I don't learn anything on HN anymore, bring back the upvote count | Hacker News Reader

I don't learn anything on HN anymore, bring back the upvote count

1,136 pointswdewind229 comments

I know 1500 people have said this before, but today I guess I'm the one who hit the boiling point.

The value of HN, from the perspective of simply learning, has been destroyed for me since upvotes were hidden.

When I first started reading HN I learned a TON very quickly and everyday about completely new stuff, and was able to do so because I could easily sort through the legitimacy of opinions based on their upvotes. Yes, the upvote system wasn't perfect, but it's a piece of information I can take with a grain of salt. Especially as an engineer who knows little about business, it was extremely helpful to get a community perspective on the startup stuff.

Now, unless I know a lot about the subject (in which case I get limited value from the community), this is just a forum with no easy way for readers to differentiate the noise except for: a) the location on the page, which is useless when comparing a parent to a child reply, and b) the overall confidence and aggressiveness of the poster, which is the exact thing we are trying to avoid judging legitimacy on.

I feel like I just come back here everyday because of inertia and habit. That wont last much longer because I don't feel like I learn anything here: it's just watching people argue now.

I expect massive downvoting but hopefully I'm not the only one...

View on HN

You need to ween yourself off your social proof addiction. The problem with vote counts is that two downvotes can start a cascade of reflex downvotes. Or at best, it won't get upvoted. When vote counts where active, this effect caused many instances where an excellent comment was found greyed out at the bottom of the page until a few smart HN folks with high karma voted it back up.

This is a minor detail that might or might or might not be significant. I personally don't see a lot of comments with negative votes that weren't deserved, but maybe you frequent other threads than I do.

On the other hand, I agree with the OP that not showing the votes has significant downsides. For instance, I don't know whether your comment is something a lot of people agree on (it has a lot of upvotes) or it has just been posted two seconds ago, and is thus the top comment. You won't know whether other people agree with my comment either.

This becomes even worse when it's a discussion about something like security - if someone posts a comment saying that you should always use md5 encryption and someone else replies that this is wrong and you should always use bcrypt which is right? I simply don't know, even though one of the answers may be obviously correct and the other obviously incorrect.

Besides - low scores are about the same as they were before, and high scores are a bit higher. Source: http://news.ycombinator.com/item?id=2465002

> 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.

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).

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.

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.

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.

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.

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

Agreed (if you don't like short comments like mine, bring back the vote count and I'll happily upvote the parent)

Whereas before this would have gotten downvoted, but there is now no meaningful way to communicate your agreement than to leave a comment.

I like the chain of comments more than an up-vote.

Why do you need to communicate your agreement? I receive no benefit from knowing that you, as an individual, like something.

Isn't that the whole point of discussion? Otherwise, all comments would only be contradicting their parent. The upvote count allowed us to balance disagreements (which usually need an counter-argument) with agreement, which don't need to re-iterate what has been said.

Now we get to have a lot of "+1" to counter-balance the criticism.

If we're speaking purely in terms of debate for purposes of Aumann-theoretic truth-finding, then there's many other things you can do to accede to someone else's point than by simply saying some moral equivalent to "Hear, hear!" For example, you can reinforce the parent post with additional justifications, or explore corollaries to its conclusions. You can provide an example if it's a generality, or generalize it if it's a concrete example. Someone I don't know saying "This!" about a subject I'm no expert on is just as informationally-empty to me as that same person posting a single-sentence "No, you're wrong." I see an opinion, but that provides me with no more basis on which to form my opinion.

But agreeing and disagreeing aren't the whole of discussion, either. The posts I enjoy the most on HN are neither of these—instead, they're just comments on the subject of the article, but with the ideas of their parent commenter serving as an additional jumping-off point for whatever tangent they find themselves on. When I think of something that's valuable to the discussion (and therefore deserves upvotes), I think of the parents to these posts—the ones that cause people to spin off tangents in all sorts of directions; those are what I want to see more of here. On the other hand, the expression of purely normative opinions, whether through the voting arrows or through long-winded sermons that still simply amount to "Yay!" or "Boo!", is something I want to see avoided altogether. I can form my own opinions by reading the facts in the discussion and doing research on them, thank-you-very-much—and if I don't have time to do that, then I shouldn't think I have time to form an opinion.

>> or it has just been posted two seconds ago, and is thus the top comment.

you can always see how much time has passed since that comment was posted.

Then you go back and read the source so you can try to run through the calculation backwards in your head to estimate, relative to the surrounding comments' timestamps, what the score might be based on its position?

Idea: a [view count] link. Once clicked, votes are ignored for that comment. You can vote, or get the social proof, but not both.

That's quite brilliant. How about up, down and abstain buttons and if you click any you get to see the vote count and you're done voting. It will encourage voting, remove the social proof influence and give the option to abstain.

Perlmonks.org does this. They only show you vote counts for nodes you've voted on. It always seemed effective to me -- I'd vote on comments I thought were good, and after I voted I'd see what their vote scores were.

How does that option help the main poster? Forcing him to vote on comments (even as abstention) is just meaningless mechanism that gets in the way.

There should be an option to abstain from voting as well, for those that don't know if it's a valid comment or not.

Isn't that just called "not voting"?

I think they mean "not voting but still being able to see how people voted on that particular comment"

About 4 seconds after that feature was implemented, someone would write a greasemonkey script that clicks all the abstain buttons for an HN article.

If you choose to run that script then you're choosing to not vote at all (assuming votes can't be changed, which they can't now.) Which means you can't contribute to the "social confirmation up/down vote pile-on" problem either. So that particular problem would remain solved.

I agree so much I'm commenting.

That defeats the entire point of having people vote who know what they're talking about.

I watch my karma on reddit and HN very closely. It is very rare that I get reflexively downvoted as a result of initial voting, and when it does happen, I mention it and clarify my post and almost always see it even back out.

Oh, are you one of those pathologic "Edit: Downvotes? Really?!?" karma gamers on reddit that seem to feel entitled to upvotes? I think that sort of meta-entitlement karma-whoring is extremely annoying. I get a little chuckle when the person playing that game makes the edit within the ninja-edit window and is missing the edit-star on their comment. Of course even on HN even with the hidden points, you see your own comment points, so feel free to go ahead and play that game here. Or just pretend you're being abused by the ignorant hordes. It's pretty easy. Nobody knows if you're actually negative here anymore.

Edit: Downvotes? WTF? I'd like to know who downvoted this. At least post some constructive criticism.

I'm a karma gamer for being disappointed when rational opinions are downvoted by people that make incorrect assumptions or whose opinion get in the way of a good discussion?

I didn't really ask for your judgement, but since you've taken it upon yourself... I don't "game" karma. The few posts that I've editted here to inquire about downvotes have usually resulted in a further discussion supporting my point, or at the very least opening up further discussion about areas of contention. As it often turns out, no one had bothered checking the child comment below me that said "No you're wrong because of [obvious statement].", when in reality it turned out that actual evidence or a simple screenshot proved my stance.

If I edit my post and point this out and inquire as to why people downvoted me without supporting their position...

That's me "gaming karma"? That's a sad stance. I should just let people say that Chrome 11 lacks H.264 support when a very simple test and very minimal amounts of research prove otherwise? What would you have me do, but edit my post and plea that people take a second to research what they think they know rather than just recalling the infamous blog post that was on HN for a week discussing a phase-out of H264?

School conditions us to care about numbers assigned to our work so it's an easy trap to believe numbers on comments represent correctness or truth. When things are graded incorrectly it's only natural to want to object. I feel the number more as readings of my ability to communicate effectively in the medium with the audience. I wouldn't take low scores personally as a reflection of your skill or domain mastery.

