3 ms·

> For instance, I don't know whether your comment is something a lot of people agree on (it has a lot of upvotes) .... That highlights a major problem right th

> For instance, I don't know whether your comment is something a lot of people agree on (it has a lot of upvotes) .... That highlights a major problem right there. I would hope that upvotes are for worthwhile contributions to the discussion. You (and others) seem to assume that they are for agreement. Different things. I would like to know whether lots of people think a comment is worthwhile. But except in the rare case of polls, I'm usually not much interested in who agrees. So as long as lots of people think in terms of agreement, I say counts can stay hidden.

hackinthebochs 16y ago
Most upvotes are going to be based on agreement, its simply psychological. We're more likely to even consider upvoting something if we agree with it. Downvotes on the other hand, have a detrimental effect such that we can stop to think if this actually deserves a downvote after we get the impulse. So even though ideally upvotes should be about adding to discussion, it's just not going to play out that way.
- dwc 16y ago
  I think it's even worse: upvotes are upvotes, and the reason why someone clicked it varies. It could mean "I agree," "this was well stated and insightful" or even "I disagree and it's dumb and mean but so represents common sentiment that I think everyone should read this." The symbol is a little triangle, and inherently says none of those things. People will click to mean whatever they mean.
- goblin89 16y ago
  Person feel obliged to register their opinion on some question. And upvoting appears the only acceptable way to do this (since "me too"-kind responses get downvoted quickly). So I think the ways to increase voting quality are: 1) allow to register user's agreement/disagreement with the post separately of the post itself (like another voting button), or 2) say in rules that you can't display your agreement or disagreement by upvoting. I think 2 is the best, since users will be forced to make sufficient contribution when they agree or disagree.
wrs 16y ago
This has been suggested before, but perhaps the ability to vote separately on "good/bad contribution" and "agree/disagree" would be a good thing. If nothing else, it might give the "agree/disagree" folks a way to express their views without affecting the sorting (which should be by "good/bad contribution", I hope most would agree).
- moe 16y ago
  I vote both "good contribution" and "agree" on your comment. Your two metrics pretty much nail what I want to know to quickly skim a dense comments page. They don't even need to be numbers, a small graphical representation might work even better. Just give my eyes something to lock on, from there I will likely read the sub-thread regardless of votes anyways. The current state is awful. Just take this thread, with currently 180 comments. No chance am I going to skim even half of them to pick out the good ones. I read the top few, scroll down a little, and then quickly lose interest. - That's the new HN for me.
jwecker 16y ago
While the two variables may be orthogonal for an individual, in aggregate they tend to be the same thing. To illustrate starting with the most stark case- try to imagine a comment that 0% of HN readers agree with but that 100% feel is a valuable contribution. I can't, and the reason I can't is because usually when I encounter the "valuable but disagree" situation individually, it's actually me thinking "some % of other HN readers may share this misconception, therefore it makes more sense to respond than to downvote." or "I disagree with this, but I know most HN readers won't." Or, rarely, "I disagree vehemently, and feel obligated to make that disagreement known because the 5% of readers who actually agree really need to know why it's [dangerous/silly/misguided...]" If the poster is the only one in HN-space that agrees, no matter how "thoughtful" (i.e., verbose, in that case; possibly like the comment I'm writing now) it is, it's not important. The fact that some percentage of readers agree* or may agree is what gives it importance. In aggregate I'd suggest that the number and ratio of upvotes to downvotes would match the importance as defined by the community. In this context I'm using the word "agree" in a sense a little broader than usual. Instead of "I take that position to be my own"- an upvote is broadened to "I can see myself taking that position as my own after further reading/thought/experiment," and finally even "This line of thinking is attractive to the kind of people I like to have in this community." Where things get muddy is when we take that even further and say to ourselves "Some percentage of people will take that position even though I will not. Therefore it is 'important' and I will upvote it." Kudos to the altruism and social empathy, but imagine if 100% of HN readers thought this way. A position that is shared by only 5% of the community would have the same number of votes as a position shared by 90% of the community, even pithy one liners and mean-spirited ad-homimems. In a more likely equilibrium, a percentage of people will do an "importance" upvote for an item separate from their agreement with a probability roughly equal to the number of people who actually agree with the comment (So, there's a 1 in 100 chance that I'll give an "importance" upvote on an item that roughly 1 in 100 people would agree with). That would keep the proportions more or less sane and only result in general karma inflation. Whew, hopefully showed my point- aggregate agreement _is_ relative importance- even if the individuals agreeing and upvoting are not the same people. I could go on with a parallel line of thought for disagreement, although with slightly different results (it's probably a lot less likely that one thinks "While agree with this, it's not important." or "I agree with this and it is important, but x number of people probably don't agree with it so I'd better give it a downvote to be fair...") In fact maybe, the difference between the two sides is key to the bigger questions of site quality degradation. So, to the bigger question: possible reasons for degradation that I've thought about and seen discussed so far- * People trying to decouple agreement from importance and therefore end up trying to channel the thoughts of others, leading to inflation and biases in all sorts of weird ways due to a social instinct that is, by itself, very positive and healthy. * The blind leading the blind (avalanching) that pg is experimenting with solutions to now. * Scariest of all because the fixes are more difficult: The community, in aggregate, is not as valuable/smart/insightful as it was. This compounds the problem with the first point- What happens when some percentage of people (say, greater than 10%) genuinely find "You're a dirtbag" insightful? It trumps the other two points even if they're perfectly solved. What happens when 80% of readers thoughtfully proving comments turns into 30% of readers, with the other 70% just there to "learn" and follow the lead of the 30%? The last point is the most difficult to solve, but there are several methods- there are tweaks such as the "only posts from old members" hack pg did, the green names for greenies, etc. Then there are more frontal attacks: Doing everything you can to lift the collective thoughtfulness and intelligence, and discouraging, sometimes via use of the downvote, comments that aren't worthy of the community. With all three happening at once, the system becomes very difficult to understand. Some suggestions: * Let your vote reflect your degree of agreement/disagreement, AFTER giving it some serious thought. Feel free to not vote either way and to post follow-up thoughts or questions instead first. * Refrain from voting as an emotional reaction or indiscriminately to simply feel more a part of the community. * Try not to take downvotes or upvotes personally. * Never pass up an opportunity to try to crystallize your own thoughts and lift the collective intelligence of the entire community by responding with care, whether expounding or disagreeing. * Never pass up an opportunity to downvote comments that genuinely degrade the intelligence of the discussion ("frist p0st", personal attacks, almost anything that is a single line and contains no wit...) * (more I'm sure but this post is getting way too long) EDIT: formatting

