site banner

Add Agree/Disagree Voting

The move from Reddit to a dedicated forum is a huge opportunity to mix things up. We should take advantage. Never let a crisis go to waste, etc.

One thing I would suggest (if technical limitations allow) would be the addition of a two-tiered voting system, somewhat like what LessWrong has implemented, where users can vote both on the quality of a post, and separately on whether or not they agree with it. I think this could have really positive effects for the kind of community and discussion the Motte was created to promote. The Motte's raison d'etre is to promote discussion and debate with people you disagree with. Separating voting on quality from voting on agreement would promote that goal in a couple different ways. Fundamentally, there is a tension between upvoting a post you think is well-done, and downvoting that same post because you disagree with its content. I think the Motte wants to be a place that encourages outsider or minority views, and separating the "quality" vote from the "agreement" vote would help promote this. From what I have noticed in this community, despite our commitments to encouraging debate and discussion with people you disagree with, posts coming from a more liberal/left-wing/social justice/woke viewpoint tend to get downvoted, even when their quality is equivalent or superior to other posts.

I'll also quote from the reasons given on the above LessWrong post about this feature, because I think the reasons given are good ones.:

I personally feel much more comfortable upvoting good comments that I disagree with or whose truth value I am highly uncertain about, because I don’t feel that my vote will be mistaken as setting the social reality of what is true.

I also feel very comfortable strong-agreeing with things while not up/downvoting on them, so as to indicate which side of an argument seems true to me without my voting being read as “this person gets to keep accruing more and more social status for just repeating a common position at length”.

Similarly to the first bullet, I think that many writers have interesting and valuable ideas but whose truth-value I am quite unsure about or even disagree with. This split allows voters to repeatedly signal that a given writer's comments are of high value, without building a false-consensus that LessWrong has high confidence that the ideas are true. (For example, many people have incompatible but valuable ideas about how AGI development will go, and I want authors to get lots of karma and visibility for excellent contributions without this ambiguity.)

There are many comments I think are bad but am averse to downvoting, because I feel that it is ambiguous whether the person is being downvoted because everyone thinks their take is unfashionable or whether it's because the person is wasting the commons with their behavior (e.g. belittling, starting bravery debates, not doing basic reading comprehension, etc). With this split I feel more comfortable downvoting bad comments without worrying that everyone else who states the position will worry if they'll also be downvoted.

I have seen some comments that previously would have been "downvoted to hell" are now on positive karma, and are instead "disagreed to hell". I won't point them out to avoid focusing on individuals, but this seems like an obvious improvement in communication ability.

Would this be a doable change? And would it be a good one? I am strongly in favor, but open to reasons why I'm wrong.

21
Jump in the discussion.

No email address required.

Theres so much speculation about whats the best voting system ranging from "no voting allowed" to "vote with emoji that can have any abstract meaning you want" or "separate vote for each fraction of the post you want to highlight"

But we could go with the archipelago system of just adding every voting system as an option you can select when making a post, and well just see what happens

There might be a different 'best' voting system for each unique combination of topics and audiences.

I know this is going to be the most expensive way of doing this and if theres not enough interest no one will want to program it, just wanted to throw this idea out

On the topic of changing voting, one thing I've liked about some forums (eg. EDAboard) is that it lists who has liked a given post. I think either public display or private display (only the commenter knows who reacted how to their post) would be beneficial to building a rapport in the community. I lean away from public display because I think it can tend to be too consensus-building and people can hesitate to react the way they want because of how it looks eg. with Facebook. Private however could at least give you an idea of who is reacting which ways so that you can engage with them accordingly and gauge the value of their reactions for yourself.

Get rid of the voting system altogether.

Rational as we think we are, we are still monkeys and big number better than small number. Discussions where you don't know if your position is popular or not via proxy of similar comments are far more natural and prevents sliding into lazy thinking. The proxy of consensus is a bug not a feature in this forum!

Voting systems are great for restaurant reviews. Not so great for anything not concerning matters of taste, We want to eliminate that variable. Imagine you had a bunch of studies to include in a meta analysis but they came with aggregated votes of what other scientists thought of those studies.. Terrible idea eh?

The discussions here on the motte are more akin to choosing the best studies than the best restaurant.

You need voting for engagement and participation so people get a dopamine effect of either anger or everyone loves me.

And I’m someone who never votes. But not getting a dopamine effect might push people to other venues.