This is the third post on HN that has had massive numbers of people saying that they use the karma for posts to judge for correctness and to add relevancy to the posts. I have stopped reading many comment threads because comment volume is simply too high and I have no filter. It's not so much that I need the ego as I get aggravated that false information is spread to anyone.

It's been suggested before, but the actual problem here is that the functions of the up/down arrows are overloaded and used differently on other sites. Some people use them for correctness, some people use them for entertainment. Personally, I think admitting defeat on the UI is the best option here. The arrows should become essentially placebos. Separate the mechanisms for rewarding insightful comments. Perhaps something more like the "thanks" mechanism in some forums where there is a scarcity to the votes--each user is allowed only 5 thanks/day whether they use them or not. The limit could be seniority based. But if I was running the site, I would leave the arrows and just give up on re-educating the masses to use the arrows differently and move on. It's a battle that cannot be won.

I want that option as an account setting, I'd gladly give up all voting rights in return for being able to see vote counts.

This makes me sad. I always thought that part of being a good user on HN was submitting feedback via comments and voting. To imagine that there are users out there who don't consider this a community, and want only to leech off the useful information contributed by others, is saddening.

In general I would agree, and I would agree that people should do their part in up (and down) voting comments. But to me, with no scores visible, I hardly notice the impact. I always read all the way down the page, and rather than trying to work out how high a comment has been voted based on its position on a page, I just ignore it.

So I'd rather chose to not have any votes and actually get something out of other people's votes than to have an impact that nobody can actually see.

Bad idea. That doesn't help skimming for info, which is the main poster's problem - you have to read the comment before you decide to find out if it's upvoted. And once you've done that, you've already committed the time. Sound nice, but it's not practical.

How does this help the situation at all? And how long before there's a plugin to "click" all of the [view count] links?

The point is that we want to be able to see karma relative to the rest of the thread. If one comment in 20 has 150 upvotes, it means we should at least look at it because what is being said might be important.

Also a [view votes] for the whole page, and you lose your right to vote on that particular topic.

I upvoted this. My own opinion. I don't know how many other silent readers agree, but hopefully I can know soon.

Building on your idea: why not also have the downvote count revealed after pressing up/down/pass? That way we can also know more than the net score tells us.

No. It's overly complicated for the problem that's trying to be solved. I would have to weigh voting on a comment against seeing how many other people voted on it for EVERY comment. I just want to scan smart comments and learn the sides of an issue quickly.

What is the problem we're trying to solve again, anyways? And do the multiple downsides mentioned here multiple times (and heavily upvoted each time) really outweigh the downside of whatever that perceived issue is?

The problem is groupthink. Even HN is not immune.

Unless you create two accounts.

All systems can be hacked and tweaked, you should never write off a system just because its possible to work around it. It's a trade off, and you have to weigh it all up.

A great idea. You get to see votes only after voting, which eliminates social proof inertia and gives an incentive to vote. As long as you can't revoke votes (a la reddit), this would work beautifully.

Pretty sure they're doing this already; have you guys looked at how this thread is sorted? It's not by time of post, implying that your vote is tallied (and counts).

Would having a socket puppet just to peek at votes be grounds for banning?

Would this even be a problem for most scenarios? I don't think most users have (or would want to spend) the time to look at two browsers for every set of comments I want to view and/or vote on.

This is a hacker community. It will take all of 24 hours for purpose-built bookmarklets or even browser extensions to appear.

Seeing vote counts could also be a karma linked feature, I suppose. That makes it enough work to the point where it's probably not worth it to game the system.

I figured that is part of how it works anyway not just upvotes.

I hate this assumption that everybody on HN is stupid. We're not. HN is one of the smartest communities on the internet.

I'm not saying that everybody here is completely immune to what you're talking about, but this isn't digg. We're not just going to heap on the downvotes because we're trying to follow the pack. In fact I've found the opposite to be true; HNers seem to be really good about correcting downvotes. Compare the bottom comments here to the bottom comments at reddit if you don't believe me.

And what you're saying isn't fixed by hiding vote counts, downvoted comments still appear grayed out.

I agree with what you're saying, but I'm not entirely sure this is a totally conscious reflex. I'd be interested in seeing a study where comments are rated with one set showing a low or negative score, the same comment with a high score, and then again with a neutral score. Given to three different groups of course.

I agree that we're smart. That's why I think we can all read things. Points add little value to a good comment, or detract substantially from a bad one. If you don't have time to read, you shouldn't be participating in an active forum like this.

Do you also apply this logic to books in a library?

How do you choose which ones to read? Do you read all of them?

I think if you don't have time to read all of the books and judge them on their merits, then you shouldn't be participating in an activity like reading books.

If I'm interested in a book, I tend to read user reviews. A "351 people said this book is great" note won't help me much. I do the same thing for electronics. And most other things too, when the number of choices isn't very limited already, come to think of it.

I definitely agree, having upvotes visible fostered a herd mentality.

What I really like is there is less of a knee jerk reaction to assigning value to a post. You can assume it's value based upon placement, but in order to determine it's value you must actually read what they say.

What I don't understand is why people need to be told what is valuable. With an ever changing community, I don't see how you could forever trust it's ability to tell you what is a quality post and what is noise.

I find myself upvoting less as a result of this change because the post must actually persuade me to upvote it without the influence of the group mentality. This means my upvotes are now more meaningful, and I think that is a good thing. I believe many people are having similar experiences.

You're right about the herd mentality, but that's the whole point: measuring the herd mentality has a value. No one is asking you to put 100% of your judgment in the point count, and you'd be a fool to do so, much as you'd be a fool to believe everything you see on TV.

Showing points does not kill HNers brains. We've already lost if it does.

It has far less value than you might think. When this community was 100% startuppers, that worked better, but now there are so many people who don't have first hand experience that are willing to upvote anything that sounds correct that it's lost a lot of its value. It's now a dangerous positive feedback loop.

High vote counts tend to be very convincing to all but the most discriminating people, it's not the mark of a fool to believe what everyone is telling them is correct, just a sign of inexperience with that topic.

What you've just told me is the community is broken, not the system.

No, what he is saying is that the system that worked for a small community of like minded individuals does not work for a large diverse community.

The upvote system used to work because it consisted of a lot of similar people saying, this is a good point, or I like this.

Today's community is much more diverse, which is inevitable when you have a quality service/product/community/yadda yadda.

>> In total, Hacker News now sees about 90,000 unique visitors per day. http://newsgrange.com/ycombinators-hacker-news-reaches-1-mil...

The old system just doesn't mean what it used to. People are looking for new strats to preserve the original culture as best they can.

I don't really disagree with your premises, but please explain how removing upvotes fixed the problem.

I could see it going a couple of different ways.

1) It actually has improved comment rankings.

2) or I am just experiencing a placebo affect.

- 1: If it actually fixed the problem, this would mean that quality comments will appear higher on the page, not cluttered by mediocre comments that enjoy the group bump. I pull this assumption based upon the way I use the upvote system. So of course, I could be wrong.

- 2: Maybe the site just feels better to me, because I don't have to see comments get high upvotes that I believe are poorly thought out, or based highly on emotion/fanboyism/or other non logical motivations.

I could see it going either way, but my money is on the first possibility.

All I know for sure is that I now rarely find myself thinking.."Whaat? People think that was a good comment? People thought that added value? People think this guy knows what he's talking about?"

I think it has made the problem worse, sharply worse.

It's just not as self-selecting as it used to be. What worked for the community in the past doesn't necessarily work now or going forward. At least a bit of regression to the mean in the community is inevitable if it grows (which it has, by a lot).

That's simply not true - "Comments with high scores seem to have slightly higher scores than they would have, but comments with low scores seem to have about the same. Probably because the -4 limit on displayed comment scores was already concealing the actual score." http://news.ycombinator.com/item?id=2465002

My observation on my own comments has been that most of them level out at a lower number, say 2-4 when they might have hit 8-10 eariler.