hackinthebochs 16y ago

Most upvotes are going to be based on agreement, its simply psychological. We're more likely to even consider upvoting something if we agree with it. Downvotes on the other hand, have a detrimental effect such that we can stop to think if this actually deserves a downvote after we get the impulse. So even though ideally upvotes should be about adding to discussion, it's just not going to play out that way.

dwc 16y ago
I think it's even worse: upvotes are upvotes, and the reason why someone clicked it varies. It could mean "I agree," "this was well stated and insightful" or even "I disagree and it's dumb and mean but so represents common sentiment that I think everyone should read this." The symbol is a little triangle, and inherently says none of those things. People will click to mean whatever they mean.
goblin89 16y ago
Person feel obliged to register their opinion on some question. And upvoting appears the only acceptable way to do this (since "me too"-kind responses get downvoted quickly). So I think the ways to increase voting quality are: 1) allow to register user's agreement/disagreement with the post separately of the post itself (like another voting button), or 2) say in rules that you can't display your agreement or disagreement by upvoting. I think 2 is the best, since users will be forced to make sufficient contribution when they agree or disagree.

wrs 16y ago

This has been suggested before, but perhaps the ability to vote separately on "good/bad contribution" and "agree/disagree" would be a good thing. If nothing else, it might give the "agree/disagree" folks a way to express their views without affecting the sorting (which should be by "good/bad contribution", I hope most would agree).

moe 16y ago
I vote both "good contribution" and "agree" on your comment. Your two metrics pretty much nail what I want to know to quickly skim a dense comments page. They don't even need to be numbers, a small graphical representation might work even better. Just give my eyes something to lock on, from there I will likely read the sub-thread regardless of votes anyways. The current state is awful. Just take this thread, with currently 180 comments. No chance am I going to skim even half of them to pick out the good ones. I read the top few, scroll down a little, and then quickly lose interest. - That's the new HN for me.

