Blindly submitting content to HN, good or bad? | Hacker News Reader

Blindly submitting content to HN, good or bad?

54 pointsbenologist44 comments

I've noticed an account that's submitting 80+ stories a day all from generally liked sites.

I think it's bad, at the least it automatically pushes everyone else a 1/2 a dozen spots further down when another batch is dumped making it that much more unlikely someone else has a fair chance. It also reminds me of digg's 'power users' where a handful of people ended up dominating the site, this isn't happening right now but when you're submitting 10% of stories the front page will reflect that eventually.

But on the other hand some of the submissions would have been made regardless by other people.

Thoughts?

View on HN

80+ stories a day is excessive. I would be in favor of PG rate-limiting submissions.

If each article takes five minutes to read, it would take the submitter 6 hours 40 minutes(!) to read all the articles he submitted. This implies the submitter is not reading the articles before submission.

This worries me because it is indicating submission by quantity, not by quality, and the thing I love about HN is the quality of the submissions.

If some of the submissions would have been done by other people, that's fine. IMHO...

a) Don't hog the karma.

b) As pointed out by benologist, the weaker submissions would not have been submitted by others. The weaker submissions are therefore competing against other submissions that are more deserving of our attention, contributing to the dreaded "deterioration of HN quality."

Don't hog the karma.

For me, I don't even care so much about the Karma but if one person is submitting 80 stories then it either creates a one dimensional feed (if the user is actually reviewing those articles) or it is automated and there is no guarantee of quality. Either way, I believe that it can only serve to decrease the quality of the site. I would be in favor of submission caps. I mean I read maybe 5 articles a day, that are worth re-posting to a larger audience. So capping it at 5-10 I think would not hurt the site, but stands a chance of improving the new feed (which is where I hang out)

Users should be rate limited based on the average score of their last x (say 100) submissions. If their last 100 submissions got an average of 1 (no points) then they should be rate limited to 3 a day. If their average score for the last 100 is above, say, 30, then it's obvious the user is providing real value to the community and their rate limit should disappear.

Maybe if everyone was rate-limited to, say, one submission per day the /newest section would be much cleaner. Ideally, it would also cost a little bit of karma to submit links.

I don't like that. If anything here smells like Digg's "power users", this is it: the more you can post, the more karma you gain, and up goes your ceiling.

That's the point of the 100 most recent. If they are power users and they can submit 80 posts a day, each of which gets at least 20 points, then they are obviously making a contribution. If they just have high Karma though, they won't necessarily be able to have a high limit. If they keep submitting crap in the hopes that a few will be upvoted, their average will go down.

Also, you could try using median rather than mean and that might produce better results.

One thing I've noticed happens quite a bit is that a thread on one story will have someone post a link to an external site. Someone takes this and then submits it as its own story. Take, for example, this comment on the "ACTA will force border searches" story:

https://news.ycombinator.com/item?id=3496269

Two hours after this was posted someone posted the link within that comment as its own story:

https://news.ycombinator.com/item?id=3496475

Normally it's not a bad thing to have a new discussion if the subject is too tangential to the main item being discussed. But this is close enough that the discussion probably ought to have happened directly within the thread.

Additionally, the person who submitted it as its own story received (as it now stands) 145 points for the effortless submission, far more than would be likely for even an incredibly well-written and researched reply within the original thread.

HN has gamified participation by rewarding people with karma. Even though it's only a bullshit number, it triggers a desire to maximize this number. As it stands, submissions are too highly rewarded relative to the value and behavior like 80+ submissions per day results.

Normally it's not a bad thing to have a discussion if the subject is too tangential to the main item being discussed. But this is close enough that the discussion probably ought to have happened directly within the thread.

Similar to this is when the same event is reported by different sites, each of which gets submitted. We end up with essentially the same discussion happening in multiple places.

Ideally there would be some magical way to have all these posts direct to a canonical discussion, but in practice it seems this requires human intervention.

In the meantime perhaps the best action is to flag the late-comer posts and post a discussion link to an existing thread.

Not necessarily. Google's news aggregation has gotten pretty good at grouping numerous stories on the same subject together. It doesn't always get things right, but helps to reduce the clutter.

