73
post #4

I think you must have arrived since the explanation of how we're going to avoid drift. http://ycombinator.com/hackernews.html Vote weighting isn't turned on yet, but for the last couple weeks I've been thinking it was time to. Incidentally, the poll problem can easily be fixed by changing the frontpage ranking algorithm to treat poll votes as worth e.g. a third as much.

Before you change the ranking, are there any specific things you like or dislike thus far? I've noticed a lot of new user accounts lately, so maybe now would be good for a state of the union.

Stuff that takes less than 30 seconds to view and vote up is a bad sign. In fact, that suggests a solution...

Re: Design: Voting up or Down is Dead

#74
post #73

Earlier quoted context omitted.

Before you change the ranking, are there any specific things you like or dislike thus far? I've noticed a lot of new user accounts lately, so maybe now would be good for a state of the union.

Stuff that takes less than 30 seconds to view and vote up is a bad sign. In fact, that suggests a solution...

Require a click and a five minute delay for all articles to be voted up?

Re: Design: Voting up or Down is Dead

#75
post #72

Earlier quoted context omitted.

It does seem like a "weighted" voting system defeats the purpose of having a voting system at all. What's the point of asking everyone to give their opinions if the only opinions that will matter are those that conform to the opinions of the editors!? Of course, in truth I might be quite happy with this system. I've liked what the editors have let through so far (though last few days less so) and I would probably rat…

Weighted voting is different from merely having editors, because there are two steps. The editors don't pick the stories. They (in effect) pick the people, who pick the stories. Sort of like the president picks a Fed chairman, but does not directly control the money supply.

Who picks the editors?

Re: Design: Voting up or Down is Dead

#76
post #4

I think you must have arrived since the explanation of how we're going to avoid drift. http://ycombinator.com/hackernews.html Vote weighting isn't turned on yet, but for the last couple weeks I've been thinking it was time to. Incidentally, the poll problem can easily be fixed by changing the frontpage ranking algorithm to treat poll votes as worth e.g. a third as much.

I think one thing that might help is if the choice was a little less binary. A one to five star rating system might help sort out the difference between "a decent way to kill some time at work, so up vote" (3 stars) and "something that will actually change the way that I look at the world, so up vote" (5 stars). Something to think about.

A better solution than that would probably be to add two different types of votes, "agree" and "disagree." Maybe even "interesting." That way you don't have upvotes meaning a bunch of different things.

Then you can indicate the level of agreement/disagreement on some sort of graphical/iconic bar. Having a number of points means almost nothing except that a bunch of people clicked the arrow. With an agree/disagree/interesting system, you can even do math in the background to determine the kinds of articles that certain people tend to like, and you can really figure out their personality that way.

Then again, polarizing things into three categories isn't a very accurate way of discovering what is truly popular in a community. But I do think it's better than selecting certain people to have heavier votes.

Re: Design: Voting up or Down is Dead

#77
post #72

Earlier quoted context omitted.

It does seem like a "weighted" voting system defeats the purpose of having a voting system at all. What's the point of asking everyone to give their opinions if the only opinions that will matter are those that conform to the opinions of the editors!? Of course, in truth I might be quite happy with this system. I've liked what the editors have let through so far (though last few days less so) and I would probably rat…

Weighted voting is different from merely having editors, because there are two steps. The editors don't pick the stories. They (in effect) pick the people, who pick the stories. Sort of like the president picks a Fed chairman, but does not directly control the money supply.

Sure, but you wind up with the monetary policy that the president wants, not the monetary policy that you would get by democratic vote. (Though perhaps the analogy with the Fed is a little akward because monetary policy is something most policy makers agree on, not a matter of contested opinion.)

With weighted voting, the stories that wind up on YC news will closely reflect the desires of the editors. That's fine, but it's not a democracy. It's more like a benevolent dictatorship! (very benevolent ;)

Re: Design: Voting up or Down is Dead

#78

Earlier quoted context omitted.

Yay for mentioning Slashdot! Slashdot has a good system, but for increased granuarity users actually need to think more about what they're voting, instead of what is possibly a subconscious, wholistic, instantaneous judgment, which seems pretty applicable for Digg. So there's a tradeoff for a system to be "easy to use" rather than "intelligent." In this sense the UI becomes a bigger challenge than the data processing…

Slashdot's commenting system is one of the best I've seen. One thing I wish they had though is the ability to filter out types of comments. You can already label comment as "funny","interesting", or whatever, but you can't filter by those labels. I want to be able to read at a threshold of 3+ and not funny.

Are you crazy? You can just set a -1 funny modifier (I set -2, Slashdot's meme humor rampant and unbearable). If you do that, and browse sorted by highest score, Slashdot has by far the best comments anywhere on the net, aside from certain niche newsgroups.

Re: Design: Voting up or Down is Dead

#79
post #30

Earlier quoted context omitted.

That's a bit generous, I think. Most people aren't smart at all. And even the ones that are don't act objectively and without bias. Ever. Although I will admit that moderated solutions almost always come out better. See slashdot's complete lack of this problem.

Slashdot may not have the problem on a story/post level, but have you ever tried to read every comment on a story? A third are hate speech or spam, another third are repetitious or illogical, a sixth are insightful/funny, and the other sixth would be if they didn't happen to be against the wider political views of the people that had mod points that day. I think the only real solution is to have a small community who…

"I think the only real solution is to have a small community whose members are actually heavily invested in the topic... basically, experts."

Commitment to a community is completely independent of investment in a topic.

Post reply on HN