Exposing the utterly meaningless number seems to hugely encourage playing to the crowd on sites where it's visible. It's one of the worst aspects of reddit that almost every thread has the same pattern. I may be wrong, but it seems like that pattern stems from playing for karma. Not for interest, not for discussion, just points which encourages the smart arse in every sub thread.

Exposing who would just further encourage tit for tat.

Agreed, having never really thought about it until this thread. I like the fact that I can't see vote count on this site

You can on your own posts though. And that is something in of itself. Particularly because you are rewarded with powers for doing “well.”

That's fairly new isn't it? I feel like HN showed scores fairly recently

Well "new" is relative, but it's been around 8 years since the comment scores were hidden.

https://news.ycombinator.com/item?id=2595605

Hmm, not since I've been on here.

It's not only tit for tat but also the herd behavior of following the masses and like the already popular content instead of deciding for yourself if you like something.

One of the most basic ways to combat that is to set the default sort to "new", not top.

Does not eliminate that behavior, but it does very significantly reduce it to almost a non issue.

Just FYI.

True, mostly. I do find it wearing that so many interesting reddit discussions would be so much easier without the 50% that's all the "your mum" jokes and "username fits" easy karma wins.

That varies considerably by subreddit. I do agree with you though. Many Redditors are not longer form type people.

I tend to skim, and when I find something compelling, I power through where the subs do little about those things.

Some subs do cultivate better, longer form writing, and or less of the Reddit "BS."

And, I vary too. It's not always annoying. Mixed in there are some real zingers. Love those.

I confess I vary too - sometimes there's a few really clever ones. I think it's most frustrating in the busier subs where the pollution gets out of hand. I;m much more forgiving of the odd digression on here or the more on topic subs.

We all have our shallow moments. I relish them. :D

I came to the same conclusion and made a Chrome extension to remove several of these numbers (upvotes, karma, etc.) from reddit: https://chrome.google.com/webstore/detail/novote/kepihilioco....

It's not a meaningless number, especially when not exposed. But you're right that exposing it causes undesired behaviour, + makes the number increasingly meaningless (c.f. Goodhart's law).

You nailed it. I appreciate that aspect of this site and it makes reading even the lower posted replies somehow more meaningful and relevant to me.

If number is utterly meaningless it should not be here at all, visible or not.

I don't think it adds anything at all having it visible on my own posts. Occasional surprise when something attracts a particularly high or controversial voting pattern isn't giving much. Keeping the overall total seems relatively harmless and helps identify the low and negative karma trolls.

Ordering on HN by vote count works, and would continue to work invisible to all. The few places it breaks - the political, the climate change discussions etc, are where there may be more voting for tribal reasons.

It risks biasing the comment - for example, even before you read it, you'll see a "7 votes" or "23 votes" and immediately take a glow to it you wouldn't otherwise. Well, I wouldn't trust that I'd be immune, at least. How can there not be a cognitive bias?

I do wish it took more than a single downvote to "grey" a comment. First of all, a single vote doesn't mean much statistically, and shouldn't change the color of text. Secondly, some stuff is just controversial. I've had comments go -2, +5, 3, down and back up.

Fully agreed.

aside: I think ones comment history plays a role in the threshold to grey. For new accounts it’s -1, but that decreases for older accounts. Some magic combo of karma sum average and/or mean, probably?

Likes are like feedback mechanism and it should be shared with intended person only. We don't go and share our exam results nor do we share how much we earn with the world.

Next they need to remove following and follower count.

Neither are apt examples in this case: not sharing exam results is inconsequential, and not sharing how much you earn disadvantages yourself and gives bosses more power to determine how much you earn :-)

I’m one who resisted the change to comment vote display here, and I’m happy with the change. Similarly, at MetaFilter, I’ve adopted the increasingly common practice of hiding favorites. A few Discourse forums hide likes, and so on. People are finding that not having this information makes them happier. That doesn’t mean you are wrong to prefer to see the votes, of course.

I would, though, like if the person you’re replying to could give you a public upvote to signal a friendly end to a conversation.

Agree. I was against it at first but have come to prefer it.

I'm ambivalent about it; remember that every metric can become a gameable target.

Knowing who upvoted seems like a fight-starter.

The old Slashdot system gave the user a choice of upvote and downvote reasons - and the reader could assign their own multiplier! So you could decide that "funny" merited -1 rather than the default +1.

I've never seen that anywhere else.

I believe the idea is that it prevents someone to upvote/downvote solely based on the existing score, and rather based on the actual content of the post.

I remember back in the day that the vote count were public, and there was a lot of resistance to the change. For me, while on a micro-scale I feel like I'm missing information, I agree that on a macro-scale it's a net positive for the community.