But there are some outliers that are getting voted much higher than they honestly deserve. I'm talking 50-70 range for something that would have been 8-10 before.

I think people might be seeing comments that contribute to the conversation in some useful way and are now upvoting them when they would have let them cap out at a medium score, if they could see the score.

They have no way to know the comment is already considerably over-upvoted.

I mean, thanks for the karma and all, but I don't know that it's really helping.

I read that slightly differently. I read that as meaning there is a greater disparity between the high voted comments and the low comments.

I don't see it addressing the quantity of highly voted comments vs quantity of lower voted comments.

Good point though. It does definitely question my assertion that the herd mentality has negatively affected the value of upvotes. But I don't believe it disproves it.

isn't herd mentality exactly piling votes onto something that's already popular? If so, then herd mentality has been growing while we haven't had visible votes.

My experience is that I upvote less now. Which means I'm more selective about what I upvote. I am no longer affected by the polling of the community.

If this is a common change in other's voting patterns, you might see a shrinking of mid ranked comments, since there are less votes to go around.

If this is the case, I would say that the herd metality has been removed. I never stopped voting on comments that I thought were really good, just on the ones that I was on the fence about, and pushed over by their rank.

The fact that the high comments have seen a bump? Maybe people are reading more in depth, which could get them a few extra up votes? It's a possiblity.

We won't know for sure until we see some actual stats about whats going on. I don't think either side could prove the other wrong at this point.

Hard to count the ways in which mmaunder is wrong here, but let's try:

>You need to ween yourself off your social proof addiction.

Judgmental and prescriptive much?

>The problem with vote counts is that two downvotes can start a cascade of reflex downvotes.

a) You're guessing b) Is that really "a problem" c) Said "cascade" can only actually amount to 3 additional downvotes before the -4 limit is hit. Not much of a "cascade".

>Or at best, it won't get upvoted.

How do you know this? And if you do know it, why are you contradicting it in your very next sentence?

>When vote counts where active, this effect caused many instances where an excellent comment was found greyed out at the bottom of the page until a few smart HN folks with high karma voted it back up.

Yeah, this sentence. Anyway, the effect also "caused" many stupid comments to be rightly downvoted. BTW, high karma does not equal "smartness"; but it's interesting that you seem to like high karma so very much, since your rating is rather high, but hate displaying the accumulated karma on actual posts. Who precisely is "addicted", here?

The -4 limit was discontinued months ago. Other than that, I agree.

  You need to ween yourself off your social proof addiction

I think you are making an unwarranted assumption about the OP.

The OP, like me, gained a great deal of value from knowing the vote counts of comments. In much the same way I (and presumably s/he) gain value from knowing the score and comment count of the post itself.

Vote counts were not perfect indicator of value, but they were better than having to guess at comment quality by observing the relative position of comments vs posting time.

[edit] Interesting. The comments from mixmax (29 minutes ago | http://news.ycombinator.com/item?id=2495535) and blhack (20 minutes ago | http://news.ycombinator.com/item?id=2495561) show higher than my brand new one. Presumably that means quite a few other people have voted them up in agreement

That is absolutely a fault in any social proof system. As I mentioned, upvotes need to be taken with a grain of salt.

I still hold they are far superior to nothing.

Edit: case in point: I'd REALLY like to know how many people upvoted mmaunders original comment.

I agree.

In terms of learning new things, the vote count helps tremendously! You can tell that a security-related suggestion earning 50 up votes is sound (of course considering context), technology-wise. I've learned a lot about passwords, plaintext, server-side hashing, salting and related best-practices solely from HN comments.

Displaying scores might make the other aspects troubling (since group-think is supported unnecessarily, often disregarding novel thoughts or disagreements). But I've learned to ignore such things, especially since most of the conflicts are "opinions" anyway, so the value addition is somewhat limited.

In terms of actual facts and expertise though, nothing can beat the vote count!

Having said that, I am still not sure which one I prefer given that there is inherently a tradeoff between the two aspects.

> In terms of learning new things, the vote count helps tremendously! You can tell that a security-related suggestion earning 50 up votes is sound (of course considering context), technology-wise.

No, absolutely not! You can only tell that other non-experts agree in some path-dependent fashion.

I'll occasionally see highly-upvoted nonsense in an area that I'm expert in. This is very bad.

This is some kind of cognitive bias.

EDIT> I should clarify that it's entirely possible for experts to disagree. So this isn't the you disagree with me so you dumb argument.

But that's a great opportunity for the expert to step in and say something valuable in reply. The high upvote count of the "wrong" response will lead to your response getting more readers.

The fact is we're already half way bought in to this "social bias". Otherwise just get rid of the voting altogether. Get rid of karma. In fact, get rid of associating usernames with comments. But I think everyone realizes, even if they don't like to admit it, that social context provides some value, even if not perfect.

> Get rid of karma. In fact, get rid of associating usernames with comments.

These two things are very different. I'd be all for getting rid of karma, but usernames provide identity.

I.e. when evaluating someone's comment it would be good to see other comments they've made that might provide insight into their biases.

1) username <-> insight into what you think over time 2) karma <-> insight into how popular your comments are

I want (1) and not (2) for my own learning. (2) is interesting in terms of studying group behaviour.

> usernames provide identity.

No the don't, they provide a social signaling function that can be just as powerful as a vote count. When a "popular" name is attached to a post it gets read more closely and frequently voted up. If HN users were truly interested in letting the "quality" of the content stand on its own then comments would not have names attached to them and the content would truly stand on its own.

> when evaluating someone's comment it would be good to see other comments they've made that might provide insight into their biases.

In other words, let the popularity of a username influence the visibility of the content. If you are suggesting that people (including yourself) actually go back and check out user comment histories with any frequency I think you are mistaken.

Karma is not just popularity, it is also a measure of perceived authority and insight that a comment provides to a discussion. The consequence of losing this signaling function may not have decreased the quality of the comments, but it has certainly decreased the utility of the comment sections at HN.

"I still hold they are far superior to nothing."

I disagree. Since the new system was implemented the signal:noise ratio seems to have gone up. The fact that HN is more difficult to read now is a good thing; it forces people to think for themselves, and improves the overall level of discussion.

I do agree that the voting should be shown after a few days, but for right now the changes seem to be making things better.

I disagree with you. The signal/noise ratio has gone down considerably. What I'm noticing now is that people are writing longer comments, but the comments have considerably less substance.

People can't stand out based on upvotes (actual community approval) so they're trying to stand out based on post-length.

At a very basic glance it appears that what you're saying is true, but I've found the opposite.

Removing the obvious noise comments like "+1", "lol", bad jokes etc. how are you defining signal and noise?

This is my entire point: I don't care about the points when I know about the topic, it's when I'm not qualified to decide what is signal and noise that they are valuable.

I can't agree that the signal noise ratio has gone up since the system changed, though this is a subjective, unmeasurable point.

It may improve the "level of discussion" in some abstract way, but it certainly does make it less useful for individual readers.

It was common, before, for me to jump into the comments of a dicussion, check out the few highest rated ones, and then read the article with that information in mind, coming back to read other comments after that. With nothing to differentiate upvoted comments (unless they happen to be at the same thread level in the same thread), there's no way for me to do that, now. I don't view this as a good thing.

Maybe it's making the site less useful to those who are making the site less useful.

You don't seem to accept that other people may have different tastes than you. I often use the vote count to be more efficient in getting a grasp of the issues quickly, without reading everything. Great for you if you like to read everything from start to finish. Not everyone does, at least not all the time.

> You don't seem to accept that other people may have different tastes than you.

