Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Website scores kill our success, waste our time (arencambre.com)
77 points by AuthorizedCust on Feb 21, 2021 | hide | past | favorite | 63 comments


On stackoverflow you can spend your reputation on bounties for certain questions (niche, difficult) that would otherwise get ignored.

Personally I spent about 1000 reputation points on those, and some of the answers were surprisingly good.


(I am the OP.) That's an excellent counterexample. Just out of curiosity, how many hours did you invest to get to the 1000 points?


Sometimes if you give a good answer it will just sit there collecting points for all eternity. I did that a couple of times and I have several thousand points after a few years.


As someone has already replied, you slowly gather passive points from your earlier questions.

My account is from 2009, with 33 questions, and 120 answers and 10K points. It's a tough estimate, but I felt like I received my money's worth, sort of speak.

However, bare in mind that contributing was significantly easier 10 years ago, hence why most of my points stem from older questions and answers.


The dynamics of the StackOverflow/StackExchange reputation system is interesting.

The way I look at it is that, if your question meets these two conditions, you shouldn't have to put a bounty on it:

1. it is phrased/framed in a way that is palpable;

2. it is intrinsically interesting.

With that in mind, I've never put any bounty on my questions on StackOverflow. But then I wonder, can reputation points be "invested"? Would it be possible to give a little push to my question by spending some points on a bounty, and then earn them all back and more via upvotes? Has anybody successfully done that?


The bounty system is good for the questions you’re asking after ten years of experience.

If you can’t figure it out, then you’re asking a question with an incredibly small audience. If you’re lucky, some more internet points are just enough to make it worth someone’s while, or as it once happened to me, attract someone who was compelled to help you with the research.


I,ve had several niche questions sit around for a week or two, then get answered within 24 hours after I put a bounty on it.

I,m a dev with QA experience, I can write them well.


The sociologist in me suspects that scoring systems are popular because they instrumentalize interaction in a sufficiently ambiguous way to do work for many audiences. It's really useful that comment scores have vague relationships to a lot of socially important qualities (interest / sentiment / etc). I 'use' changes in my HN karma as a cue that someone is interacting with one of my comments. For me the karma system could be totally replaced by reply notifications - but others might use it differently.

Keeping the score system basic leaves each person to have whatever thoughts they want about their karma without the website telling them how to feel. There's a dark pattern aspect to this where web sites can foster unhealthy levels of interaction while saying that's not what they intended. There's also an emergent behavior side where scoring systems are used in ways the makers can't predict.


(I am the OP.)

The cynic in me says it's just gamification to maximize site owner's net worth.

But maybe it's like "teaching to the test", where if you're teaching to a well regarded test, you're actually doing a good class.

Extending the analogy, if the site's value is well aligned with your self-value (quality of the test), then playing the site's score game (teaching to the test) may be beneficial to you?

Here's an example: I invested some time a few years ago to improve the quality of a few Wikipedia articles. I later found very similar reasoning to what I got into these articles in legislative debates. I think I had a hand in influencing something important to me by updating Wikipedia.


I totally agree that many sites put the scores in to maximize revenue. I'm just trying to consider scoring systems in the wider context of user <-> website interaction techniques.

If score systems were just to generate revenue, I'd imagine that users would have a roundly negative view of them like dark patterns that help revenue (spam, tracking, sponcon). I think you're absolutely correct about scoring systems reflecting your position in the ontology of the sites' creator, and so if you want what the site wants you want your score to go up. I suspect that the popularity of scoring systems comes from users understanding they could be bad or good and being drawn in to the enjoyable play of discovering what their score reflects based on their own experience.


I am with you on the karma being about the same as a reply notification for me, but then I don't care how many likes my holiday photos gets on Facebook.

I know people for who these systems has quite the negative impact. Below are two examples from Facebook.

Example 1: Someone gets really depressed if they don't get about 10+ likes on photos on Facebook. Positive comments are ignored in favor of the number of likes.

Example 2: Someone got really upset and depressed that only a few people wished them a happy birthday on their Facebook wall. Everyone else had done it either in person or via private chat, yet they went down the "no one cares about me" rabbit hole.


HN karma isn't important? I've wasted my life.


I hear if you collect 10 million karma you can trade it in for a pack of Juicy Fruit or Big Red, but that could be entirely hearsay. I'm still optomistic and hoping for the best.


Back in the day they would prioritize high karma HN users for YC startup school invites. That’s how I first met pg.


I like this site, but honestly the most annoying thing is that there isn't a way to opt-out of the karma system visibly. Why can't there be a check box that stops showing me my karma and greying out downvoted posts? There's no way it would take more than an hour to implement this.