Here's the classic referential thread on the matter

https://news.ycombinator.com/item?id=2595605

Why do you TRUST the numbers you see on social media sites (including Hacker News!)? They should not be your tools for reasoning about the world (not even a little bit) - there's no way you can know if those numbers were created by the community, bots, the site owners, or "other".

Being influenced at all by numbers where there is no accountability for site owners of their origins seems increasingly like a bad idea.

Agree totally with this. It also encourages people to not think for themselves, further increasing the herd mentality. The lame defense sometimes given for this is - Well people do not have time so the numbers help them decide. That is another load of crap, the illusion that people don't have time. People should consistently question what they read and 'like' things they truly love.

That said, I am trying to think how this feature of 'hiding likes' can also be tampered with because that is human nature

It also reinforces us to confuse popularity with quality, a common mistake on the Internet going all the way back to PageRank.

I believe it's fine to hide this information if the algorithm to rank posts is obvious and easy to reason about. For example, you know on Hacker News that the top posts are the ones with the most votes. More detail doesn't give much more information.

But Instagram is like an Ocean of content, there is no clear boundaries between topics, it's an endless stream of content and events happening in parallel, AND the algorithms are known to be opaque. So the information of number of Likes are valuable to relate to other events which happened in the past. This also applies to Twitter.

By the way, we can see the total votes on the front page of Hacker News, because it helps to put things in perspective according to past events.

Instagram should hide the numbers of like in messages in the discussion of picture, but IMHO they should keep at least the order of magnitude of Likes under the pictures. ("10s Likes", "100s Likes", "1k Likes", etc)

It looks like you probably weren’t around at the time the change was made (though you may have been lurking). There was a lot of popularity voting at the time, which didn’t enhance the site. Patio11 could post anything and get 300 upvotes, just because, patio11. And arguments tended to be really overloaded with votes in both directions that caused more inflammation than anything else. I’m happy with the way it is now.

Well, another big site explored this. DailyKos made votes public.

You are not wrong. The info was very useful. One could get a map of users and a real sense of their politics, preferences, etc...

I did, and learned a TON about online dialog. Knowledge I have put to use elsewhere to great effect.

But the cost! Damn. The truth is that data generates an insane amount of meta. You name it. Cliques, brigades, retribution, beauty contests, the works!

Considering(from professional experience) how much vote manipulation occurs on online forums like HN, I'm certain no companies with sites like HN will ever implement that. I know plenty of people that make a living(or at least side money) providing voting for sites like reddit and HN and if you had that data you would realize how few of these votes are organic in nature.

It would be interesting to know the outcome the other way round: still use the numbers and the upvotes in the back end, but hide the up one count from everyone including the poster.

My guess would be a higher quality discussion, but a less addictive ui.

It would create some level of accountability, similarly if they're going to keep downvotes - showing who did it could add value; if retaliation is a worry by people downvoting, then perhaps they should reconsider not adding a downvote - and simply letting a comment fall away, or consider putting effort into writing a comment, using critical thinking to explain the dislike, as opposed to satisfying and processing their urge via a quick click of a downvote - which benefits no one.

You can't reason about the world through voting numbers in a forum primarily visited by tech people from SV. The bias is just too strong.

That's what meta-analysis is for!

>It feels like the system is hiding information from me that I would find interesting

Sometimes the behavioral impact matter more than full transparency. I definitely think it contributes to HN being higher quality than reddit for example. People say things for the sake of it not to gather 100 meme points

Has anyone done any studies on the effect of the change at HN?

Ir seems like it would be a really rich dataset. On the simple side of the analysis is simply what it did to the vote counts. Did they stay the same, or go down? What happened to the ratio? On the more complex side you could do some semantic analysis to see if the comments have changed in character at all.

It's one of my favourite features.

I feel like it does a great job of stopping the focus on karma. People who write bad comments still get downvoted/flagged/treated appropriately but everybody else doesn't feel they must write a comment with karma in mind.

And this is why you never put an engineer in charge of UX. It's such a coder answer to ask to see all the data, and it ignores everything about user behavior and human psychology. As much as we'd like to think that we're above such things, the reality is that seeing the numbers biases users and has a clear negative impact on how you read the comments.

Being engineer or coder does not mean that they don't understand human psychology and behaviour!

Public upvote counts would allow me to create a data set for analysis that would then allow me to manipulate HN to get lots of upvotes and control / influence the discussion using the knowledge I have data mined.

I am not to be trusted with such a capability.

Nor is anyone else on here.

To me, if a comment is faded or not, then that's enough info.

I enjoy hackernews for precisely this reason