Counterpoint, I check upvotes on my own posts some time afterwards and use them to retroactively approximate how much minor engagement they received: people who read the post and liked it but didn't feel like they needed to comment/respond. There's a difference between a post with 2 karma and 0 replies versus one with 8 karma and 0 replies. Assuming Although technically this can't distinguish between a pot with 8 upvotes minus 6 downvotes, and one with just 2 upvotes, but assuming I haven't written anything especially controversial then downvote proportion is probably low and it gives a proxy.

So I can in some sense using this as feedback to tailor future posts and determine which topics people do and don't want to hear about, or methods of writing, or length, or whatever. On the other hand, this might just be my monkey brain trying to get positive feedback for agreeing with consensus. But I think there's something valuable here that would be lost, even if something else would be gained. A voting system that could only be seen by the commenter would give this feedback without influencing public opinions, but that seems kind of silly and I'm not sure voters would care enough to use it.

The question being should a discussion forum optimize for bidirectional communication or unidirectional communication.

There have been times I had nothing to add to a discussion but did enjoy reading a post. And a voting system lets the writer know that he has readers, which is fair. But as I said it does have its tradeoffs. And one can make the argument that;

  1. Readers benefit from posts and discussions. Making discussions better at the cost of posts might not be a net loss to readers.

  2. Why do we want to encourage posts that get no replies? It's a discussion forum, not a soapbox.

Which demographic the motte should optimize for..? Idk. Given the mods have a lot more control now, they can probably run A/B tests and surveys to get feedback on that.

I don't necessarily think we need two parallel voting systems, because we'll probably just have the double downvoting issue as people have pointed out. However, I think leaving the current voting system as an agree/disagree is fine, but there should be an option to double upvote posts that are particularly high-quality. Basically, it's like a Quality Contribution nomination, but more visible and public. Having additional voting is fine, but the added one should only be for upvotes IMO, sort of like deltas in the CMV subreddit, except in this case for good posts instead of just ones that "change minds". Maybe (like deltas) there ought to be a slightly janky way of giving them so people don't distribute them like candy.

I think this will lead to most people just downvoting it twice without much consideration as to the nuance between disagreement and well-argued disagreement. Well-argued but unpopular opinions can still do well here, so it's not like group-think is necessarily a problem that needs a new voting system.

Worst-case scenario that means it retains the current status quo, because the way LessWrong implemented it the agree/disagree vote has no effect on making a post more/less visible.

I'd rather completely get rid of user feedback and let the text stand for itself. Internet points and their consequences have been a disaster for online discourse.

What started as a benign experiment in lazy community moderation has amplified the worst aspects of tribalism and made unpopular opinions, either impossible to express or lauded in echo chambers. Even as they are some of the most important parts of discourse.

This community is great because people don't play into that particular system too much and are ready to hear things they don't like and not just try to hide them by any means, but I think encouraging out basest instincts in that manner is terrible.

It is just me, but I don't even want to know if people agree with something or not outside of a formal poll. It's poison to the fair evaluation of the ideas for themselves in my opinion.

I agreed with your sentiment, and then out of force of habbit gave you an internet point.

That says a lot about society.

How about Favorites, Engaging, and Rulebreaking? This would be <3, 💡, and a judge’s gavel as symbols.

Voting a post as engaging pushes it higher in the default comment hierarchy, but we can keep the score hidden, and the default effect will be to increase the position in the comment hierarchy. For instance if a post gets 30 direct child comments, the engaging+ will boost the position of the most engaging comment. “Engaging” is the right way to think about it, because it could be low effort but still very engaging, and high effort but not very engaging. The most engaging three comments in the chain could have a unique identifier so that a casual reader with limited time can quickly read through “generally good” comments. To make up for new comments not being defaulted, maybe each post can have a New: X toggle, so you can only see New or only see Engaging. I read most comment, but busier users certainly do not, and there’s no reason to make it so difficult for busy users to find the most engaging comments.

Favoriting a comment will give the user a notification that it has been favorited. This is positive reinforcement without positive punishment, and the science shows that positive punishment reduces engagement. Downvotes are a form of punishment. For added benefit, the user can get a notification that they received +10 or +25 favorites in intervals, along with “milestone” imagery and so on. This is simply to encourage participation, nothing else.