No. I'm well aware that others have different tastes.

What I'm arguing is that this feature that some people find convenient and harmless is not in fact harmless, and actually systemically degrades the site over time.

It's a systems problem, not a problem with any one individual.

Well, thanks for being direct, at least.

If points aren't being shown, why have them in the first place? What's the point?

Points should reflect the quality of the comment. If it's not being used for that, it's worthless to have.

Showing them days later, after the discussion is finished, is also fairly worthless.

I'd much rather have some indicator of quality posts. We already have an indicator for poorer quality posts. Their is value in knowing what your peers think.

I must agree with this, I find this feature one of the most missed. It doesn't even have to show the number of upvotes, just anything that will tell me which of the posts is more highly upvoted in a relative manner (e.g. normalise all comments based on the highest ones on the page and use a 5-number/color/whatever scale).

Having the information normalized as you suggest would be great. Breaking it up into a ~5 point scale (especially if it took advantage of visual cues) would save some mental effort. Perhaps relative font sizes?

> The problem with vote counts is that two downvotes can start a cascade of reflex downvotes.

So? As you said, this is usually corrected. And, if points don't matter as so many people want to suggest, where is the harm in having some indicator of a posts quality. Please note indicator does not mean points.

I've been suggesting for some time now that high quality posts should be marked in some manner. No, rising to the top of the page is not a good manner, as it only affects parent posts.

Downvoted posts already get this treatment, even under the current system.

In fact, without some sort of quality indicator, points are useless.

Finally, their is a lot of discussion about how lack of points force people to read and research. That's all well and good, but I'm smart enough to know that I can't research everything, and their is value in knowing what the community thinks of a topic. It helps to steer me in the right direction.

Basically...

If you aren't going to use points to highlight good comments, then having points is useless, and they only serve those who think karma is worth anything.

What if the net number of votes a comment has received is hidden whenever it is between -10 and +10 (for example)? That way you are able to tell that a comment hasn't yet been rated highly, but also can't jump on the bandwagon of voting whichever direction the first couple of voters do. When a comment has a votecount somewhere near zero, you still have to think about it yourself (if you want to read it), but can also choose to ignore it if you only want to read already highly rated ones. If a comment at some point has, say, 12 votes and then receives three downvotes, it would return to the nebulous "small vote-count" status.

Another scheme I've liked on some sites is to show the counts for both up and downvotes, along with the total. A comment's having 10 upvotes and zero downvotes can be quite different from its having 100 upvotes and 90 downvotes, even though both equate to a net count of +10.

You put that excellently. Indeed, Hacker News was and should be a place where social proof is not desired.

Everyone agrees that social proof is deleterious.

Not true, otherwise we wouldn't be having this discussion - the OP is saying that it's a necessary filter.

How can I see who's winning this argument without seeing the vote-count!

"Everyone agrees"? :)

Yeah, I should have more faith in the commenters. Not used to attempts at humor here...

Everyone agrees that my sarcastic jokes are funny. Especially the metajokes.

Only in comments, or should we apply that to submissions also? If it weren't for using social proof to decide which stories should be highlighted, why do we have a voting-based front page?

Indeed, we should all wander through the net with no inkling of others opinions in order to mitigate some dubious risk. Enlightened people read every article, decide for themselves and then sit silently.

And yet what we've left in is the feedback mechanism that provides people a self-centered understanding of how well the community receives them. ie: People are still "scoring points," though maybe we have made the points less valuable.

> The problem with vote counts is that two downvotes can start a cascade of reflex downvotes.

If the comment is good, eventually its going to float up to the top anyway. An excellent article submitted at a time when most HN folks are away is less likely to hit the front page. This is a problem similar to what you mention. Hiding comment scores, IMHO is not the solution. Rather, it just kills the HN experience (at least for me).

On the whole, I concur. However, for comments like your own, the votes are interesting, as a rough measure of whether the HN community agrees with you or not. (It's a rough measure because upvoting!=agreeing.)

Clarification: I don't want to bring back vote counts, or even a general indicator of votes, like others have suggested. I approve of the hiding votes, but a creative compromise could be better than what we have now.

* a rough measure of whether the HN community agrees with you or not.*

It shouldn't be. Like you said, upvoting!=agreeing. It should be a rough measure of you making your point in a good way.

I think there is a problem with the “upvote doesn’t imply agreement” dictum. It doesn’t make all that much sense to me, at least not in all situations.

I can imagine an eloquent, well written movie review with which I don’t agree but which I think is so well written that it should be read by more people. That’s taste and you should certainly not downvote taste, but most discussions on HN are not about taste but rather about questions with right and wrong answers.

When I’m disagreeing with someone about those I must be thinking that what the person writes is based on wrong facts or is illogical. Should users be encouraged to upvote eloquent and well written comments even if they think those comments are spreading false information and are illogical? (Whether or not such comments should be downvoted is a different question. I tend not to.)

It’s never quite that clear cut in the real world. You might feel that the facts are wrong or that some illogical argument is hiding somewhere in the comment but you might not be sure about it. Upvoting may be the right thing to do, then – the comment can be seen by a wider audience and it can come under more scrutiny. The comment and its analysis can help all parties to get closer to the truth.

There is no doubt to me that there are situations in which upvoting despite disagreement can be a good idea. I also think that upvoting because of agreement and not upvoting because of disagreement can be a good idea.

On a certain level, agreement and “making your point in a good way” are one and the same. No matter how eloquent, how can a comment with factual errors that undermine its central thesis ever be an example of “making your point in a good way”?

ugh has a very good definition of what a high vote count means: the comment is a must read. It doesn't mean everyone agrees, but it's a key point in the overall discussion.

You took away such an important buoy for us to navigate HN!

When vote counts where active, this effect caused many instances where an excellent comment was found greyed out at the bottom of the page until a few smart HN folks with high karma voted it back up.

Why would the people need "high karma" to vote it back up? I ask this as someone with high karma. "showdead" is an option available to anyone, isn't it? I don't know the answer but I've always had it turned on at least.

But then, you need to have a little faith in HN and PG. Best posts are usually shown first. Try that: Check a new post and try to guess the votes for the posts at the top of the page. It's still the same algorithm as before.. posts with -4 are shown in light gray.. high voted posts are shown first, etc.

But even you admit that eventually the excellent comment would eventually make it's way back up.

Maybe the solution is to only show positive vote counts. We were already stopped at -4, so not showing any count unless it's above 1 isn't that far off.

I've noticed the same thing. I'll see a debate which is basically a thread of two people replying to each other with no other contributors. In this circumstance, it's impossible to distinguish between a situation where one person gets 10 upvotes and holds a well-agreed opinion and the other person gets 1 or 2 upvotes and is arguing his opinion just well enough to avoid being downvoted.

I understand that we're supposed to form our own opinions as to who we agree with, but sometimes its just not reasonable to take the time to do enough research. Sometimes, you want to learn from someone that actually knows what they're talking about.

Without some sort of vote indicator, it's hard to tell who has the most accurate opinion, except often in subjects like law, security, and seo where there are known experts that often chime in (e.g. grellas, tptacek, patio11).

I'd suggest that a form of "fuzzy vote counts" be implemented. Something to indicate either a relative score ("this comment is substantially higher voted than its parent") or just an approximate value ("unvoted", "few votes", "many votes") without a specific score.

You assume that votes are a good indicator of who holds the most accurate opinion. My experience says that they are merely an indicator of who holds the most popular opinion, or who argues their point in the most convincing fashion. The latter is a particularly insidious case, since accurate information does not always go hand in hand with good debating skills.

