My “recent karma” (sum of karma values of last 20 comments) is positive
I guess that it counts comment’s karma minus your implicit vote, and that sums to around zero.
Edit: I thought your implicit vote strength was 3 instead of 2. My next best hypothesis is that limits recalculate only when posting comment; will check code now.
(The opacity of this feature is, I feel, a major strike against it.)
Yeah, agree, we really should build a better rate-limiting dashboard and UI for explaining what is going on. This is also making our job as moderators harder! I might try to prioritize it this coming week, though there are many things to do.
How does it make sense to just run the rate limiter robot equally on everyone no matter how old their account and how much total karma they have? It might make sense for new users, as a crude mechanism to make them learn the ropes or find out the forum isn’t a good fit for them. But presumably you want long-term commenters with large net-positive karma staying around and not be annoyed by the site UI by default.
A long-term commenter suddenly spewing actual nonsense comments where rate-limiting does make sense sounds more like an ongoing psychotic break, in which case a human moderator should probably intervene. Alternatively, if they’re getting downvoted a lot it might mean they’re engaged in some actually interesting discussion with lots of disagreement going around, and you should just let them tough it out and treat the votes as a signal for sentiment instead of a signal for signal like you do for the new accounts.
I would also point out a perverse consequence of applying the rate limiter to old high-karma determined commenters: because it takes two to tango, a rate-limiter necessarily applies to the non-rate-limited person almost as much as the rate-limited person...
You know what’s even more annoying than spending some time debating Said Achmiz? Spending time debating him when he pings you exactly once a day for the indefinite future as you both are forced to conduct it in slow-motion molasses. (I expect it is also quite annoying for anyone looking at that page, or the site comments in general.)
I actually happen to prefer it in once a day spurts, and think this generalizes some to others. I don’t think it’s obvious in general which way is better on this dimension though.
Are the dialogues rate limited too? If not, they might be a more suitable medium. They are admittedly harder to branch, but the object-level point of Said vs GSW case has been lost already.
Almost every comment rate limit stricter than “once per hour” is in fact conditional in some way on the user’s karma, and above 500 karma you can’t even be (automatically) restricted to less than one comment per day:
// 3 comments per day rate limits
{
...timeframe('3 Comments per 1 days'),
appliesToOwnPosts: false,
rateLimitType: "newUserDefault",
isActive: user => (user.karma < 5),
rateLimitMessage: `Users with less than 5 karma can write up to 3 comments a day.<br/>${lwDefaultMessage}`,
},
{
...timeframe('3 Comments per 1 days'), // semi-established users can make up to 20 posts/comments without getting upvoted, before hitting a 3/day comment rate limit
appliesToOwnPosts: false,
isActive: (user, features) => (
user.karma < 2000 &&
features.last20Karma < 1
), // requires 1 weak upvote from a 1000+ karma user, or two new user upvotes, but at 2000+ karma I trust you more to go on long conversations
rateLimitMessage: `You've recently posted a lot without getting upvoted. Users are limited to 3 comments/day unless their last ${RECENT_CONTENT_COUNT} posts/comments have at least 2+ net-karma.<br/>${lwDefaultMessage}`,
},
// 1 comment per day rate limits
{
...timeframe('1 Comments per 1 days'),
appliesToOwnPosts: false,
isActive: user => (user.karma < -2),
rateLimitMessage: `Users with less than -2 karma can write up to 1 comment per day.<br/>${lwDefaultMessage}`
},
{
...timeframe('1 Comments per 1 days'),
appliesToOwnPosts: false,
isActive: (user, features) => (
features.last20Karma < -5 &&
features.downvoterCount >= (user.karma < 2000 ? 4 : 7)
), // at 2000+ karma, I think your downvotes are more likely to be from people who disagree with you, rather than from people who think you're a troll
rateLimitMessage: `Users with less than -5 karma on recent posts/comments can write up to 1 comment per day.<br/>${lwDefaultMessage}`
},
// 1 comment per 3 days rate limits
{
...timeframe('1 Comments per 3 days'),
appliesToOwnPosts: false,
isActive: (user, features) => (
user.karma < 500 &&
features.last20Karma < -15 &&
features.downvoterCount >= 5
),
rateLimitMessage: `Users with less than -15 karma on recent posts/comments can write up to 1 comment every 3 days. ${lwDefaultMessage}`
},
// 1 comment per week rate limits
{
...timeframe('1 Comments per 1 weeks'),
appliesToOwnPosts: false,
isActive: (user, features) => (
user.karma < 0 &&
features.last20Karma < -1 &&
features.lastMonthDownvoterCount >= 5 &&
features.lastMonthKarma <= -30
),
// Added as a hedge against someone with positive karma coming back after some period of inactivity and immediately getting into an argument
rateLimitMessage: `Users with -30 or less karma on recent posts/comments can write up to one comment per week. ${lwDefaultMessage}`
},
I think you could make an argument that being rate limited to one comment per day is too strict given its conditions, but I don’t particularly buy this as argument against rate limiting long-term commenters in general.
