Propose HN: 'Countless' experiment – hiding all karmic integers

[floatingatoll]

I'd like HN to experiment with a "countless" view of the site for a week. As in: all vote, karma, and comment counts are removed from all pages sitewide. An example of this idea is here: https://i.imgur.com/OzE8qgq.png

I suggest this as a mandatory-for-all sitewide test, as otherwise the compulsion to see integer counts (such as karma) will fork HN into two distinct communities operating within the same shared space. I believe that this experiment is valuable only if everyone participates, and since the data will remain present on the back end, no harm is done to anyone's preferred integers of choice while the experiment is underway.

I believe this would be a good experiment for HN as I worry that HN is unintentionally causing addictive feedback loops with its pervasive use of counts. Counts are already today hidden in scenarios where the admins seem to have determined that people behave more appropriately when they are unable to see a precise integer ranking of other's approval. I'd like to give the HN community a chance to experience the value of HN without the pressure of integer-ranking altogether.

The voting algorithms need not be modified for this experiment, as we already are able to flag comments and posts effectively without being able to see how many flags are present. The front-page links that say whether comments are present or not would simply be replaced with the word 'discuss' to remove the pressure to 'be seen' participating in posts, permitting one's interest in the post and/or topic to become the sole determining factor for participation.

I hope that HN is willing to entertain and implement this experiment and am curious to hear how the community feels about the idea.

[DoreenMichele]

permitting one's interest in the post and/or topic to become the sole determining factor for participation.

An awful lot of people here are more interested in the comments than the articles per se. Plenty of people read the comments first and then decide if they want to read the article. Many people participate in discussion without ever reading the article. (There are both good ways and bad ways to do that.)

If you are relating to the site in a particular manner and not happy with it, I suggest you do a private experiment to try to find some new way to interact with the site.

If you want to "offer" people the chance to experience this, I suggest you code up something that lets people try this.

[johnmaguire]

When browsing the titles of Hacker News, I look at the vote and comment count to determine whether an article is high quality or not. I am not sure how I would filter the signal from the noise without this.

[fyfy18 → johnmaguire]

I don't really bother about the vote count, but more the comments. A lot of stories I don't even read, I just go straight to the comments as I find them a lot more valuable (most stories you can get the basis just from the title).

[localhostdotdev → johnmaguire]

you could use https://news.ycombinator.com/over?points=200 or maybe the experiment could allow opt-out

[heavenlyblue → localhostdotdev]

But you can use that one to then brute-force the points for every story.

[localhostdotdev → heavenlyblue]

not so easy for lower scores :)

[JoshTriplett]

> The front-page links that say whether comments are present or not would simply be replaced with the word 'discuss'

Please do not do this. This makes HN far less convenient to browse.

I regularly use the "N comments" information to find discussions worth reading. This has nothing to do with "being seen", it has to do with quickly finding interesting comments. The front-page screenshot you show completely removes that signal.

For a similar example, consider how many email clients show you, without opening the mail, when it was sent and whether it has an attachment. This information helps with prioritization and processing, particularly when you have many emails. Comment counts, and to some degree story scores, serve a similar function, for people who are never going to read every story on even one viewing of the front page.

I go to HN, I open a few stories and comment sections in tabs based on a combination of title, comments, and score, then close the front page and start reading.

This isn't similar to the removal of scores from comments. For those, users were already on a page full of comments, comments themselves are relatively short and shown inline, and even without scores they're still reasonably sorted by some HN-internal metric of interestingness, making it possible to just read downwards and stop when your expected value of further comments doesn't seem worth the time of continuing.

Removing comment counts and scores from stories, on the other hand, makes the front page far less usable.

To look at this from another perspective: consider the myriad stories and discussions here on HN, about Facebook and Twitter and similar trying desperately to get their users to read content they don't want to read. "Let's stop sorting timelines reverse-chronological. Let's show random things from people you don't follow. Let's make it impossible to figure out where you left off. Let's make it so when you refresh (or click-then-go-back) you get completely different content and can't find the tweet you just saw a moment ago." No, stop, don't.

[intended → JoshTriplett]

But that would be part of the point - a barrier to interrupt the ‘addictive’ aspect of a website.

[JoshTriplett → intended]

If you make HN less usable, people will certainly spend less time on it.

That argues for an option, similar to the noprocrast options, not a site-wide change.

[whatshisface → intended]

Should the experience of everyone be degraded to help the subset of people that are especially susceptible to addiction?