For comments that truly ought to be punished, this seems to overlap with comments that ought to be modded. In which case there should be a unique button for that distinct from Engaging/Favorites, maybe the button looks like a judge’s gavel. This would be hidden from users but seen by mods. This makes sense because most of us are probably reading bad comments anyway, so the downvote system doesn’t really do its intended thing of orienting the best comments. Instead of something obviously punishing like being at -80 (like when I made fun of Magnus Carlsen on the chess subreddit yesterday), it’s sufficient to leave the comment at the bottom of the comment hierarchy, and if it should be modded there’s a mechanism for that.

From my experience reading on LessWrong, I think this does actually work relatively well. It isn't perfect at avoiding the 'I think they are wrong/misinterpreting/whatever, thus downvote', but I've personally found that it helps me.

However. I'm less certain how useful it is on a relatively more 'adversarial'/'political' site like TheMotte, I'd expect it to not work as well but probably still work some.

Why wouldn't it work as well here? People would downvote political enemies to suppress their ideas?

I actually didn't put that into words in my own head, so that's embarrassing.

I think the primary thing driving that intuition is that on political discussion sites, things tend to be two opposite extremes going at it. While, on LessWrong, there's a good amount of 'yes I see what you mean, however I think you are missing this important bit', which works nicer for agree/disagree. While on TheMotte, there seems to be a larger degree of 'here is my view on the issue', which still works but is often less self-contained? I'm not entirely sure I'm getting my intuition into words nicely.

Agreed, upvoted, considered high-quality. But what is «Quality»? (We could add any number of voting dimensions really, while we're at it. Even nonsense ones, «bananas» or something, or little hearts and diamonds like on gaudy old forums. Zorba, don't take this seriously).

I admit being that guy who frequently uses downvote button as a disagreement button.

That's because this place does not even more effort and eloquence in peddling bad takes. Reddit maybe (let normies cut their teeth on mock debates), not The Motte. Basic literacy and politeness are requirements for participation, not virtues. This is not a Powerful Takes Magazine to boast of your VQ and radicalism; a well-polished turd should be getting two downvotes at once. And if a take is good (by my account), then I usually don't disagree all that much, it's just a... respectable difference in opinion.

But on the other hand, a genuine high-quality post may not get an upvote if I disagree with it and wish its ideas would have less influence acquired by peer pressure. If there was a button to signal instead «this is a substantial effort on a crucial topic, this deserves engagement, but ultimately seems wrong-headed and worthy of a comparably well-written takedown» – then there'd be many posts upvoted on this dimension by me. And it's equivalent to rating the post [-1; 1].

And perhaps my own posts would come to score more on this metric. If nothing else, it's a interesting hypothesis to test, personally.

The best online discussions I've had over the 20+ years I've been having them have almost all been in old phpBB-type forums or (further back) on Usenet, where there were no scoring systems. I don't believe this is a coincidence. Even though rationally people shouldn't care that much about fake Internet points, they do, and there's a tendency to pander to an understood consensus, either by not raising arguments you think will be unpopular in the first place, or by prematurely terminating exchanges where you've discovered the consensus opposes you.

So my preference would be to simply eliminate voting, or, failing that, to hide comment scores from non-moderators, including from comment authors.

This seems right on one dimension. I agree that it does feel like older-style forums, without voting, seemed to have better discussions. But when I read sites like SSC that purposely hide comment scores, I find it annoying not to be able to sort by "best" and find myself missing upvotes, without really feeling the conversation is all that better without them. Maybe this is just a function of the internet getting worse in general...

I personally would like to see some kind of emoji reaction system like discord. A clown face, smile, or confused face can all be quite descriptive to gauge public opinion on a post. I don't downvote at all, unless something is especially poorly written.

I actually do have this on my wishlist. The tricky part is that I want to make sure it's hard to brigade with it; I don't want, like, "upvote" or "downvote" equivalents, nothing that says "you are right" or "you are an idiot".

I'm really tempted to set up some kind of paid emoji-submission system similar to Ye Olde Something Awful, and just cull it mercilessly if things turn into problems.

All of this is long after Improve The Comment Page Layout, though.

Could you really imagine, say, a well done argument for homeopathy? I could imagine an argument for homeopathy that deceives in a well done way, but I can't imagine an honest argument for homeopathy that is well done.