Right, while this is technically possible (and really, anything is technically possible when it comes to programming) I'm not sure if PG and whoever else maintains the site have the time, energy, and resources to set up such a system. I'm not sure what it would really take but I assume it'd be quite a bit of work to put it conservatively. The way things are run now are, in some ways, base on a compromise between the site maintainers and the users. It's as if to say "We'll manually intervene on occasion if everyone promises to play nice". Once you have to start implementing the kind of automatic quality controls we're talking about here programmatically instead of relying n good behavior, that's when you know the community just lost the quality it was known for.

> Ideally there would be some magical way to have all these posts direct to a canonical discussion, but in practice it seems this requires human intervention.

Some, but just to flag the duplication; how to process the dup is a matter of policy and could be largely automated once the humans had annotated it. Some ideas in that vein: http://news.ycombinator.com/item?id=1975950

Actually, I find it most annoying if the link to a original source gets less upvotes/discussion than a rehash of the same story from a popular news site. Therefore, a link to the original source of a story should never be seen as a dupe, even if it is a late-comer.

I am the author of the comment you mention. While I also wondered about this coincidence, I do not see the re-post as a problem. The re-submitter made this link more accessible to a lot of people. Overall, that's a good thing. I actually think I have found this link somewhere in a HN discussion, too, but could not find the discussion again.

(For the record, I collected 13 points for my comment so far.)

When we notice people doing this we rate-limit them. I just did.

What are your thoughts on allowing people to downvote posts (akin to how comments can be downvoted)?

I believe that is what the flag button is for. Down vote button on a story could lead to people downvoting something they don't like because it is the opposite of upvoting. Flag is more of a "this is an inappropriate story" word.

But this problem also affects comments - shouldn't we remove the downvote button for them?

There's less of a feedback loop there - people don't know if their submissions have been flagged or not.

On behalf of authors everywhere, please don't use the flag button as a downvote button. This, for several reasons:

1. It's overpowered. Based on some experiments I conducted a year or so ago, it seems like 1-2 flags equal about 10-20 places dropped in the rankings. It's brutal. A single person in a bad mood can drop a story like a stone. I don't know the exact maths, but a single flag seems to have the effect of many downvotes.

2. It's not what it was designed for! According to the guidelines, flagging is meant to be for "spam or offtopic", not for stuff you don't like or disagree with and want to downvote. If you disagree with it, write a comment to explain why, or find a comment that agrees with your disagreement and upvote that. Flagging is not a way to express disagreement!

3. It's mean. I regularly get articles on the front page, and I don't care that much what happens to them (though of course I like to see the discussion they generate!), but once upon a time I did, and I'm sure that's still the case for many. When you put your heart and soul into an article, and it miraculously got picked up by HN, and it's on the front page, and people are finally checking out your blog, and it's getting more visitors in one hour than it had in the last 6 months... and you see the story suddenly vanish to the second page because someone flagged it... wow, that feels like a punch to the stomach. So please don't flag "real" articles, only flag spam/trash.

If you think an article sucks, post a comment to explain why you think so. Maybe you misunderstood. In any case, the feedback will be clearer. Please don't use the flagging sledgehammer to express disagreement.

(note: I don't know if pg would agree with the above, but these are my feelings)

> It's not what it was designed for! According to the guidelines, flagging is meant to be for "spam or offtopic", not for stuff you don't like or disagree with and want to downvote. If you disagree with it, write a comment to explain why, or find a comment that agrees with your disagreement and upvote that. Flagging is not a way to express disagreement!

Disagreeing with something is not a reason to downvote either.

In your perspective, what are valid reasons for downvoting?

In the perspective of many people here, downvotes are to indicate that you don't think the comment contributes to the conversation in a positive way. For example, I upvote comments I disagree with as long as they make the point well and give me something to think about. Similarly, I downvote comments I agree with if they're snide or poorly constructed.

Yes, I think many people are under the mistaken impression that the upvote/downvote buttons are functionally symmetrical, which is understandable given that they are presented as visually symmetrical. But upvoting serves to avoid redundant 'I agree.' posts, while if users downvote to disagree, it robs the discussion of valuable content (i.e., why they disagree). It would help if the UI were modified to signal this asymmetry.

I'd fear you'd end up like Reddit where there is a large fraction of the submitters who downvote all other posts in a misguided effort to make their own reach the front page.

Paul - Why not create a "comment karma threshold" to remove the incentive for new users to blindly submit stories?

For instance, users won't get karma points for stories they submit until they reach some threshold for karma from their comments. This would be beneficial in a number of ways:

  * encourage better comments from new users
  * prevent downvote capability for new users that only submit a lot of stories
  * fewer low-quality stories will be submitted

This would seem to imply that comments are somehow "superior" to stories, and I'm not sure that's supposed to be the case.