I would agree with you. Most topics don't even really have a clear measurement of objective accuracy. And, even when they do, it's rare that votes reflect expert's analysis of accuracy. Rather, votes on HN typically reflect one of the following:

O How well known is the individual - I'm even guilty of mindlessly clicking on a grellas comment when it has to do with Law, Security with tptacek, or patio11 on all sorts of things (Japan, Startups, SEO).

O How popular is their opinion - once again, I sometimes do this myself - when I want an endorphine rush, or my day is slow, I'll go create a comment that I know adheres to HN Philiosophy, just to enjoy racking up 40 or 50 points. Juvenile, I know, but it _feels_ good.

O How effective are they presenting - this is a bit of a mixed bag, and I'm happy to see that frequently very well argued positions are downvoted into oblivion because they are nonsense. :-)

But, with that said - the first approach actually isn't too bad - grellas/tptacek/patio11 actually _are_ worth reading, and their opinions really do count for more than a random individual - so maybe the "this personis well known and has a good track record" vote does have merit.

Regarding your second bullet - do you have an alter ego to test anti-HN sentiments as well? Not really in good taste, I know, but it would be interesting as a social experiment.

I notice that a lot of posters prefix their comments with a "damn-the-downvotes" remark, which shows that a lot of disussion is skewed toward karma-preservation. Both up and downvoting are fraught and maybe you really need both or neither one.

In this particular forum I think votes are a fairly good predictor of accuratese. They're definitely better than nothing.

For instance, I know nothing about security - without votes how am I to know that tptacek is an expert at security and not just some crazy script kiddie that has nothing better to do than post here with stuff he makes up? It seems I hold this community in higher regard than you since I would infer from massive upvotes that what he says is probably correct, and not that he has good debating skills.

That's a very fair point. Unfortunately, when I don't know the subject matter at hand, social proof is pretty much all I have to determine good opinion from bad. Keep in mind that the title of this submission has to do with learning from HN, so that's the perspective I'm taking as well. If you don't know anything on a subject, the most popular opinion isn't necessarily a bad jumping off point for further research. I'm also not saying that you should ignore lesser voted answers on that basis alone. I'm simply saying that in a vacuum of other indicators, vote count works in a pinch.

You still have that rough idea of social proof in the comment ordering. Of course there's the occasional very recent comment sprinkled in, but I think the error this introduces isn't worse that popular, but slight wrong comments upvoted with herd-thinking.

I like it this way even for discovering new things, especially since the lack of comment points prevent my brain from going into autopilot mode. Trying to grasp new ideas without a concerning mind just doesn't cut it for me ...

If you look at my original comment, you'll see I used the example of a conversation with only two participants as an example. In this situation, you don't even have the rough social proof of the comment order.

Ordering is flawed even alone, because it has certain biases in comment age as well. I've seen new comments with no votes shown above other, very highly voted comments.

I should mention that I'm not necessarily campaigning to bring back the vote counts, nor do I necessarily want them back (for reasons you describe well), but I do think they did have value, and we need a way to bring back that value while minimizing the bias. I'm not sure how to do that, but I do really like the "upvote, downvote, or abstain" concept mentioned in another comment on this submission, wherein the votes are hidden until you perform one of the listed actions. You can still view the votes, but you can't influence them after you know what they are.

No one should be suggesting votes = accuracy of opinion, this is demonstrably incorrect.

What I'm arguing is that how popular an answer is is actually a valuable piece of information, outside of its direct tie to how truthful a comment is. Having nothing there doesn't help me determine how truthful a comment is more than knowing popular opinion.

To my understanding, it is generally accepted that the primary purpose of karma points on HN is to encourage good discussions. One of the most common reasons I've seen proposed by people who want to bring back visible scores for comments is "it'll make threads easier for me to skim". Another common reason I've seen is "how can I know what I should think without visible karma points?" I feel that it is worth bearing in mind that neither of those reasons is "bringing back visible comment scores will improve the quality of discussion." I also feel that (with the exception of borderline-trollish comments like http://news.ycombinator.com/item?id=2495648 and "Bring back visible scores because I'm angry today!" threads) that discussion quality has improved since comment scores where made invisible, though obviously that position is up for debate. But, if we accept that the primary purpose of karma points is to encourage good discussion (and making, e.g. skimming comment threads easier is at best a secondary purpose), and (for the sake of argument) we accept that making comment scores invisible has improved discussion quality, then it is clear that the correct decision is to keep comment scores invisible, imho.

You assume that votes are a good indicator of who holds the most accurate opinion.

Mathematically, that's a good bet to make. An information cascade usually produces a good answer (probabilistically).

It's not perfect. But without it, anybody can just throw in their opinion or troll comment and if you're unfamiliar with the topic, you won't be able to tell the difference.

Votes are like a simplified version of facial expressions when talking in a group. Going off of a group's reaction is flawed too, but we do it anyway. It has more value to observers than no group reaction at all.

What about a request for comments feature?

I.e. I see a 2-party thread. I can't decide who's right. I flag it RFQ to attract the attention of other contributors.

Essentially "hey, there's a good argument going on here" pile on (not with votes, but with criticism / argument)

I feel the opposite - I'm learning more, because if I'm spending the time to read something, I have to think critically about it and do research sometimes in order to have a good interpretation. It takes more time, but it's real learning, instead of echo-chamber reading.

It's like it's gone from a cherry-picking exercise to an actual reading exercise. It requires more of a commitment, and it seems to keep the dialog on a more contemplative level.

We're all speaking about "seems" and "feels" here, so it's coming down to impressions (not to mention taste differences), but some ideas for measuring the effect:

- change in the "story point count"-to-"comment count" ratio

- change in average current (and future?) karma of the people who do comment

- change in average comment length

None of these is an objective measure of quality, but they do indicate "change" which might help Paul or whoever make a decision about whether that change was desirable.

I know I'm not really supposed to have an opinion on these things, being a recently-created account and all, but I must respectfully disagree. If you really want to learn, using upvotes as proxy for correctness is a suboptimal way to do so, especially regarding topics without any clear cut answers. Being new to this site, I've actually spent roughly equal amounts of time with and without vote counts an I've noticed a pretty big difference in my own habits. In particular, rather than acting like HN is an omniscient font of knowledge, I treat it much more like wikipedia: a springboard for further exploration of topics.

I absolutely agree - and since it's impossible to know whether a top comment in a thread is there because it was just posted or because it has a lot of upvotes the sorting of the comments doesn't really give me any information. This begs the question: If I can't see the votes and have no easy way of knowing which comments are highly rated why have a voting system at all? What is the purpose if not to serve as a guideline for the users?

I don't use upvotes to decide what I agree with, I use them to decide what to read in the first place. Hiding the votes means I can no longer skim through a thread to pick out the comments that are gems; I'd have to read all the bad comments too. And this means that reading comments here is no longer worth my time, so I don't do it. This is definitely not the desired effect.

This is exactly how I feel. It's no longer an efficient use of my time.

I completely agree. I don't have time to read every comment. I miss being able to skim to crowd-sourced, high up-voted comments.

These high up-voted comments are still exactly where they were before - at the top of the thread.

Not always, that is only used to sort comments at the same level. Often a root comment may have had 2 or 3 points and a counter argument contained some useful insight and would be upvoted to 30+ Skim reading would allow me to spot this comment contains something worth reading...

The highly voted comments still float to the top of the page don't they? When I was only reading the highly voted commetns I started to feel like I was only getting the popular opinion, especially so when it came to touchy topics. If HN does decide to turn the count back on I'd love an option to turn it back off in my own profile.

Highly voted comments don't always stand up when the discussion gets into deep threads. I remember seeing a few threads with 20+ comments and one really insightful response in the middle changing the flow of the discussion.

I feel it is now difficult to have an idea of this kind of "flow" at a glance.

I'd love an option too show or hide the count. Or perhaps show the count when hovering or touching a link for a few seconds.