Dang actually considers the comment fading an absolutely essential feature:

> The fading is an important element of HN's design—I don't mean its web design, I mean its community design. It's one critical way in which the community signals that it has found something wrong with a comment.

> These signals are extremely important for calibrating users' perceptions of the community—especially new users.

> I think it was one of pg's master strokes actually. For all its annoyance, it creates important feedback loops, both within the community and between the community and the outside world.

https://news.ycombinator.com/item?id=24334263


The first personal browser extensión I ever wrote was to let me hover greyed out HN posts to read them. Dang’s justification is no reason to make them illegible. There’s just as much reason then to let people read it and see what behavior is looked down on here.


The fading disappears if you click "X minutes ago" above the comment, so they're always readable. (Your extension sounds like a nice solution too, though.)


Perhaps it's just your display/eyes? I find them readable even on maximum fade.


I get it, but this position makes HN definitively anti-free speech. You restrict speech and you restrict how people think. Calibrating speech is not free speech and neither is fading speech.

> the community signals that it has found something wrong with a comment.

A bully fat shaming would be "the community signaling that there is something wrong". The HN equivalent would be PC shaming.

What someone thinks of something has no bearing on right or wrong. Also "the community" isn't downvoting. It's always an individual.

A sincere course correcting response is the simple and natural fix, and what already happens. Downvoting helps with reducing responses, but comments are people. People do not enjoy being reduced.

> extremely important for calibrating users' perceptions of the community

I have enough experience now that I can write in HN mode. I cannot however, think in HN mode. I try, but whenever I think, comments like this one happens. This comment also violates the "ideology battle" clause from the HN guidelines because I am not allowed to fight for free speech. But stories about politics and ideologies make the top page all the time, and aren't most good disagreements ideological? Tabs versus spaces, React versus Vue, Open versus Closed source. Downvoting.

The technical part is never that deep. The deepest it can get is deeply technical.

With all that said, everything else about HN is the best on the internet hands down.


> I get it, but this position makes HN definitively anti-free speech

This is absolutely necessary to have a community. The moderation, both centralised in dang and distributed among the downvoters, is what makes HN HN and not a chan board. Removing comments deemed "bad" and thereby discouraging people from posting them in the first place adds value to the site.

It is far from perfect, but it's better than nothing.

> People do not enjoy being reduced.

No, but people don't enjoy being insulted either, and nor do they enjoy the "oh not this again" feeling that makes them reach for the downvote button in the first place.

(Disclaimer: I'm #24 in HN karma and #32 on electronics.stackexchange, so undoubtedly count as deeply embedded with the ""establishment"" downvoting all the edgy rebels)


> This is absolutely necessary to have a community.

Do you want a community/bubble or be open to other's idea ?


If it's far from perfect, let's improve it. It's not downvote or nothing.

Downvotes are meant to tame the insults and offensive comments. But what if downvotes themselves are insulting and offensive? Are there really no other options?

But the even deeper issue is that this may be the HN the moderators want, and it does sound like it is.

Deeper issues are often ideological discussions. Yet, ideological battles are not to be fought on HN.

But the ideology needs to be settled before technical solutions can commence. Implementing a better downvote or an alternative is a technical problem. Deciding to do so requires ideological decisions, such as, prioritize free speech.


So what exactly is wrong with this answer that it deserves to be downvoted?


So I am being asked to self-censor by "the community"? Was this comment downvoted because we're not allowed to talk about downvotes? But the topic is downvoting.

I'm genuinely curious and believe there is a better way to treat users on the internet other than handing out wrist slapping rights to everyone, and permitting petty downvoting without recourse or discourse.

I am not trolling, not spreading propaganda, have no hidden agenda, my analysis is fact-based, and I care. I may not be the best at it, but I try to make my points as accurately and respectfully as possible. But none of that matters.

I'm dead.


I downvote comments which I think make bad arguments. They either posit something without justifying it, or simply re-state what they said in their last comment.

This is not a judgement of the commenter's self-worth—merely of what I think other users ought to see. This way, the most interesting comments rise to the top, the least interesting comments fade away, and HN becomes a more interesting place with a higher signal-to-noise ratio.

In the case of your comment, I personally felt it was just a rehash of your prior comment (which does not appear to have a negative score as of this writing), without adding anything new. The comment reads "it's not downvote or nothing," but doesn't offer a plausible alternative for maintaining HN's high level of discourse.

Because the comment didn't contain anything new (as far as I could tell), I couldn't respond without perpetuating a circuitous thread. My response would likely have been downvoted as well, and rightly so.

I also downvoted your followup "why was I downvoted" comment, because the guidelines say not to discuss comment scores. I like this guideline, because arguing about comment votes is boring and uninteresting. Furthermore, if I answered your question, I'd be breaking the guideline myself. I'm now doing it anyway because you asked multiple times. :)

It's very possible that the other downvoters and I all misunderstood your comment. Then again, if we all misunderstood, it's likely other readers will too, and I'm thinking of their reading experience.

