[tptacek]

I can see why you'd want feedback that worked this way, but HN comment scores never did. Also, +20 isn't a "really good" comment score. Depending on the thread, it's:

* A comment made by someone with name recognition whose comments are read in RSS by 20 people.

* A comment on a political thread that states a clear polarizing opinion for people to glom onto.

* A mystical winged unicorn "good comment".

* A mediocre comment on a buzzy thread ("TechCrunch Says Apple App Store Approval Process Makes iOS Better Than Android!").

My best comments on the site are not, as a rule, my top-scored comments.

Meanwhile, public scores clearly do create problems: they promote groupthink, they prod reactionary voting or, worse, reactionary commenting, and they act as nerd pheromones driving tangential discussions to the tops of threads.

The harm of public scores outweighs the good, in my opinion. Reasonable people can disagree about that point. But I'm not sure they can disagree that there's nothing bad about public scores (which is not an argument you made).

[wheels]

My best comments on the site are not, as a rule, my top-scored comments.

My best comments usually get about 5 points. That's because the best things that I write are usually on very specific topics that I know a lot about. However, the way that breaks down is that there's a tiny number of people who also know enough about those things to know if I'm just talking out of my ass or actually saying something useful. And even if they could tell, there's often not enough backstory for someone who doesn't at least have a passing interest in those areas to make sense of them.

On the other hand, my 50+ point comments tend to be some combination of well-timed, snarky and generic. Basically, they're the sort of thing that HN wants to discourage. As such I'm generally a fan of the scores being hidden.

[tptacek]

Strong agree.

One of my highest-rated comments (inexplicable; I'm embarrassed by it):

http://news.ycombinator.com/item?id=2184067

Meanwhile, here's a comment I actually had to do research to write. Skim it and take a guess how its karma relates to the previous comment:

http://news.ycombinator.com/item?id=1994085

Hint book: ZBER GUNA RVTUG GVZRF YRFF XNEZN.

[pkteison]

Why are you embarassed by the first one? Targeted specific criticism is fantastic and hard to come by. The world needs more "this is bad -and heres why-" and less "yo Dawg I'm just not feeling it Dawg"

[tptacek]

Can we just suffice it to say that I think the scores on these two comments should be --- at least --- reversed?

[jcr]

Of course the scores should be reversed on a place called "Hacker News" but you know the reason why the scores are the way they are; As the popularity of this site increases, the average skill set of the users deceases. Your "embarrassing" post is comprehensible to unskilled people, but your beautiful post is well beyond the skills of most users to understand.

[JoachimSchipper]

Perhaps HN actually prefers "infotainment" over genuinely difficult information? Or prefers things that are within or near the shared knowledge to things it knows little about?

[tptacek]

It clearly does prefer infobullshit over real information, which is why the top of the BankSimple thread is about Dave Fayram leaving the company and not about BankSimple.

[JoachimSchipper]

Yes, that was painful.

[makeramen]

+1 to this.

Among my first comments to HN was a single word: "Yes." http://news.ycombinator.com/item?id=1533506

It has 34 points.

Albeit witty and appropriate in context, I think we can agree it's highly overrated. Meanwhile some of my much longer (and I think better) posts sit with only 1 point.

[gnosis]

Comment timing and choice of article to comment on are huge factors for how highly rated comments are.

If you manage to comment relatively early on an article that winds up being popular, your comment is much more likely to itself be upvoted (or downvoted) a lot.

The later you comment, and the less popular the articles you comment on, the less chance your comment will be voted on at all.

I've often commented on articles which were "old" (by HN time) and on articles which I personally found interesting but which which hadn't received many upvotes, and those comments were very rarely up or downvoted (probably because most people never even saw them).

Comment placement is another important factor. "Top-level" comments (ie. comments without any parent comments) are more likely to be voted on than comments buried deep within a thread. Though replies to high rated top-level comments are often more likely to get voted on than top-level comments which appear later on an article with a lot of discussion going on.

[smosher]

Amen, brother.

Most of the comments where I feel like I said something insightful or genuinely useful end up with a score of 5 or 6. The others are higher or lower (depending on timing.) I don't have any 50+ comments though, I haven't posted much and I try to avoid making the kind of low-signal comment that I expect would get that kind of result.

[guptaneil]

I agree that there isn't a proper correlation between score and quality on most comments, but it's comments about specific topics that I don't know a lot about that I would want to see scores. How am I supposed to know if you're talking out of your ass or actually saying something useful if I know nothing about the topic? The scores allowed me to quickly determine the trustworthiness of a comment.

[tptacek]

I think you just missed the whole thread of people explaining why scores did not allow you to determine the trustworthiness of a comment.

[guptaneil]

There is a subtle difference between quality and trustworthiness. It's possible I missed the thread you're talking about, but as far as I can see, most of the discussions here are discussing quality. Like I said, I agree score does not correlate to quality necessarily. I was simply pointing out that in the case of specialized topics like wheels mentioned, even a score of 5 is a useful signal in telling me that it's trustworthy, versus reading a comment with a score of 1 that could be complete BS and should be taken with a grain of salt.

[jeromec]

even a score of 5 is a useful signal in telling me that it's trustworthy

But that can be a false signal. It's quite possible to have eloquent writers make total crap written in an authoritative style sound believable enough to have 5 people nod and upvote. Having no scores forces people to do their own research and form their own conclusion about trustworthiness, which is more work, but better.

[alain94040]

Agreed. We have been through that argument a hundred times: a high score means the comment is a must-read. Period.

Why anyone wants to hide useful information is beyond me.

[blhack]

Here's a comment I made with +41 points:

http://news.ycombinator.com/item?id=2556647

This isn't a joke, nobody recognizes my name, it's just a comment that I did a bit of research on before posting.

I'm sure there are plenty more like this. Of course some comments get upvoted "incorrectly" (A comment on made about "Winklevoss" being a verb got a lot of points).

How does hiding the points negate this? In my opinion, it does the opposite. If you can see scores, you can decide that a comment is "high enough" and ignore it.

[tptacek]

Be cruel about someone else's infographic and see if you can't quadruple that score.