[Can_Not → intended]

But I use that information to make my limited time on this site more efficient.

[crawfordcomeaux → JoshTriplett]

Would you be willing to try it for a day or a week? If so, why? If not, why not?

[JoshTriplett → crawfordcomeaux]

No, I'm not interested in HN being less usable and less convenient. I value being able to quickly get interesting news and discussion on HN, and making that less optimized would not be welcome.

[sgillen → crawfordcomeaux]

I think the comment more than answers your question.

[crawfordcomeaux → sgillen]

The comment does not say whether or not they may be amenable to, say, stepping away from the screen when they'd normally be on HN. Nor does it explicit say no amount of time is acceptable nor why.

[jeethsuresh]

I took 5 minutes and wrote this as a userscript:

// ==UserScript==

// @name Remove HN Scores

// @version 1

// @grant none

// ==/UserScript==

nodes = document.querySelectorAll(".score"); for (x = 0; x < nodes.length; x++) { nodes[x].remove()}

[localhostdotdev → jeethsuresh]

css version tested with stylus:

    .score { display: none } /* stories and own comments */
    .pagetop { color: transparent !important; } /* hide karma top right (also hides username and logout) */
    #me, #logout { color: black } /* show username and logout */
    #hnmain > tbody > tr:nth-child(3) > td > form > table > tbody > tr:nth-child(3):not(.athing) { display: none } /* own profile */
    #hnmain > tbody > tr:nth-child(3) > td > table > tbody > tr:nth-child(3):not(.athing) { display: none } /* other user profile */

I don't think it could easily be truly hidden but a css experiment might be interesting.

[jeethsuresh → localhostdotdev]

Nice! I don't use stylus, but it's definitely good to have options for those who do

[onion2k → jeethsuresh]

While that's helpful I think the point of the experiment is to see whether or not it changes what other people post, how we interact with HN, what stories make it to the front page etc. Hiding karma for yourself doesn't do that.

It would be interesting but I doubt a week is enough to see a real difference, and I'd be sad if I didn't get more internet points.

[jeethsuresh → onion2k]

Agreed on the point of the experiment, but I can only influence my own behavior :)

You'll notice that I'm only deleting the score node, and leaving the number of comments, post age, and author intact. Those add considerable value to my HN browsing, and allow me to avoid visiting the comment section or clicking a link that I'll gain marginal (or negative) value from.

In fact, I've always thought of HN (and other link aggregators) as a hub for discussion more than a collection of interesting links - the links themselves are often biased and it's important to learn about multiple sides of the topic lest the bias rubs off on me.

I figured I'd share to lower the bar for those who wanted to see the effects for themselves - anecdotally, it hasn't changed much for me since the points themselves are less important than the way they affect the algorithm. I generally don't post/comment, and rarely vote, so I'm not really concerned about my karma total. Maybe I'm not the target market for this experiment?

[rossenberg79]

If this really is what you want then I propose eliminating usernames entirely. Just enter a password to log in if you want some continuity.

People rarely refer to each other by name here. It’s not important. Only content is important. Besides, a username in no way verifies your identity, you can say you are whoever you want to be.

[ansible → rossenberg79]

There are some consistently good commenters whose usernames I recognize. I will give anything they say slightly more weight than a random stranger.

So usernames are useful to me.

[eps → ansible]

And that's the real problem with HN discussions!

HN already boosts comments from high-karma users. Then people like yourself come in, see them and boost them further because of the poster rather than the content, so the top of every discussion is perpetually filled with the same names.

If anything, a good experiment would be to obscure the usernames and then see if "good commenters" still get the same share of votes - this time based on the actual contents rather than their glorious posting history. It this happens to be true, bring back the usernames. If not, then see how to remedy this.

[dang → eps]

> HN already boosts comments from high-karma users

That's not true. That is, if you're talking about the HN software, it does nothing like that. Ranking algorithms don't look at karma. Karma is used for things like spam filter exemptions and turning on flagging (at > 30) and downvoting (at > 500).

[AnimalMuppet → eps]

I do the same as ansible. But, to answer your objection, I do it because I recognize the name as someone whose opinions I have noted as being worth listening to. I don't base it on their karma numbers.

[dusted → ansible]

Actually, I think that's an excellent point for hiding usernames, since they skew the perception of the actual comment. It might be interesting to be forced to consider what is written entirely on its own merit, and not be biased by who has written it.

[rossenberg79 → ansible]

If their comments are consistently good then they should stand on their own merit. Else you just get idol worship and groupthink.