Anyway, I hope you'll stick around and I definitely hope you did not die!


> Dang actually considers the comment fading an absolutely essential feature:

All rulers like the tools helping them reign over their fiefdom.

edit: I actually do most of my HN reading through premii.com, so I ignore all the fading and all the stupid pagination.


the same could be said about flagged comments but you can opt out of the default behavior there. there's basically no excuse imo. if this site was open source a I'd create a pull request, but alas.


You can opt into seeing dead comments, but they're still clearly delineated as dead, alongside maximum fading.


yes, exactly - if what I described existed you could still see your karma by clicking your profile and downvoted comments could still be marked as such by simply saying so instead of ruining accessibility and graying it out, making it harder to read. the ability to click the comment and see it no longer grayed out isn't really a solution since it's still harder to read the comment and ascertain whether it would be worth clicking it to begin with.


I thought that sounded like a fun project to learn web extensions and decided to make a Chrome extension for that today: https://github.com/bdibs/DeKarmaHN. Hope it works for you.


Seems like a user style sheet would fix this.


Personally I couldn’t care less about all those scores. Why should they mean anything to me ? I’ve been on all the same websites as you for years, when I can contribute I do it. When I find a good answer or topic that unstuck me or help me I’m glad. Beyond that it’s noise that I choose to ignore. I chose that online reputation will not have any importance for me and I value this. I don’t want people to « know » me neither companies beyond that’s necessary. My value is not defined by those systems and those virtual points and the idea of chasing them makes me uncomfortable. I used to do all that but like any game, it’s almost pointless and a waste of time in the end so for me I choose not to play, even if I’m « no one » or « invisible ». That’s ok


"When a measure becomes a target, it ceases to be a good measure".

That's not to say anything about any beneficial side-effects, though.

[0]: https://en.wikipedia.org/wiki/Goodhart%27s_law#Generalizatio...


(I am the OP.) That is certainly valid for many simplistic measures. But is it possible to develop measures that are resistant to being engineered? That may negate the article's premise, though...


This one, maybe? https://xkcd.com/810/


The author is trying to fight human nature.

Deprecating scores via cold turkey won’t work. You have to iteratively create newer scoring systems until we finally end up with a scoring system we all agree is trivial and ridiculous (e.g HS Diploma). That’s if you really want to break systems.

So take Leetcode - I’d like to see a new system that cuts it down to about 500 questions across 8 classes (maybe we can live without topological sort questions). As it stands now, Leetcode is adding more and more questions and more and more topics, so the machine gets stronger. Breaking the system requires an iterative approach in the opposite direction.

That’s if anyone’s serious about it. But alas, what is the motivation to be serious about it?

People gain their status by passing LC, or having high HN karma, so what incentive do the winners, ergo gate-keepers, have in wanting to break the system? It would take sheer principle, which brings me back to my initial point. You are fighting human nature.

We are not enlightened, competition is a breeding ground for all things reptilian that are left over from evolution.

Gets back on high horse, rides away (flings cape back so the wind glamorously catches it).


Those scores are a part of the game, but that doesn’t mean they are meaningless. They are just as meaningful as goals in soccer, or stats in RPG, or gold coins in platformer games.

That means: realize this is a game, understand rules, and decide if you want to play this game for the score. If not, don’t waste time of them. If yes, don’t forget to periodically re-evaluate if this game is still worth your time.


For github, there's no point in scores.

For HN, I'd rather it be hidden, something people can peek at if they really want to.

For StackOverflow - a simple rank system would be nice, because you really want to know the cred of someone piping in. You get points (rouhgly hidden) and then a public badge for rank. Very simple, like 4 or 5 ranks max, that's it. I use StackOverflow but have never bothered to investigate the point system deeply I find it confusing and not hugely relevant and I am not alone.

For Twitter/Instagram/TikTok - there actually is something to likes. Perhaps they could flag content as being 'big' 'super big' and 'viral' or something so that people don't worry fret about numbers but at the same time we have some instinct as to how popular something is, even if we concede that it's a very crude measure of anything.


> For StackOverflow - a simple rank system would be nice

There is one which is based on your rank as a user against every other user. The problem is that SO has 15+ million users, anyone who's able to gain a reputation of more than a few thousand ends up in the top 0.1% or so because there's such a long tail of low/no-rep users. They should probably work on fixing this because it's a bit artificial.

I'm one of 976 users out of these 15M+ that has a rep of over 100k (112k), I don't contribute much now, but still accrue points for upvotes on past answers (rightly or wrongly). Though there is a daily rep cap to prevent this getting out of control, but John Skeet :/, that said I did put a lot of work into these answers at the time, so why not keep accruing points, even if my participation has waned. I dunno, I don't really care much now.