> The highly voted comments still float to the top of the page don't they?

Highly voted parent comments. Child comments remain children of a parent.

The issue with this is we have no way to compare replies. I don't actually care how nodes on the same level of the tree compare to each other, I only care about how they compare to their children replies (ie: counter arguments).

I've said this in some of the other threads on this topic, but I'll say it again:

The huge (and incorrect) assumption that people are making about upvotes is that everybody reading the comments is stupid. We're not using the upvote count as a 100% perfect indicator of if somebody is correct or not, we're using it as an indicator of how many upvotes the comment has gotten. It's just one of several things that we can use to judge a comment's merit.

This information is useful, and I cannot fathom a benefit to withholding it.

On stackoverflow the criteria for voting are considerably more concrete, yet it's clear that voting there is strongly biased by vote counts. HN might have more rational voters (though look who's saying so), but they're still human, and the fuzzier criteria, shorter active lives of posts, and volume of comments to scan through probably more than make up for this.

(I could've made this same reply to other comments here.)

There are so much more possibilities than just "show vote counts" and "hide vote counts": different sort orders based on more than just votes, collapsing threads, marking people as friends, or adding them to a kill-file. Giving votes different weights based on karma, or average comment score. Tagging of articles, and filtering or sorting based on that. And so on, and so on.

I understand that pg hasn't got the time to do all kinds of experiments, but this is HN, where more than half of us are great programmers.

Give us a simple API and let us do our own experiments. That's all I want.

I guess only pg can answer this, but I'm wondering how the change has impacted voting behavior. Have there been more or less votes per comment? Has the proportion of upvotes to downvotes changed?

I get the sense that my comments have received more rather than less votes since the change. Previously, when a decent comment of mine had say 10 upvotes, I felt like people concluded "good enough" and didn't bother upvoting. Now without the score feedback, I actually feel like some of my mediocre comments have gotten more upvotes than they deserved.

Some behavior data would go a long way toward confirming or denying hypotheses like these.

why can't we just have a setting -- those who don't want it can turn it off...

I've been trying to give it a chance, but so far it doesn't seem like hiding the vote counts has done anything to improve the quality of comments on HN. Honestly, I think it's had the opposite effect. The number of in-depth comments seems to have plummeted.

It's pretty hard to quantify quality of comments, and I've seen at least one other comment in this thread saying the quality of comments has gone up (personally, I'm undecided).

I wonder how much one's perception of the change in comment quality is colored by their initial opinion about hidden comment scores.

I've noticed that the people who seem to think that comment quality has gone down are the same people who used to skim and only read the high-voted comments.

It seems like the people who used to read all of the comments regardless of vote, seem to think that comment quality has gone up.

It's possible that the former groups impression of comment quality might be attributable to the fact that they are now reading lower quality comments that were always there, but which they weren't exposed to given their reading style.

Since HNers are so divided on this issue (http://news.ycombinator.com/item?id=2445039), why not do one month on, one month off for a time to do a sort of A/B testing over a sustained period while keeping everyone at least semi-happy half the time? This would have the added benefit of letting cumulative totals get updated on searchyc from time to time.

And yet the founder of HN said 25 days ago that there is a problem with HN, "the decreasing quality of comment threads on HN."

http://news.ycombinator.com/item?id=2403696

He summed up the problem by saying "The problem has several components: comments that are (a) mean and/or (b) dumb that (c) get massively upvoted." If pg observed a situation like that, isn't it a bad idea to "sort through the legitimacy of opinions based on their upvotes"?

A link and comment in another recent metadiscussion thread largely sums up the back-and-forth about visible comment scores as a signal on comments in active threads:

http://news.ycombinator.com/item?id=2465357

>> Please bring back the comment scores. It helps a lot in parsing the comments and assigning a proportional weight to each when reading them.

> I had to think about this a bit, and I disagree so far. I'm finding that I'm not pre-judging comments as much. It's nice to be able to read someone's comment without knowing first that 70 or 80 or 3 other people thought it was worthwhile.

Once I had thought about that a bit, I reached the second conclusion. Readers on HN are gaining more quality comments as readers look at comments according to their inherent value and not upvoting or downvoting based on the crowd appeal of what someone else has already voted. Cognitive scientists have done a lot of research on what is called anchoring bias,

http://www.sciencedaily.com/articles/a/anchoring.htm

and I get the impression, after 892 days as a registered user of HN, that comment upvotes and downvotes for the last year or so have not been based on the same careful consideration of comment quality as they were in the early days of HN, but rather have been based too much on what net karma score is already displayed as people vote. I like the new comment score system of not displaying net comment scores on OTHER people's posts (of course I can still see my own comment scores) better than the previous system. In the spirit of pg's statement, I try to help the quality of HN by actively upvoting thoughtful, helpful comments, and being on the lookout for mean comments (which pg hopes will not be upvoted by anyone) and dumb comments (which surely don't help any users relying on comment scores to learn new facts) for downvoting those. A comment that is both mean and dumb ought to be downvoted, not upvoted. We can all do our part to help the quality of the community.

In his thread, pg mentioned comments that are mean or dumb "that (c) get massively upvoted." With that condition of HN less than a month ago in mind, how do the highest-voted comments visible in the bestcomments list

http://news.ycombinator.com/bestcomments

look to all of you recently? Are there fewer mean comments than before? Are there fewer dumb comments than before? Are the comments that are "massively upvoted" since the experiment began mostly comments that are reasonably kind and well-informed, helpful comments on the whole? In most of the treads you visit, do helpful, thoughtful comments seem to rise to a position of prominence, while mean or dumb comments gray out?

Remember, pg's claim is that recently HN has not been a place where there is an "easy way for readers to differentiate the noise," but rather a place where the noise has had an attached badge of being signal rather than noise. That isn't good for anyone reading HN. As you correctly point out, we STILL have "the location on the page," which is useful at least for comparing multiple comments at the same reply level as a comment thread develops, and the actual sources and reasoning used by one or another user in supporting points made in comments. People who LOOK UP what the facts are can find out a lot about who is taking care to do good research and who is just making stuff up. And that's always the safest path to learning, to check the sources to verify other people's factual claims.

To be totally honest with you I haven't noticed much of a change in the overall thread quality at all, except that I feel like I'm missing half the picture. The community has definitely felt some growing pains, but it is absolutely not clear that hiding comment scores has helped.

Looking at the best comments page, no I don't think it looks particularly better (or worse) than it did a few months ago.

On that best comments page, look at how many of them are on threads (including this one) asking to bring scores back.

When pg put it to a vote, more people wanted to see them than not.

Since the community was the backbone and validity of the voting system, by taking away the visibility voting system we are taking away quite a bit of faith in the community. If we don't think the voting system is valuable for viewable scores, why is it valuable for sorting the threads? Adding to the fact that a significant majority wanted them shown, it doesn't really seem like we've got a lot of faith in the community as a whole. And if we don't have faith in the community, doesn't that mean HN is kind of dead and we just don't know it yet?

I don't think it is, and I think the upvote system is imperfect but worked much better than nothing. And maybe exact scores aren't what we need, but just sorting the comments isn't enough.

Without the score being displayed I simply can't go through the hundreds of comments that get displayed on HN every day. It would be perfect if I could, but I don't have the time and wonder who has. Yes, displaying the votes supports groupthink, but it's better than the current solution, which is throwing out the baby with the bath water.

I feel compelled to respond to one particular point: Yes, you can LOOK UP facts for yourself rather than rely on the vote count of a comment. But honestly, the expansive tome of information that is HN is not just a series of threaded facts. There is a vast amount of advice, warnings, interpretations, and arguments based on deep experiences and personal trials that are not subject to a simple Google search.