If I have position X, that implies that I think that opponents of X have bad evidence, are making logical fallacies, etc. If I didn't think that, I wouldn't have position X. Of course, in everyday life where a lot of people believe things for social reasons, they may believe X without considering evidence or arguments at all. I can imagine such a person thinking an opposing argument is well done. But I find it hard to imagine that to anyone who arrives at their positions by rational thinking, they could really find an opposing argument well done.

There may be an edge case where someone uses premises I disagree with but they argue well based on these premises, but in practice nobody seems to argue that way. Even abortion opponents don't say "I am trying to convince you that abortion is wrong, based on the premise that a fetus is a human being". They try to convince other people using those people's premises, regardless of whether they share them themselves.

Furthermore, quite aside from all that, the problem with any measure to prevent the use of votes as weapons is that votes work as weapons. If your system hides posts or otherwise does things weapon-like based on votes, you'll never solve the problem.

Meta irony: i disagree with your point, but find it well argued.

I agree that this doesn't solve the problem of people voting "politically," as it were, but I think it might mitigate it slightly.

As to your other point , I strongly disagree. Rationality can only work once certain premises have been accepted. There is no rational way to choose what premises you start an argument with. In the abortion context, for example, if someone starts from the premise that it is always wrong to take an innocent human life, no matter how much suffering it might experience or cause, they'll reach a certain set of conclusions, and if someone starts from a utilitarian set of premises, they'll reach another. Yes, you can argue about the rational basis for premises to some extent, but at a certain point you just hit intuition. Thus, two equally rational people could reach wildly different conclusions simply because they have different intuitions about the premises of the argument.

Rationality can only work once certain premises have been accepted.

But that's not how arguments to convince other people work. If you want to convince other people, you either have to go by their premises (which leads to the problem I described), or you need to sneak in your own premises (which makes it inherently a low quality argument because of the sneakiness). Just openly arguing based on premises that another person doesn't share won't convince him.

Nobody says "Well, I'm going to convince you of homeopathy. You just need to assume that molecular patterns are a thing...."

In the abortion context, for example, if someone starts from the premise that it is always wrong to take an innocent human life

That doesn't even work. You can think it's wrong to take an innocent human life, but disagree that a fetus is an innocent human life (or a human life at all).

Your claim was that you "find it hard to imagine that to anyone who arrives at their positions by rational thinking . . . could really find an opposing argument well done."

All I'm saying is that a rational person could easily find an opposing argument "well done" if their only problem with the argument was that it began from premises derived from intuitions they don't share.

Just about nobody will make such an argument openly, because it wouldn't be convincing. If they make it sneakily, it's not a good argument because it's sneaky. So while this is possible in theory given a certain kind of mistake, it's a rare edge case.

deleted

If I have position X, that implies that I think that opponents of X have bad evidence, are making logical fallacies, etc. If I didn't think that, I wouldn't have position X. Of course, in everyday life where a lot of people believe things for social reasons, they may believe X without considering evidence or arguments at all. I can imagine such a person thinking an opposing argument is well done. But I find it hard to imagine that to anyone who arrives at their positions by rational thinking, they could really find an opposing argument well done.

But belief (or at least confidence) isn't an all-or-nothing thing, surely. Sure, based on the information I have now and the chains of logic I have used, I have reached my own conclusions, but I know I don't know everything, and I don't reason perfectly, so I mustn't come into every situation wholly closed-minded.

I can imagine an argument that I think is a good one that still doesn't convince me: it would be one that shifts the balance of probabilities in their direction, but still not enough.

I'm afraid I don't see what the value of this place would be if one doesn't think that they could ever really find an opposing argument well-done. Perhaps it would seem a useful resource to keep abreast of the Fiendish Tricks Of The Enemy, but surely there are lots of better resources for that.

Every internet community I've been in with user comment feedback navel gazes about this. It's amounted to nothing. So LessWrong has come up with their own version of Slashdot's "Interesting/Insightful" voting, with even more galaxy-brained schemes in the comments. Other social media experiments suggest you can't hack your way around the human psychology of using feedback UI as an "agree" button. LWers will just now "up-right" vote or "left-down" vote, with two-axis voting forcing them to click a second time.

It's not worth the energy to think or argue about this, let alone design or code it, unless another community invents a voting scheme that brings home the bacon in terms of discourse in a huge, obvious way.

If you have fun talking about ideas that would take hundreds of coding hours with no promise of success for a tiny internet subforum, go ahead. But I strongly suggest anyone who might be tempted to waste otherwise productive manhours on it close this thread and go about their day.

