I log in this morning on a whim, and notice I have -15 karma.  I dig around for a bit and find this:

http://lesswrong.com/lw/nsm/open_thread_jul_25_jul_31_2016/ddjm

To be clear, that's a block of four comments, each at -10, for no apparent obvious good reason other than eugine nier has a vendetta against Elo.  I've apparently just been hit as splash damage, since I had the gall to try posting on an Elo comment thread.

I dig a little more, and I find this:

http://lesswrong.com/user/Elo/overview/

That's Elo's page, and I see a pile of discussion-grade posts that are all bulk downvoted below visibility, again for no apparent obvious good reason.

I find myself incredibly disincentivized to post or comment as a result of this.  My feeble amount of karma has taken literally years to build up, and to see sizable fractions of it wiped out any time I step on a eugine nier landmine is bullshit.  Sure, it's silly to value karma, but I value it anyway and if a year of incidental effort can be burned in two days because one guy wants to be an asshole to me, then I'm done here.

This has been going on for months.  Years even.

I understand the staff of LW are pressed for time.  I understand nobody understands how the code works.  I understand that maintaining the site is hard. However, reality is that which does not go away when we close our eyes, and reality does not care:  no matter how difficult the problems are, the fact remains that this sort of thing is abusive and it is actively driving people off the site.

If you value LW, fix this.  Use the force harder, site owners.

On the other hand, if you want LW to turn into another Kuro5hin, then keep doing what you're doing.

Prediction:  50% odds this post will be downvoted below visibility within two days due to eugine, and will basically disappear without trace.

Prediction:  if this isn't dealt with soon, 50% odds I'll stop visiting LW completely other than as an article archive by year end, because there's no goddamned point in trying to use the discussion system.

New Comment
15 comments, sorted by Click to highlight new comments since:

There is some progress on this, but overall changes to the codebase have been slow-going. I've pushed for doing things the right way, even if it takes longer, rather than quicker attempts that are less likely to work.

Of the three pieces that I think are useful, one has been implemented, another written but not yet merged (it needs a bit more work), and a third has not yet been written. If you'd like to contribute coding effort, this issue is my highest priority of the open issues with no pull requests and seems like it should be fairly simple to me.

I definitely think we should ban downvotes, at least temporarily. Also, it is clear that Eugene has an army of automated sockpuppet accounts that are repeatedly downvoting this entire thread. At a later stage something should be done about this, for example limiting the ability of people to spam accounts (e.g. the google "I'm not a robot" button) and limiting the ability of new accounts to downvote. Perhaps only accounts that have made a discussion post with a few upvotes should be allowed to downvote at all, and even then with limits per week and per user to be downvoted, and also perhaps there should be a per-user limit on downvoting of sufficiently old comments, so that even with an army of bots you cannot mass downvote people by attacking all their old content.

Overall it seems we have given out downvoting privileges like candy and now we are reaping the consequences...

As a general rule, if something is a problem, the solution needs to deal with the problem, not with its proxy. The problem is "army of sockpuppet accounts", not "downvotes" per se, therefore a successful solution must somehow address the sockpuppetting itself.

I don't want to give Eugine new ideas, but banning the downvotes would probably just make him change strategy. I can imagine two powerful attack strategies that would work if (a) downvotes are banned, or even (b) all votes are banned.

The successful solution must:

  • identify the sockpuppets; and
  • remove the sockpuppets or otherwise render them harmless

I think there are two essential approaches to this:

  • blacklisting = suspected sockpuppets (detected e.g. by their IP address or behavior) are removed and their votes reversed; or
  • whitelisting = only a set of "trusted users" can vote

These two can come with different flavors and combinations. For example, we could have an invisible whitelist of trusted users, in general treat votes by trusted and untrusted voters equally, but also provide an automatical warning to the moderators if the votes given by trusted vs untrusted voters differ dramatically (for example, if 9 of 10 trusted users upvoted a comment, but 30 of 40 untrusted users downvoted it). This is just an example; it could be made more sophisticated, but that would require more programming resources and computing power.