jwecker 16y ago

While the two variables may be orthogonal for an individual, in aggregate they tend to be the same thing. To illustrate starting with the most stark case- try to imagine a comment that 0% of HN readers agree with but that 100% feel is a valuable contribution. I can't, and the reason I can't is because usually when I encounter the "valuable but disagree" situation individually, it's actually me thinking "some % of other HN readers may share this misconception, therefore it makes more sense to respond than to downvote." or "I disagree with this, but I know most HN readers won't." Or, rarely, "I disagree vehemently, and feel obligated to make that disagreement known because the 5% of readers who actually agree really need to know why it's [dangerous/silly/misguided...]" If the poster is the only one in HN-space that agrees, no matter how "thoughtful" (i.e., verbose, in that case; possibly like the comment I'm writing now) it is, it's not important. The fact that some percentage of readers agree* or may agree is what gives it importance. In aggregate I'd suggest that the number and ratio of upvotes to downvotes would match the importance as defined by the community. In this context I'm using the word "agree" in a sense a little broader than usual. Instead of "I take that position to be my own"- an upvote is broadened to "I can see myself taking that position as my own after further reading/thought/experiment," and finally even "This line of thinking is attractive to the kind of people I like to have in this community." Where things get muddy is when we take that even further and say to ourselves "Some percentage of people will take that position even though I will not. Therefore it is 'important' and I will upvote it." Kudos to the altruism and social empathy, but imagine if 100% of HN readers thought this way. A position that is shared by only 5% of the community would have the same number of votes as a position shared by 90% of the community, even pithy one liners and mean-spirited ad-homimems. In a more likely equilibrium, a percentage of people will do an "importance" upvote for an item separate from their agreement with a probability roughly equal to the number of people who actually agree with the comment (So, there's a 1 in 100 chance that I'll give an "importance" upvote on an item that roughly 1 in 100 people would agree with). That would keep the proportions more or less sane and only result in general karma inflation. Whew, hopefully showed my point- aggregate agreement _is_ relative importance- even if the individuals agreeing and upvoting are not the same people. I could go on with a parallel line of thought for disagreement, although with slightly different results (it's probably a lot less likely that one thinks "While agree with this, it's not important." or "I agree with this and it is important, but x number of people probably don't agree with it so I'd better give it a downvote to be fair...") In fact maybe, the difference between the two sides is key to the bigger questions of site quality degradation. So, to the bigger question: possible reasons for degradation that I've thought about and seen discussed so far- * People trying to decouple agreement from importance and therefore end up trying to channel the thoughts of others, leading to inflation and biases in all sorts of weird ways due to a social instinct that is, by itself, very positive and healthy. * The blind leading the blind (avalanching) that pg is experimenting with solutions to now. * Scariest of all because the fixes are more difficult: The community, in aggregate, is not as valuable/smart/insightful as it was. This compounds the problem with the first point- What happens when some percentage of people (say, greater than 10%) genuinely find "You're a dirtbag" insightful? It trumps the other two points even if they're perfectly solved. What happens when 80% of readers thoughtfully proving comments turns into 30% of readers, with the other 70% just there to "learn" and follow the lead of the 30%? The last point is the most difficult to solve, but there are several methods- there are tweaks such as the "only posts from old members" hack pg did, the green names for greenies, etc. Then there are more frontal attacks: Doing everything you can to lift the collective thoughtfulness and intelligence, and discouraging, sometimes via use of the downvote, comments that aren't worthy of the community. With all three happening at once, the system becomes very difficult to understand. Some suggestions: * Let your vote reflect your degree of agreement/disagreement, AFTER giving it some serious thought. Feel free to not vote either way and to post follow-up thoughts or questions instead first. * Refrain from voting as an emotional reaction or indiscriminately to simply feel more a part of the community. * Try not to take downvotes or upvotes personally. * Never pass up an opportunity to try to crystallize your own thoughts and lift the collective intelligence of the entire community by responding with care, whether expounding or disagreeing. * Never pass up an opportunity to downvote comments that genuinely degrade the intelligence of the discussion ("frist p0st", personal attacks, almost anything that is a single line and contains no wit...) * (more I'm sure but this post is getting way too long) EDIT: formatting