We hold up individual critical thinking to such high esteem that we seem to dismiss the value in the collected opinions of others.

You're missing that even if both popular + insightful get upvoted, you can still use this as a prefilter, look at the high upvote counts, and dismiss that which is mean. It's how I used the system. Right now we just get pages an pages of text with little in the way of friendly direction if you don't have time to examine all the commentry.

How about this solution to improve the readability of HN threads, without degrading the quality of comments/discussion?

1. Show comment-ranges. (1-, 0, 1, 2-10, 10-50 etc.), instead of absolute values. This indicates the quality of a comment, without promoting/prompting voting, based on existing votes (that much.)

2. Do not show username (use a placeholder for profile link). This will solve the problem of "instinctive" up(/down?)voting from commenter's fan base.

3. Any related changes required in the UX, to bubble the better comments more visibility.

Usernames can be very valuable though. If it's a story about Delicious, I'd like to see joshu's comments. If it's about Google search, I'll be looking for Matt Cuts.

I do like the first idea though. It would be a loose indicator of quality without giving too much away.

Well, It is true, but they are exceptions, not the rule.

Potential solutions

1. If the author feels like he is the authority and/or is the important part of the comment, he can always declare it in the comment itself. (The profile-link can always be verified by the users, to ensure that any such claim is correct.)

2. or you can click the links of the comments.

3. or even there might a functionality implemented to show all the usernames on the thread (by default, the author names would not be shown). Later PG can see, if default off makes sense, based on the data captured and user trends (eg. if 99% people are always want to see the commenter name) .

I agree with point #2. I use karma points when scanning comments and read a selection of "pre-filtered" comments. ...user names don't provide much value and I am always slightly disappointed when the most popular comment was made by a popular user.

Valuing ideas should not be a popularity context.

I think the best solution would be for everyone to just start posting "better" commentary (and submissions), and revert back to the old ways of downvoting things that add no value to the discussion.

The "old" HN was a lot drier (for lack of a better word), but the discussions were generally much more informative and/or insightful. I remember people complaining about getting downvoted for a comment that would have probably gotten many upvotes on other sites, and I was always pleased to that these in turn were downvoted as well. This doesn't seem to happen any more though, which is a shame IMO.

The point you make is totally valid; and now that I think about it, when I was able to see the upvotes I could at a glance identify the answers with one or more of the following:

1. Correct Answer (real, like a founder answering about his app). 2. Popular Answer (comedy, something people "lol" to).

Identifying the difference between those two is done with only common sense; but supposing you still lack of that, you can still feel what the community liked by looking at the numbers.

"50 upvoted this" that must mean something vs "3 upvoted this"...

I don't need to know a comment's score, just the score's magnitude provides a lot of utility. In the same way that a grayed-out comment signifies a poor reply, find a way to show me whether a comment is neutral, slightly positive, very positive, or stupendously positive.

I like not seeing vote count, but agree with you that it's a handy metric to sift through the cruft. How about a color scheme to show popularity instead of the specific vote count?

Sure - to be clear I'm not tied to upvotes as a number specifically, and I doubt anyone else is. But give us SOMETHING.

How about sorting comments in any subtree by the number of votes? That way hierarchy indicates the comments with the highest number of votes, while hiding the actual number itself.

That's the easiest and most effective solution, I think. But we could also do with an optional visual cue, since a lot of people want it.

What I found weird is that the points went away but the greying effect stayed. The truth is that that goes a lot further in making sure I don't ready unpopular views (with my eyes I often have to copy paste it to a text editor before I can read it comfortably). I would much prefer all comments be the same font color and being able to see the score so I can make my own decision as to whether to read it or not.

The fact that a comment is grayed out makes me want to read it.

But seeing it say -10 wouldn't accomplish the same thing? Or are you implying that this is literally a swimming upstream situation for you and the shear physical difficulty of the task makes it appealing?

I miss them also. Was a weighted voting discussed as an alternative? One where the up/down vote is weighted against the karma of the voter? Seems like that would help elevate big-karma users to sort of meta-moderators and might help soften the concerns about reflexive voting?

And yeah, maybe for display the numbers aren't the best option.. just some sort of watered down "+" "++" "+++" type scheme...

"because I could easily sort through the legitimacy of opinions based on their upvotes. "

This is a mistake. People frequently vote down things that are right because they don't understand the material as well as they think they do.

Except the mechanic for voting down (dimming out the comment) remains.

Also, my experience is that the down votes are usually countered. Good comments get a good number of points.

By removing points from comments, you've made points worthless for determining comment value. In that way, having them is pointless.

I, literally, stopped reading comments and even started to stop reading hacker news because I didn't feel like spending extra time to read through and filter out the informational comments.

Thanks for posting this. This is exactly how I feel about it.

The one upside to this change I guess is I'm spending much less time on H-N now.

I miss the voting. I don't even bother reading comments anymore.

I could see this change working for other sites, but HN actually does have a sort of collective wisdom.

I've also stopped looking at comments entirely. Before, I would sometimes click through to just read the comments. There was value there. It's not necessarily gone now, but it is much harder to find.

Yes please.

Also, the poll at http://news.ycombinator.com/item?id=2445039 showed that most people agree.

And yet, an astonishing number prefer without. I'd argue that as the particular problem being worked on is increasing group think, this would be a perfect example of herd mentality or popularity contest leading us in the wrong direction.

Throughout threads such as these, the general reason for votes is to avoid reading or to be told without thinking what to think, while the general reason given for preferring them gone is an elevation in discussion.

The second was the stated goal for hiding them in the first place, and the first is the problem to combat -- glib votes.

i agree. i miss being able to skim faster and determine what the community judges as a good reply, I know many others who feel the same way.

Nowdays I just don't read beyond the first few comments. Request to please bring back the upvote count.

Can we at least maybe have a star rating or color rating or something that lets us differentiate high voted comments from mediocre comments? It really does simply take too long to scan and read all the comments. Most of the time I'm just looking for interesting snippets of information, I simply don't have the time to read everything.

Please bring something back, I find myself never reading the comments now.

The only thing you're no longer learning is groupthink. Be glad you aren't learning it.

It's easy to learn in person, just parrot what everyone else is saying. You don't need HN to learn it.

You realize that people game the upvote system right? Writing karma whore comments is so easy when you know the votes. If you really want to know the votes, just reverse engineer the algorithm, it will show you the relative votes.

Lately, I've been reading the "new" headlines as opposed to the "top" headlines to find those hidden gems that _I_ think are interesting.

Why not have a (high) karma threshold, above which you can choose whether or not you want to see vote counts?

I think respected, established members of the community would know themselves well enough to decide for themselves whether it's good or bad to personally have vote counts.

FWIW: Below the threshold I think it's better to not have counts.

I prefer not seeing the number of votes, and think lukifer's idea of a "view count or vote" model is great. A visible score attached to each comment seems to me somehow demeaning. But I've never been much for grades...

And has been stated multiple times, the highly upvoted comments filter to the top already.

I am very sorry, but that is just wrong.

Votes are not a reliable indication of whether or not a post is correct.

They are a reliable indication of whether or not a post is popular.

It does sometimes happen that a post is both correct and popular, but that is not something it is ever safe to rely on.

'Social Proof' is a fantasy that does not exist.

I feel like all the articles that get posted on HN are ones that I come across through a variety of other sources (SAI, TechCrunch, etc.) on Twitter. I find HN to be much more useful for discussions that are started based off the aforementioned articles from other sites.

Hm, this post is only 2 hours old and is already 14th on the front page despite having 600+ points. Do all local posts disappear so fast from the front page, or is this one treated specially? Or maybe its upvote/downvote ratio is low?

One issue I'd like to see explored more deeply is the path-dependence of a comment's final score.