I get what you're saying, but I am not sure it does.

When I first arrived on HN, I was unsure about what was appropriate to post. I learned by reading the comments and contributing where I had something useful to share.

Now, after a year or two, I often hear or read a story and think HN has to see this. Then I step back and really weigh whether it fits HN or if I just think it's cool. Often, I end up not submitting. By starting out slow I realized HN isn't popularity contest, but an important community.

That was my experience anyway, YMMV.

Check the source code for news.arc

https://github.com/nex3/arc/blob/master/news.arc

to see what is already done about this. The account submitting so many stories may decide to be more selective after considering the implications of site algorithms.

Can some arc hacker explain the algorithm for the rest of us here. As far as I know the rank of post declines exponentially over time and improves with upvotes. Does the decay also depend upon the rate of submission and the points earned by previous submissions?

If you're referring to the "oversubmitting" algorithm,

a) this appears to be turned off by default (line 1535)

b) if turned on it only applies to new users (lines 1539, 1540)

(Disclaimer: this is the first time I'm looking at the source, please correct me if I'm missing something.)

Ha. This is a classic community phenomenon. You can also notice this in Foursquare, Twitter, and almost every other community based site. I think us as a community need to start encouraging progressive and meaningful arguments, rather than just the top news. I would like to come to HN to have a conversation/ discussion about something new rather than a link to NYT.

One idea would be to create separate sections for discussions, news, blog links, at all.

And how would that discourage the behavior of autosubmitting?

The speed in which new URLs are submitted by the same user should be limited according to their karma. (Assuming users with high points are trustworthy, which is the primary reason this metric is made for.)

New users could start with a 24 hour limit between submissions (86,400 seconds) and get 20 seconds removed for every karma point. Someone with 2000 karma would be able to submit every 12.8 hours. By 4000 you've probably learned that quality beats quantity.

That seems fair, though it might be better to use average story karma. The bots in question probably have a fair amount of karma just from the sheer number of stories submitted.

Multiply the value by their average. It boosts overall submission quality automatically by giving better commenters more chances for a front page article.

edit: Though the value would need to be much lower than 20 per karma with that modifier. My average of 2.73 with 2011 karma would let me post as much as I want.

karma * 3.7 * average would come to around 20,000 for me. That would let me submit an article every 18 hours.

I can't see any obvious disadvantages to having a set limit of, for example, 10 article submissions a day.

Or even two a day.

As someone who almost never submits stuff, I find it really frustrating that the few times I do find something that I think would be great, I see it get knocked down on the "new" page within an hour because there are so many submissions coming in. The rate of new submissions seems like it makes it almost random what just happens to get picked up in the brief window of time and what doesn't.

as with all regulation, don't forget the unintentional side effects: spammers with many sockpuppet accounts will be more present in proportion to typical users that only have one account

I definitely disagree with auto submitting but don't believe that Hacker News should be more locked down.

There is always a lot of complaints about content but there is an inherent system allowing posters to know what the mass wants rather than one commenter and making sure the best content rises.

Don't vote up anything you believe doesn't deserve to be on the front page or flag anything that deserves to be flagged. If there are very strict content guidelines some of the most interesting and obscure content won't be recommended for posting.

Voting or not only addresses the person submitting a ton of content, it does nothing for all the stories prematurely bumped down to page 2 or 3 of the new section where they are effectively dead.

I had the same reaction. HN is already fairly strict with content. If an article has value, it will be on the front page.

Not to mention, most of the stuff I comment on, I want to engage in a decent discussion. If a post doesn't have a lot of comments, chances are the people here don't think it's all that important.

A slightly related previous discussion [1] on the aspect of blindly submitting stories.

Even not doing it blindly (say, you save up links that are good and post them all at peak times for better results) is not in good form either if it is more than say 5 at a time.

[1] http://news.ycombinator.com/item?id=2386443

Agreed; surely there must be a limit on submissions per day?

I am not sure if limiting the submissions per day should be our goal. It's obstruction of freedom of speech- which is do not stand for. Even if we limit the user, how are we making the site productive? The real problem is not that someone has 100K karma, rather, it's that because of this excessive submissions real stories are lost in the mix.

How do we help good stories/discussions come to light? Karma is a vanity metric, let's look beyond it ya'll

Reply on news.ycombinator.com