Perhaps only accounts that have made a discussion post with a few upvotes should be allowed to downvote at all

Eugine would simply upvote posts made by his sockpuppets by his other sockpuppets. In the very best case, this would force him to write one half-decent post per sockpuppet.

limits per week and per user to be downvoted

Limits per sockpuppet = more sockpuppets.

perhaps there should be a per-user limit on downvoting of sufficiently old comments

Maybe downvoting of sufficiently old comments should be limited in general, not just per user. Just like on Reddit you cannot vote on too old stuff. (Question is, how old is "sufficiently old"; on Reddit that means a few months.)

Dealing with bots is hard, banning downvotes is easy. Ideally, with infinite resources, the bots would be eliminated.

In the very best case, this would force him to write one half-decent post per sockpuppet.

that's actually a huge inconvenience for him - writing a good post, using the puppet to mass downvote, then an hour later the puppet is caught because it is mass downvoting and the votes are reverted.

Limits per sockpuppet = more sockpuppets

presumably there is a cost per puppet. Combining this with my above suggestion would mean more articles for him to write...

At this point, if we can't undo mass-voting, it would be better to just disable downvotes or voting entirely.

It's not the only cause of LW's decline, but it does contribute.

[-]9eB1250

LessWrong is no longer even large or active enough for downvoting to be necessary. The activity of posts here is similar to Usenet, which had no moderation.

Usenet, which had no moderation

Some newsgroups were / are moderated. Most of the sci. soc.

Also newsreaders had killfiles so you could ignore people.

Same problems for decades: 4chan -> moderation -> echo-chamber

I wouldn't have thought of this before reading Dagon's comment, but I agree that 'disable downvoting completely' seems like a solution with few or no drawbacks at this point. Certainly I think it would be a massive improvement over the status quo, at very small cost (even if the result is just hacked in, it should be pretty straightforward to hack in.)

I think some people don't see this as a big deal because they generally ignore karma, or they just instruct others to ignore karma, but some of us (perhaps from a history of reading reddit, who knows) have it ingrained enough that there's at least a real "trivial inconvenience" associated with ignoring that signal, and there are associated negative feelings towards the site which are growing over time at a System 1 level.

Comment 1: If anyone wants to comment or reply here, but can’t afford the karma hit, feel free to PM me and I’ll comment for you without listing your name. I have 169 karma to burn (97% positive!), from comments going back to Feb 2015. However, I’ve wanted to update to a different username, so I don’t mind destroying this one.


Comment 2: It might be wise not to discuss tactics where eugine can read it. (Also, causing lots of discussion might be his goal, but so far we haven’t talked about it much and it’s just been a background annoyance.)

Is there interest in a skype call or some other private forum to discuss possible solutions?

[pollid:1160]

Could the administrators create protected accounts that could be up-voted but not down-voted?

They probably could, but that ends up being a very toil-based setup as new targets are found and selected. I wouldn't consider this anything more than a short term stopgap.

As an example, even if Elo was protected, it's pretty clear the eugine is willing to downvote anyone who comments on Elo material.

Also, if some comments by Elo are good and some are bad, we would lose the ability to organically downvote the bad ones. (Maybe the protected users should still be able to downvote each other?)

Thank you for posting this. I agree that LW is a potentially valuable site, and I'm very angry that one person has been able to nearly wreck it.

MarsColony_in10years, thank you for working on a venue. I recommend text for easier reference rather than a phone call system, though both text and voice might be a good idea if people want voice for hashing out ideas.

Vaniver, thank you for working on this and posting about it. I didn't know progress was happening at all.

I agree with Dagon and the others who said this that if it is going to take a significant amount of time to fix what needs to be fixed, it would be better to disable downvoting completely, at least temporarily, if that can be done easily. Of course it would be better if people didn't care so much about karma. But that is not a thing which is easily changed, and we can see from this post that people really do quit for that reason.

yeah, we definitely need to ban downvotes. This is ridiculous.