There's this often unstated assumption that "good" comments will rise to the top.

What about this:

1) Display all comment scores, all the time 2) A new comment is posted 3) Segment the readership somehow, so that the comment has a (possibly different) score for each segment. E.g. Allow a comment to have 4 different scores. 4) Start the comment off with a random score in the range [-2, +2], but one that's different for each viewer segment 5) track (but hide) the number of upvotes and downvotes for each score 6) display just init + upvotes - downvotes to each segment

I'm not a heavy user of social networks so I don't know if this exists somewhere but why can't I have a customized view where posters I like are weighted more than others. Then I can do this:

- The poster's weighting would show up as a color (good,better,best)

- The product of the poster's weight and the popularity of the comment determines the order on the page

Now I can look at the top of the page where very popular comments show up and then scroll toward the bottom and quickly identify any posters I really like

I realize this wouldn't be too hard to do by scraping the comment page and maintaining my own database of favorite commenters.

Maybe there should be 2 modes.

One for those just wanting to read and learn where they can see the votes but where the user is not allowed to vote.

And one for those wishing to vote where they can't see and be influenced by the votes of others.

Here's a thought: maybe it's time to close HN. That's right, close it down. PG said all along HN was a production experiment for Arc, so maybe it's time to evaluate the hypothesis and conclude the experiment.

All this hand-wringing about 'the community' and signal/noise indicates the membership has outmanoeuvred any means of wrangling it into some pre-conceived notion of 'quality'. For better or for worse, the thundering hoards of the internet have arrived and are drowning out the elitist clique. If that is bad for everybody, then it's time to shut up shop.

every time someone posts this I add my 2 cents and say the same, please bring back the karma most useful.

Wow. I never realized that people paid attention to upvotes. This explains a lot...

> Especially as an engineer who knows little about business, it was extremely helpful to get a community perspective on the startup stuff.

It's not a community perspective. It's a positive feedback loop.

> The value of HN, from the perspective of simply learning, has been destroyed for me since upvotes were hidden.

Now, you're learning.

Perhaps pg should set a point where the points are revealed, but additional voting is locked out (but not comments)? Perhaps 24-48 hours after an article is submitted? Then provide a link to those articles as they decay into that group for people who really need point confirmation to catch up on what the rest of the community thought?

Or at least shove it in a div or span and hide it by default, so those of us who want it back can style it unhidden.

That would completely defeat the purpose.

because...?

I hate to be "that guy," but I don't see why the count display isn't an option in our account panel.

In the previous polls, people seemed to be split 50/50-ish. Unless pg, as HN Overlord, feels that either removing counts has a performance or pedagogical purpose, this option seems like a perfect candidate for an account setting.

You've hit upon the answer yourself: it can't be a preference because hiding the vote count is intended to have a corrective effect on the voting and commenting behavior of the community at large.

If anyone could switch their vote count on, the majority would -- including myself -- to the detriment of us all.

Deleted comment

I have never. Ever. EVER. witnessed the top comment on an active HN thread be something "witty remarks that a lot of people found funny." Granted, I've only been here a year or so, but comment threads here do come very close to being a true meritocracy, where merit is a function is insight, relevance, wit and brevity.

HN only feels like Reddit comment threads to me on these introspective topics. I avoid them for the most part but your comment in particular stood out to me.

Take any site like this, they all suffer from this problem when the number of users reach a certain point.

I like the slashdot way of organizing stuff. Mods who actively moderate and categorize comments. It's more work, but the end result will be of higher quality.

Learn to read critically and think critically. You don't need prior knowledge on a subject to evaluate a comment. Look for sound logic. Identify and evaluate the assertions and any hidden assumptions.

Vote counts are won't ever reliably reveal truth.

To the people who think that comment points should not be visible:

How do you go about choosing which books to read? Do you read all of them, are you exclusive to specific authors, or do you depend on your peers to make suggestions?

To new HN readers, the comments appear to be sorted randomly. There's no explanation for why some posts are at the top or bottom. That's a definite flaw of hiding the votecounts and could dissuade newcomers.

It is also annoying when you post a reply and the comments have shifted order. I guess PG expects you to start reading the thread from the top again.

Hacker News needs a Meta HN.

Hacker News has long had a rather rarely used Meta HN, the Feature Requests thread linked to from the bottom of most of the main pages on HN, such as the home page.

http://news.ycombinator.com/item?id=363

I don't understand this. I have no problem discerning good quality content and comments from bad quality. I just read stuff. It sounds intellectually lazy and quite possibly self-defeating to rely on the trivialities of other people's votes to determine what is important to YOU.

After all, what is important to ME or not important to ME has nothing to do with what's important to YOU. The fact that you have stopped learning anything on HN, as you put it, has to do with YOU and not ME. So, don't blame the lack of votes. I wonder how you can get along in this world where you have to evaluate things at face value. In the "real world", there aren't point values and upvote/downvote nonsense on things.

I hope HN gets rid of the point system altogether. I'm sure I'm in the minority.

Although I learn a lot from the comments on submissions, I tend to learn a lot more from the submissions themselves.

Maybe the Hacker News team is trying to place more emphasis on submissions instead of comments?

How come this post is not in the first page after 8 hours while the following post for example still is?

Joel Spolsky: Lunch (joelonsoftware.com) 275 points by alexlmiller 12 hours ago | flag | 175 comments

"Ask HN"-types posts (i.e. without a link) are heavily disadvantaged. They take more upvotes to get to the front page, and fall down faster.

is there an explanation for that?

Will setting downvote display limit to 0 instead of -4 help?

Comment age is also a factor: an older comment has more points than a similarly ordered newer comment.

So, when scanning downwards, score is inversely proportional to age.

Karma is a form of currency. It's influence should be exploited not suffocated! There are ways to use it that would benefit the community.

So have two types of browsers -- those who choose to see vote counts, and those that don't. And discount the votes of those that do.

I kind of like hiding the points for active threads, maybe after 24 hours it could show the points. This would be helpful when

I did not even notice they removed the vote count. It doesn't bother me. I don't know how to downvote either though

Having it as an option for everyone to choose if he/she wants it enabled or disabled could as well be a "solution"?

I think you either need an upvote count or the ability to collapse comment threads.

Maybe make the up / down votes on a comment a reward for participating on that thread.

That is, if you up vote / down vote a comment and make a comment on that comment you can view that comments vote count. Does that make sense or did it come out sounding bad?

Just a thought. I think I like it without knowing the vote count.

Edit: Clarification.

Agreed, votes are a good way to filter out the most of noise from the signal.

I disagree. It is forcing you to think for yourself.

What do you want to learn here? What most HNers believe to be true, or do you want to learn to be independent?

By "Be independent", do you mean "go and learn what people who have written books believe to be true"?

Or do you mean "go and learn what people who have written forum posts believe to be true"?

Can we get a user preference, as the case with show_dead and topcolor? I don't see that could be bad for anyone.

"Show score on posts: (yes/no)"

How about the Slashdot system? +1/5 Insightful/Funny/etc?

Could we have a hybrid? We don't need to see the exact number... what if massively upvoted (100+ points) got a green highlight on the "link" URL, and massively downvoted (-10 or more) get shaded, like they already do?

This keeps the content from being so quickly judged, but eventually shows strong content that has been upvoted by hundreds without bias.

Good comments are still put at the top. Mentally filter out the recent comments, and you should be able to see quite easily how "good" a comment is. I honestly don't see the big difference.

What are the incentives for up/down voting?

here here!

Suggestion: Try getting dopamine from reading the comments instead of from game mechanics layered imperfectly on top of the forum.

I did see that as '..bring back the upvote c*unt'. Thought it was a little harsh.

Reply on news.ycombinator.com