But presumably you want long-term commenters with large net-positive karma staying around and not be annoyed by the site UI by default.
A substantial design motivation behind the rate limits, beyond throttling newer users who haven’t yet learned the ropes, was to reduce the incidence and blast radius of demon threads. There might be other ways of accomplishing this, but it does require somehow discouraging or preventing users (even older, high-karma users) from contributing to them. (I agree that it’s reasonable to be annoyed by how the rate limits are currently communicated, which is a separate question from being annoyed at the rate limits existing at all.)
Almost every comment rate limit stricter than “once per hour” is in fact conditional in some way on the user’s karma
That’s a heck of an “almost”, given that it excludes the specific example that we are in fact discussing…
I think you could make an argument that being rate limited to one comment per day is too strict given its conditions
I definitely think so.
A substantial design motivation behind the rate limits, beyond throttling newer users who haven’t yet learned the ropes, was to reduce the incidence and blast radius of demon threads.
I’d just like to register my strong objection to the very concept of “demon threads”. I consider it to be both misleading descriptively and detrimental to understanding forum dynamics and to building good discussion environments.
At least in my experience having high karma is very little evidence of being a good commenter. It’s almost exclusively evidence of being a frequent commenter, which also happens to be the people most important to rate limit. We have some rules that apply on recent karma, and they are generally less harsh, and then some much harsher rules based on total karma, so we do take this into account, but overall I think it’s crucial for rate limiting to apply to high karma accounts as well.
We have considered applying rate limits based on high average karma, but haven’t done so because I don’t want to disincentivize productive niche conversations, but it seems like a better start.
Why don’t you make it so karma from posts gives much higher boost to total karma than karma from comments (maybe 5x, possibly even 10x)? This has seemed like an obvious improvement to me for a while. (If you don’t wanna inflate total karma, you could instead do 13x vs. 3x or something.)
Sometimes I wonder whether there should just be nonlinear returns to karma on any item. Like, a 100 karma post should count for much more than 20 5-karma comments / posts.
I feel like it’s a thing where you should use human moderator judgment once the account isn’t new. Figure out how the person is being counterproductive, warn them about it, and if they keep doing the thing, ban them. Ongoing mechanisms like this make sense for something like Reddit where there is basically zero community at this point, but on LW if someone is sufficiently detached from the forum and community that it actually makes sense to apply a mechanical paper cut like the rate limit on them after years of them being on site and accumulating positive karma, they probably shouldn’t be here to begin with.
The basic problem is that it’s not treating the person as a person, like a human moderator actually talking to them and going “hey, we think you’re not helping here, here’s why … in the future could you …” (and then proceeding to a ban if there’s no improvement) would be. People occasionally respond well the moderator feedback, but being hit by the rate limiter robot is pretty likely to piss off just about any invested and competent person and might also make them go “cool, then I’ll treat your thing as less of a community made of people and more like a video game to beat on my end as well”, which makes it less likely for things to be improved in the future.
I think the impartiality really helps. The default thing that happens if we threaten moderation action on any specific individual with a long history on LW is that they feel personally persecuted, complain about it publicly, and try to generally rile up a bunch of social momentum to defend against the prosecution, which then produces a lot of distrust and paranoia and stress for everyone involved.
A nice thing about automatic rate limits is that it’s really transparent we are not doing some kind of strategic purging of dissenters or are trying to find the most politically convenient pretense by which to ban someone, which many people tend to be worried about (I think not without reason given the outside view on the realities of human politics). I think for many people it is much less stressful to interact with a deterministic machine than a human who could potentially be pulling some kind of galaxy brained strategic moves at each step.
they probably shouldn’t be here to begin with.
Many people get triggered for a while. LessWrong commenters change in quality. People get caught in some horrible demon-thread where they feel like they have to keep saying things or lose face. Temporary rate-limits do actually catch many of those cases reasonably well.
The basic problem is that it’s not treating the person as a person, like a human moderator actually talking to them and going “hey, we think you’re not helping here, here’s why … in the future could you …” (and then proceeding to a ban if there’s no improvement) would be.
To be clear, the thing I would do instead of a ban in most cases is an intense rate limit. They just have much better properties in terms of not completely banning certain viewpoints from the site for most cases.
I also think you vastly overestimate our ability to give people constructive feedback. New content review and moderation currently already takes up around one full-time equivalent on average. We don’t have time to do much more of that.