Why do you say it hasn't worked for LessWrong?

How are LessWrong comments better than they were two months ago? Is there a good library you can point to of comment threads where the top karma-sorted post is fiercely downvoted? Or where highly upvoted, low effort comments are near the bottom because they have 0 karma? That's a good minimum standard for saying the change "worked".

Rediquette is stupid, always has been always will be. The downvote button is used to signal disagreement, or perhaps even more accurately: "I wish this person to experience negative feelings." There's no accountability, so the "disagree" and "quality" tools will just be used as "I don't like this person" buttons.

If y'all want to do a goofy voting system, it should be one with stakes for the person voting. There needs to be a cost to make it worth the effort.

Vote economy with new points distributed periodically and bonuses to AAQC winners?

(Something something quadratic voting).

For this to work, people commenting would have to use both votes together: 'I disagree with this post (so downvote) but I think it is well-argued (so upvote)'.

In practice, I think it would happen that people would use "upvote/downvote for agree/disagree" or "upvote/downvote for quality", but not both, so we'd end up once again with the problem of "was this post downvoted because people think the idea is stupid/they dislike the person posting, or was it downvoted because it was a good idea but poorly expressed?"

The trouble with all voting systems is that they do end up as popularity contests, and on here the danger is popularity of an idea, more than popularity of a single person. There is already the history of complaints that left-wing ideas get dogpiled, it's even mentioned within the original post:

posts coming from a more liberal/left-wing/social justice/woke viewpoint tend to get downvoted, even when their quality is equivalent or superior to other posts.

So we'd end up with the worst of both worlds: people arguing that the large amount of upvotes for right-wing idea means that everyone on here agrees with the content (not merely the presentation) and that this is indeed unfriendly and hostile to left-wing/liberal ideas. How many critics are going to carefully parse out "this post on the Holocaust got 200 upvotes for quality, but 500 downvotes for disagreement"? How many of them will take it as read that "upvoted for quality" means "agree with it"? End result being yet more coverage of "over there a post on Holocaust denial got 200 upvotes, we told you they were all far-right and now this proves it!"

In practice, I think it would happen that people would use "upvote/downvote for agree/disagree" or "upvote/downvote for quality", but not both, so we'd end up once again with the problem of "was this post downvoted because people think the idea is stupid/they dislike the person posting, or was it downvoted because it was a good idea but poorly expressed?"

I agree for sure, in general, but I also think that a population like this one might be about as resistant to such behavior as a population can be. (Is that enough? I don't know!)

I don't think we need interaction buttons beyond "AAQC" and "Hey mods!". The thread is already sorted chronologically, so upvotes don't matter that much anyway.

How do you propose this factor in to post ranking? Posts are sorted by new by default, but the top, bottom and hot sort are all still there. Would you only use quality? Just add quality and agreement? or create a different sort for each metric? Either way I like the idea of adding complexity to the vote system, as it is the primary way a poster can glean information about the community reaction to the post, and complexity means more information to draw from.

Well in my ideal world you'd be able to sort by top - quality and top - agreement, but if they wanted to keep it simple they could just use the quality metric. I think that's what LessWrong does.

Completely agree about the complexity angle.

I'll voice my thoughts here, I guess:

I kinda wish there was a feature where you could upvote/approve individual parts/sentences of a post. So often, I see a post I would have updooted if not for a part I either disagreed with or thought was too spicy. The more upvoted a section is, the more highlighted it'll become. I don't know if this is possible with current tech.

There's one big issue I have with this, and this is going to sound dumb, but bear with me, this is a gamedev's perspective:

Easy contribution options are a great way to get people to feel ownership and buyin of a community, and that keeps people coming back.

This is entirely an argument around trying to attract people. But I think that's important - hell, I think it may be the single most important thing that we need to deal with in the next few months. Every bit of complexity that's added drives people away. That isn't always bad, but we have to be super-careful about driving the right people away. In an ideal world I want to attract 100% of the people who are interested in complicated and difficult debate, and 0% of the people who aren't, and I'm not convinced that a more complicated front-page voting system is going to cleave those percentages in the way we want.

This is not to say the current voting system is here to stay. I do have quite a few ideas on ways to improve it, taking inspiration from things like Discord and Slashdot. But right now that's just on the backburner, I'm afraid; we have somehow avoided the frying pan but are now on a slow conveyor to the trash can, and that's what I gotta deal with first.