Names are not as useful as you think.

[ansible → rossenberg79]

> If their comments are consistently good then they should stand on their own merit. Else you just get idol worship and groupthink.

> Names are not as useful as you think.

Ummm... no.

I don't have the time and desire to completely verify everything that everyone says in every forum that I look at.

I doubt that anyone has time for that.

So I will use shortcuts, as does every other human on the planet. One of these shortcuts is reputation.

If someone has a good reputation with me, then that increases the credulity of everything they say. That doesn't mean that if one of the commentators I hold in high opinion says that "Goblins are behind the election hacking" I will just believe it.

However, if one of these highly-esteemed commentators has an opinion about Docker vs. CoreOS Rkt, I will likely take them at their word, if it isn't a critical issue. This doesn't obviate me from doing my on research on technology decisions, but can help guide my efforts.

I don't worship anyone or anything. Everyone's going to be wrong about something, but some are right more often than others, and should be given some benefit of the doubt.

[Tomte → rossenberg79]

Names are incredibly useful, because I know that tptacek does security, cpercival does cryptography and DannyBee does law.

In a thread about legal issues I'll certainly put more weight towards DannyBee's comments, even if he might have totally wrong views on, say, beekeeping.

[nisuni → rossenberg79]

You’ll get something like 4chan.

[nathcd → nisuni]

This reminds me of a thought I've had recently. It'd be interesting to have an anonymous forum like 4chan but with the strong moderation/guidelines of HN or /r/askhistorians.

[nathcd → nathcd]

Also, now that I'm thinking about it, I feel like tree style comment layouts are showing their age as well. I wonder if something like Zulip's layout could work in a more forum-y type of setting.

[isoprophlex → rossenberg79]

yeah, good idea, but I'd suggest hashed/pseudonymized usernames that are salted with the topic id for local consistency. I do value being able to read who is who in a dialogue

[rossenberg79 → isoprophlex]

That could be a good compromise.

[_bxg1]

I definitely find myself having an addictive experience here; the dopamine rush from having an opinion validated is similar to what others may feel on Instagram or the like. Sometimes I refresh compulsively, hoping to see an integer go up.

I don't really see a point to include the number of comments in this omission though; I don't see how that would create a dopamine loop for anyone, and it seems to be the main thing that commenters in this thread are averse to losing.

[brudgers]

Writing a browser extension hiding 'karma' might be a way to explore your theory without a dependency on anyone else working. A blog post or maybe a 'Show HN' could be used to share your experiences with the larger community. The dynamics would be opt-in. Such an experiment would clarify the problem of page position as a proxy for karma, i.e. the role of karma moving items up and down the page.

[bichiliad → brudgers]

Perhaps, but I think a lot of the power of an experiment like this would be to see what topics get voted up without the visual indication of karma (aside from the fact that it's on the front page, of course). You can only really get that with a critical mass of people using the plugin, or with buy-in from the site implementer.

[zzo38computer]

I can do without the vote and karma numbers, but comment numbers are important (mainly in order to see whether or not there are any comments, and if there are a lot of comments; not to determine whether or not the article itself is worth anything, which is something that the reader must decide for himself).

If some users want those numbers visible and some not, one way to do is to add <span class="karma"> around those numbers in order that users can hide them if they want to do.

[dangrossman]

This experiment was run 10 years ago.

https://news.ycombinator.com/item?id=844979

[mkl]

I think this is a terrible idea.

It would drastically reduce the value of the site for me. Titles on HN are often not very clear, especially with the policy of removing editorialising by using the site's own title (which is often pretty meaningless out of context). The vote count is very important for guiding the limited time I can spend on the site.

[chuckschemer]

you'll have more luck with a chrome extension. this place will never do such a thing. "they know best", they will never change from an outside suggestion.

in case you haven't noticed the forum culture is not flexible. hn is a very specifically flawed lens to the world. they're not about to "fix" anything because "nothing is broken"

they will patiently explain to you why you're wrong, then get on with being the unchanging same thing they are.

a funny paradox. a non innovating lens onto a "highly innovative" part of the world.

[dang → chuckschemer]

We actually suggested to floatingatoll that he post this submission, as part of an email conversation we were having. I'm not closed to trying experiments like this.

Would you please stop creating all these accounts and, especially, stop breaking the site guidelines on a regular basis? That's something we actually do feel strongly about.

https://news.ycombinator.com/newsguidelines.html