And lastly, I also think you just underestimate the internet’s tendencies to desperately freak out if you ever try to ban anyone. Every time we consider banning any long-term contributor, no matter how obviously harmful they seem for the site, we have dozens of people who otherwise leave good comments come out of the wood work, strong-vote downvote anything even remotely adjacent to the discussion that tries to explain the rationale, complain in like 15 different places, threaten to leave the site, threaten to become enemies of LessWrong forever, and all kinds of things. I think some of that instinct is healthy, but I really think you are vastly vastly underestimating the cost associated with banning a long-time contributor.
Indeed, that is true now, but the total was different when I posted the grandparent comment!
Here’s my suspicion, tell me if it’s right: the rate limit kicks in if the “recent karma” ever dips below the threshold, and then does not get withdrawn if the “recent karma” rises back up over the threshold. True?
I don’t think so. The code to display rate limit warnings gets run anew every time the page loads, and the final check whether a user is allowed to post gets run on comment submission, querying their last 20 comments, with no intermediate caching as far as I can tell:
I just did a quick check by-hand, I got −5 net karma on last 20 comments. [Edit: Ah, Habryka edited in his own recent karma sum, my comment is now redundant.]
Why am I still rate-limited in commenting? None of the criteria described in the post about Automatic Rate Limiting on Less Wrong seem to apply:
My total karma is obviously not less than 5
My “recent karma” (sum of karma values of last 20 comments) is positive
My “last post karma” is positive
My “last month karma” is positive
What gives?
I guess that it counts comment’s karma minus your implicit vote, and that sums to around zero.Edit: I thought your implicit vote strength was 3 instead of 2. My next best hypothesis is that limits recalculate only when posting comment; will check code now.
(FYI no one’s implicit vote strength is 3.)
No, I accounted for that. Anyhow, it seems like the rate limit has ended now, so maybe there’s just some sort of… delay? Batching? Who knows…
(The opacity of this feature is, I feel, a major strike against it.)
Yeah, agree, we really should build a better rate-limiting dashboard and UI for explaining what is going on. This is also making our job as moderators harder! I might try to prioritize it this coming week, though there are many things to do.
Looks like the rate limit is back, even though (I just checked!) my “recent karma” is again positive, etc.
It seems to me that this feature does not, in fact, work in the way that it is described as working, in the post that I linked.
Seems like it’s recent karma from last 20 posts/comments:
Here is the karma for your 20 most recent comments:
Subtracting 2 from all of these numbers for your own small-upvote to get the net karma from them, I get:
3 - 14 - 5 + 1 - 5 - 17 + 3 + 4 + 3 + 1 - 2 + 8 + 2 + 1 + 3 + 11 - 3 + 2 - 2 + 0 = -6So I don’t see any errors in the logic. Seems like it works exactly as advertised in the “comment rate limit” section of the post:
-5 recent karma (4+ downvoters)
1 comment per day
How does it make sense to just run the rate limiter robot equally on everyone no matter how old their account and how much total karma they have? It might make sense for new users, as a crude mechanism to make them learn the ropes or find out the forum isn’t a good fit for them. But presumably you want long-term commenters with large net-positive karma staying around and not be annoyed by the site UI by default.
A long-term commenter suddenly spewing actual nonsense comments where rate-limiting does make sense sounds more like an ongoing psychotic break, in which case a human moderator should probably intervene. Alternatively, if they’re getting downvoted a lot it might mean they’re engaged in some actually interesting discussion with lots of disagreement going around, and you should just let them tough it out and treat the votes as a signal for sentiment instead of a signal for signal like you do for the new accounts.
I would also point out a perverse consequence of applying the rate limiter to old high-karma determined commenters: because it takes two to tango, a rate-limiter necessarily applies to the non-rate-limited person almost as much as the rate-limited person...
You know what’s even more annoying than spending some time debating Said Achmiz? Spending time debating him when he pings you exactly once a day for the indefinite future as you both are forced to conduct it in slow-motion molasses. (I expect it is also quite annoying for anyone looking at that page, or the site comments in general.)
I actually happen to prefer it in once a day spurts, and think this generalizes some to others. I don’t think it’s obvious in general which way is better on this dimension though.
Are the dialogues rate limited too? If not, they might be a more suitable medium. They are admittedly harder to branch, but the object-level point of Said vs GSW case has been lost already.
Almost every comment rate limit stricter than “once per hour” is in fact conditional in some way on the user’s karma, and above 500 karma you can’t even be (automatically) restricted to less than one comment per day:
https://github.com/ForumMagnum/ForumMagnum/blob/master/packages/lesswrong/lib/rateLimits/constants.ts#L108
I think you could make an argument that being rate limited to one comment per day is too strict given its conditions, but I don’t particularly buy this as argument against rate limiting long-term commenters in general.
A substantial design motivation behind the rate limits, beyond throttling newer users who haven’t yet learned the ropes, was to reduce the incidence and blast radius of demon threads. There might be other ways of accomplishing this, but it does require somehow discouraging or preventing users (even older, high-karma users) from contributing to them. (I agree that it’s reasonable to be annoyed by how the rate limits are currently communicated, which is a separate question from being annoyed at the rate limits existing at all.)