I think there are some web scores that are actually useful. The ssllabs one comes to mind.


I think scores on github, stackoverflow and HN may be a net win for all concerned.

Twitter and Facebook scores on the other hand... a massive net loss for all but Twitter, FB, their employees and shareholders.


I'm not really convinced GitHub scoring is realistic or even really representative of much. Kubernetes has tons of stars, yet how many stars do all the packages that Kubernetes consumes have? The ingredients and the arrangement of the parts are often more important than the whole; yet look who is rewarded.

HN scores, to me, stop mattering past about 500, maybe 1000. By then you should know the guidelines, how to conduct some proper discourse, and should have dropped baggage from previous websites like Twitter or Reddit.


I like having more green boxes than not on github. But at least 10% of my little green boxes are just commits little nonsense commits I made for the sake of getting a green box.

Is a potential employer ever going to look over every one of my commits to see what percentage are meaningful? Nope. Will a potential employer look and see a lot of green boxes and think that makes me more qualified or passionate? You bet your ass they will and have.

Is that good? I don't think so but I'll keep doing it as long as it helps me get paid.


> Is a potential employer ever going to look over every one of my commits to see what percentage are meaningful? Nope.

What they'll do instead is look at a semi-random sampling of your repositories and commits. They don't need to look at every commit to get a statistically meaningful sampling of what your commits look like, and they can look at even less to get a vague, rough idea of what you're up to.


That’s the sort of thing I thought employers did when I was younger, but let me disabuse you of that superstition before you waste more of your time. And if you really do believe it, then you can fill in a 100% green grid with a trivial script.


I'm about a month into my first full time dev job. I know that in my job search the activity on github played a role. It is likely much less important or noted for someone with more experience.


It's good for me to see which is the more popular repo when I have to select between 2 with the same functionality. Popolar repos tend to be better maintained.


(I am the OP.) To me, that's an example of you benefitting off others' scores. The premise of the article is to not get overly concerned with your score.


It's great for creating network effects / getting popular.

I understand that some people can't handle attention well, but people who can use it well clearly benefit from it by creating a business out of it, connections, or getting a job.


FWIW, I find the belief that GitHub stars mean anything to be highly offensive.

However, there are plenty that do. And I don’t understand them at all.


It’s basic social proof. Social proof doesn’t tell you everything, but it tells you something. The fact that anyone uses a library is a signal in itself that puts it ahead of a zero star library.

People who have a problem with that are usually, in my experience, the ones placing too much importance on stars, and then projecting that onto everyone else who actually have more apathetic, reasonable views.


While I agree in principle, I don't think these three sites in particular are good examples against scores.

If you completely ignore the scores, just using these three sites "correctly" will naturally increase your karma. I strongly suspect the people who are on any of these sites for maximizing the score are in a vanishingly small minority.


I find HN score partially useful - as an aggregate it tells me nothing, but the score per comment tells me when people disagree, it is a feedback loop. Sometimes I don't figure or disagree with the feedback, but it is a lot better to have it than not to have it.

I don't care of any other score in any other place, it tells me nothing.


I tend to see it as having an impact with new users, but almost everyone stops caring fairly quickly.

You see the same sort of effect with some video games where they make highly visible score or damage numbers, which is motivational early on, but a lot of people find the setting to turn it off if they play it for a while.


OP, you’ll be interested in my proposal from a couple years back: https://news.ycombinator.com/item?id=19745267


Scores often don't just benefit the website owners, but arguably also the other users (and thus also you). Without karma, HN could be a different place.


I think the initial ramp of karma up to whatever the threshold is to allow downvotes is helpful. I'm not sure it has much actual utility past that.


> I think the initial ramp of karma up to whatever the threshold is to allow downvotes is helpful. I'm not sure it has much actual utility past that.

I have to agree (at least, in my case). I know that when I first started here, I was a lot more reactionary and frequently found myself wishing for a way to downvote people who disagreed with my perspective, but I couldn't act on that immaturity because of the required ramp up. I'm still human, I'm still reactive- but instead of just saying "I disagree; you're canceled," it's forced me to take a moment and actually focus on both perspectives and the merits that other paradigms offer. I'm not anywhere close to perfect, but it's certainly helped me grow :).


Is it a fixed numner or it depends on the average upvotes?

I find much more people misusing it than before, but maybe it's just more visible as the total number of votes increased a lot in the past few years.


on my sites, the user,s karma score is directly proportional to their voting weight.


Yeah. Fuck scores. Downvote to oblivion baby!


Is this guy saying you can't pay rent with internet points?

Huh...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: