What's the main difference on the quality of itens classification using a voting system that allow only up votes versus up and down votes (digg vs. reddit)? YC.news changed from up and down to only up. Why?
I noticed that on reddit the voting on my essays always had a higher proportion of downvotes after 10 minutes than 10 hours. I know there are reflexive upvoters as well as reflexive downvoters, but from the way the ratio changed over time it seemed that a higher percentage of downvoters were reflexive than upvoters. If so then the downarrow injects more stupidity into the system than the uparrow.
Shorter version: downvotes are more likely to be thoughtless.
The best reason is because downvotes piss people off. They have been the cause of more flamewars on reddit than anything else. When you know you are right but everyone is downvoting you and upvoting the "wrong" guy, I know very few people who won't react aggressively.
This is why democracy + free speech (in theory, at least) works. No matter how unpopular, a voice cannot be silenced. If you don't like a person's viewpoints, you can't shut them down by downvoting them; you can only upvote the viewpoint you do like, and / or present your own position yourself, if no one else has done it.
The entire up vote/down vote thing needs to go away. Voting on a story should be related to an action that the user is taking which implies an up/down vote.
For example, in my Facebook application Wildfire stories are up voted only if you pass then on to your friends. As you mentioned to me yesterday, delicious/popular is interesting because it's based on pages people bookmarked for themselves.
I think the hardest part is finding what behavior to track. Click-through is obviously bad. Bookmarking and passing to friends will also slant the stories in a way that isn't necessarily good. I have a lot of friends that don't really care about hacker news, and tend to bookmark more reference-type material.
Comments are useful, but that slants towards controversial topics.
Simple idea someone might want to found something on: provide a little highlighter extension for browsers/in a wraparound frame/as a bookmarklet. Track pages people highlight; title and describe them as the few most highlighted words.
Not sure if I'm the exception to the rule, but I definitely highlight random crap while I read a page (just something for my mouse hand to do)... so I'd ruin your stats in a mean way.
The problem inevitably becomes what do you do once the community has become larger? How do you keep the topics and the level of discussion from sliding downward? This is an issue that Lambda the Ultimate's been facing for the last couple of years as less academic, more industry people join the community, and they've been handling it fairly well but it's taken a lot of work on everyone's part. Reddit's facing the same issue as well. If there are more dumb people than smart ones in a community, they can dictate what articles make it to the front page and overwhelm the good discussion with drivel.
Could always have a downvote, but weight it as far less than an upvote.
Personally I miss having the option of a downvote. Yes I agree downvotes can be an uninformed reflex, but removing any voice of disagreement seems overly harsh.
I'd say have upvotes that count as 10 points, and downvotes that count as 1 point or something like that.
I disagree, as noted further up the thread. Preventing downvotes doesn't remove the dissenting opinion. In fact, it allows and encourages it. Downvotes allow groupthink to silence disagreement. If you disagree with something, instead of downvoting, post a comment explaining your position.
Likewise if you agree with something--then posts can just be sorted by comment count. Of course, it could be said that the upvote button is just a collapsed and made-convenient form of replying "Me too!"
Wouldn't it be interesting, then, if instead of editing a metadatum, clicking the arrow made a duplicate of the post (or, rather, a Post object with the votee as its prototype), which then stacked under the original...
So what happens if something stupid or banal gets voted up to the front page. What if it's something that I don't think is relevant or useful.
Should I upvote every other article to express my opinion? :)
People who agree with an article/think it's important/useful have an immediate way to express that in a useful way - an upvote.
You don't have the same capability if you strongly disagree with an article, see it as unimportant, not relevant, or off-topic. I still think that's a shame.
An alternative would be using flags, like reddit's 'report', the idea being you only use them for stuff that is really unsuitable, like spam, trolls and stupid stuff.
Relatedly, I've submitted quite a few of my blog posts both here and on programming.reddit.com, and they'll often get a good number of points here, but on reddit someone will almost instantly downvote them to zero, and they of course never come back up. So I like News.YC's system.
I rarely downvote on something but I think that having that option does no harm.
Maybe on reddit downvoters are more likely to be thoughtless, I don't think that is a problem with news.ycombinator readers.
It would help a lot to have the same statistics for downvoters on Y combinator.
If you rarely downvote and others allways downvote items they don't like, your opinion is going to be underrepresented. I think that is a harm. As I said elsewhere, downvotes are too ambiguous in meaning.
Won't it be better for ycnews to allow downvotes and silently not count them against a submission but use them for other purposes like measuring thoughtlessness of a user / analyzing the impulse-response of the crowd here etc?
You could make down votes more expensive by forcing someone to type a rationale of why, or, if the dot product of their down vote to the community vote on a post is way off, discard. Finally, maybe have an up vote/down vote bank that can be exhausted for a given time period. Since you have so much free time, anyway.
I don't like downvoting because downvotes are so ambiguous in meaning. An upvote clearly means "read this!". A downvote could mean disagreement on the subject, dislike of the author or just disinterest. I might want to read a story with wich many people disagree, but I might not want to read one that everybody finds completely irrelevant or shallow.
The trouble is that the meanings of up and downvotes are different for news stories and comments respectively. For a comment, the votes almost exclusively mean agreement/disagreement. Only in rare cases do I disagree with a comment but still find it so important that I vote it up. This is different from, say, a scientific paper where this happens frequently.
Overall I think downvotes are unnecessary and cause a lot of bad blood.
One thing I noticed from K5 is that having both options leads to lobbying and information cascades. That is, if one or two early voters post comments saying the article is really good or really bad then that hugely affects whether later voters hit +1 or -1. With only the option to vote up, people might point out flaws in the article or praise it, but the comments are more intellectual and less designed to influence the voting of others. Of course K5 is much worse because the design strongly encourages people to vote one way or another, whereas on this site the bad articles will quickly disappear if people just ignore them.
I think a method of ranking news based on the number and quality of comments would be better than voting up and down. This method would be somewhat analogous to page rank in that the rank of a story (webpage) is indirectly determined by the comments (links) and quality of the comments. Writing a comment is arguable less reflexive than clicking an arrow.
Also interesting news that was generating a good discussion could tend to persist and stay near the top. This would encourge thoughtful comments as you are not penalized for taking some time to think about the matter rather than hastily shooting from the hip.
The problem with this is firstly that it creates a feedback loop, where the more commented something is, the higher it climbs, the more commented it is...
Secondly, it'd be hard to extract a quality measure from comments alone - there are a number of metrics (number of unique users commenting, rate of commenting, amount of voting...) but how do you get from that to a quality measure? How do you distinguish controversy from quality? Or comment threads full or jokes and snark?
Also, commenting on anything you disagree with or dislike would result in that link getting a boost, which can't be right.
All that said, I think comments could be used, but modestly, as part of a bigger algorithm.
But a feedback loop is exactly what is desired. The hard part is to amplify the signal while suppressing the noise. Up/Down voting also creates a feedback loop that I would argue provides a very noisy signal of quality at best and an irrelevant signal at worst. The thesis under consideration is whether comments can provide a better signal corresponding to the item commented on. If this is true, it is desirable that items commented on climb higher and generate more comments,
If you consider comments to be analogous to links, the quality of a news item (or comment) can be extracted as easily as pagerank determines the quality of a webpage. Controversy can not be distinguished from quality because they are orthoganol measures. Many controversial issues (to VC or not to VC) are of high interest to readers of this site and would be considered high quality while off topic political controversies would probably not be considered high quality for this site. In general, most interesting things are controversial.
The existence of noisy comment threads (jokes and snark) is a good point. I would think that some form of Bayesian spam filtering could penalize these comments. Off topic comments could be penalized similiarly if a Bayesian filter was trained using the text of the news item.
I disagree that disagreeing or disliking is disagreable. I would call it conversation.
I don't think a bigger algorithm is the answer. One with a good impedance match would suffice.
I've never heard of anything like this, though I've thought of it quite a long while back for Reddit. The problem is that it won't solve anything. Most people will still downvote / leftvote things they don't agree with and upvote / rightvote things they do agree with. If the desire to silence a dissenting opinion could be controlled by the majority of people, free speech would not be such a rare thing, let alone need to be written into law.
I agree. That is why I purposely did not equate quality with up and down voting, but left it undefined. Perhaps a solution could be that the quality of a comment was ranked by the number and quality of sub-comments. We would have the beginning of a highly ranked comment thread, for example.
I believe the problem with the down voting button is it often reflect more on the opinion of the voter rather than how well the article was written or its usefulness.
Good article that agree with the user = comment/up vote
Bad article that agree with the user = ignore
Good article that disagree with user = comment/down vote
Bad article that disagree with the user = down vote
Unless everyone actually vote up/down purely base on how good the article is and not their opinion, removing the down vote button is not a bad idea...
The voting button is a good idea if it's a poll, but the "hotness" of something is probably better tracked by the comments and other actions a user take...
Perhaps you could start a news.yc Vote Exchange website. I liked the funny linux command line story, so in exchange for me not upvoting it, you can not upvote something else!
related question: i've seen comments here that have, say, -2 points. if you can only vote up, how is it possible for a comment to ever go below one point? my guess is maybe you get a down-vote button once your karma goes above a certain score?
Summary: up only voting measures personal interest. It tends to promote anything that has a constituency. Up and down voting measures community approval. It tends to promote group-think.
I've thought of doing that; I assume when you mean "market" it implies trade. So one way is to invest in shares of a piece of information and get returns to fluctuate based on the demand for that piece of information. But that would die given enough users, so implemented something simpler, but similar.
A prediction market like http://us.newsfutures.com/ works by having an ultimate arbiter of value: Either the event in question happens or it does not happen.
I wonder how one would implement a market based system for recommending articles. Could you elaborate on your ideas?
Sure. If all information was traded like stocks, then, the person who speaks with the most authority would get the most investment. If they become senile, then they lose value. In terms of the recommendation system, then, it would simply recommend whatever/whoever has the highest market value, plus whatever metrics you added in.
You don't have an arbiter though, because information keeps increasing, and at very little cost, and is not limited by resources in the actual economy. Also, you aren't making a yes/no prediction, so you are not measuring something like accuracy (which is measurable). You are measuring a subjective valuation. Steve Jobs might be only 60% correct that the iPod is the best music player in the world, but a fan would think he's 120% correct. Even though the prediction market says he's 60% correct, in the actual market, he earns lots of money. Something like that.
It's not so far from the karma system, but the karma system is inherently flawed because it's based on democratic principles. When you see a likeable merchandise, you don't think in terms of "buy" or "not buy." You think in terms of value, relative to some baseline. A better system would give users the power to express this.
Shorter version: downvotes are more likely to be thoughtless.