That’s a heck of an “almost”, given that it excludes the specific example that we are in fact discussing…
I definitely think so.
I’d just like to register my strong objection to the very concept of “demon threads”. I consider it to be both misleading descriptively and detrimental to understanding forum dynamics and to building good discussion environments.
At least in my experience having high karma is very little evidence of being a good commenter. It’s almost exclusively evidence of being a frequent commenter, which also happens to be the people most important to rate limit. We have some rules that apply on recent karma, and they are generally less harsh, and then some much harsher rules based on total karma, so we do take this into account, but overall I think it’s crucial for rate limiting to apply to high karma accounts as well.
We have considered applying rate limits based on high average karma, but haven’t done so because I don’t want to disincentivize productive niche conversations, but it seems like a better start.
Why don’t you make it so karma from posts gives much higher boost to total karma than karma from comments (maybe 5x, possibly even 10x)? This has seemed like an obvious improvement to me for a while. (If you don’t wanna inflate total karma, you could instead do 13x vs. 3x or something.)
Something a bit awkward about this is it incentivizes making long comments and quick takes into low-effort posts.
Sometimes I wonder whether there should just be nonlinear returns to karma on any item. Like, a 100 karma post should count for much more than 20 5-karma comments / posts.
I feel like it’s a thing where you should use human moderator judgment once the account isn’t new. Figure out how the person is being counterproductive, warn them about it, and if they keep doing the thing, ban them. Ongoing mechanisms like this make sense for something like Reddit where there is basically zero community at this point, but on LW if someone is sufficiently detached from the forum and community that it actually makes sense to apply a mechanical paper cut like the rate limit on them after years of them being on site and accumulating positive karma, they probably shouldn’t be here to begin with.
The basic problem is that it’s not treating the person as a person, like a human moderator actually talking to them and going “hey, we think you’re not helping here, here’s why … in the future could you …” (and then proceeding to a ban if there’s no improvement) would be. People occasionally respond well the moderator feedback, but being hit by the rate limiter robot is pretty likely to piss off just about any invested and competent person and might also make them go “cool, then I’ll treat your thing as less of a community made of people and more like a video game to beat on my end as well”, which makes it less likely for things to be improved in the future.
I think the impartiality really helps. The default thing that happens if we threaten moderation action on any specific individual with a long history on LW is that they feel personally persecuted, complain about it publicly, and try to generally rile up a bunch of social momentum to defend against the prosecution, which then produces a lot of distrust and paranoia and stress for everyone involved.
A nice thing about automatic rate limits is that it’s really transparent we are not doing some kind of strategic purging of dissenters or are trying to find the most politically convenient pretense by which to ban someone, which many people tend to be worried about (I think not without reason given the outside view on the realities of human politics). I think for many people it is much less stressful to interact with a deterministic machine than a human who could potentially be pulling some kind of galaxy brained strategic moves at each step.
Many people get triggered for a while. LessWrong commenters change in quality. People get caught in some horrible demon-thread where they feel like they have to keep saying things or lose face. Temporary rate-limits do actually catch many of those cases reasonably well.
To be clear, the thing I would do instead of a ban in most cases is an intense rate limit. They just have much better properties in terms of not completely banning certain viewpoints from the site for most cases.
I also think you vastly overestimate our ability to give people constructive feedback. New content review and moderation currently already takes up around one full-time equivalent on average. We don’t have time to do much more of that.
And lastly, I also think you just underestimate the internet’s tendencies to desperately freak out if you ever try to ban anyone. Every time we consider banning any long-term contributor, no matter how obviously harmful they seem for the site, we have dozens of people who otherwise leave good comments come out of the wood work, strong-vote downvote anything even remotely adjacent to the discussion that tries to explain the rationale, complain in like 15 different places, threaten to leave the site, threaten to become enemies of LessWrong forever, and all kinds of things. I think some of that instinct is healthy, but I really think you are vastly vastly underestimating the cost associated with banning a long-time contributor.
Indeed, that is true now, but the total was different when I posted the grandparent comment!
Here’s my suspicion, tell me if it’s right: the rate limit kicks in if the “recent karma” ever dips below the threshold, and then does not get withdrawn if the “recent karma” rises back up over the threshold. True?
I don’t think so. The code to display rate limit warnings gets run anew every time the page loads, and the final check whether a user is allowed to post gets run on comment submission, querying their last 20 comments, with no intermediate caching as far as I can tell:
I see… well, this really just highlights the aforementioned fact that the UI for this stuff really needs work, which I guess you already know.
I just did a quick check by-hand, I got −5 net karma on last 20 comments. [Edit: Ah, Habryka edited in his own recent karma sum, my comment is now redundant.]