Follow-up to: What have you recently tried, and failed at?

Related-to: Challenging the Difficult Sequence

ialdabaoth's post about blockdownvoting and its threads have prompted me to keep an eye on controversial topics and community norms on LessWrong. I noticed some things.

I was motivated: My own postings are also sometimes controversial. I know beforehand which might be (this one possibly). Why do I post them nonetheless? Do I want to wreak havoc? Or do I want to foster productive discussion of unresolved but polarized questions? Or do I want to call in question some point the community may have a blind spot on or possibly has taken something for granted too early.

Of course I see postings that provide objective information. These are almost never downvoted. Some members specialize in these. "Write a lukeprog-style post on X" is an appeal to invest time to provide a benefit (information) for the community. No problem here. Neither for a large body of 'typical' LW posts.

But some postings which I associate with newbies or aspiring rationalists (me included) which are sometimes personal, sometime objective often get a share of downvotes because they don't match some standards or norms. I mostly just don't vote these. I upvote them if I recognize that the poster has made a genuine effort I want to honor.

But progress is made by the discussion of controversial topics, where a consensus or synthesis has not (yet) been reached. Agreed - there are topics that are inherently ambiguous because different people have different values (and I don't mean politics which is abuse of topics for us-them-games). These topics can accumulate a sizeable share of downvotes. But even on these topics agreement should be possible at least on the meta-level of acceptance of the existence of different values. It is not necessary to downvote just because you have another position on this topic than taken or implied by the poster.

Beside these direct on-track controversial topics which are mostly civil with regard to voting (except possibly if a strong stance is taken by some party) we also have another kind of post - or often threads. These are posts and (sub) threads about topics like religion, recreation, politics, status, real life pragmatism, relationships, dealing with newbies and trolls (which actually is a range and precise placement is difficult initially). What is the reason for these seemingly low quality posts (measured by karma)? What keeps those posting these at LessWrong? Who are they?

I think some amount of these topics are a necessary part of a healthy community and somebody has to tend to them. Some are more inclined to do so. Maybe these are housekeeping gnomes who do not really get the reward for their work - as downvoting these topics (to limit them) cannot be differentiated from the need to handle them somehow. If you depreciate these posters you depreciate those topics. Do you want to reduce this area of LW topics which connects LW to real life and help LW to keep alive?

One can wonder why the karma mechanism hasn't driven those posters and the topics away already. As always what gets measured gets optimized. In this case the karma mechanism ensures that no <0 karma poster remains. But as long as you consistently achieve > 50% positive you can stay. This suggests that we should see long-term members at all levels between 50% to 100%. Maybe someone with access to the database could provide a histogram of users by positive percentage. Do we have (long-term) members with near 50% karma?

It can be that there are some social processes at work which reinforce downvoting near 50% (thus skewing the distribution) possibly by those who see karma as a proxy for status and further push those near 50% away (though from a status perspective this in contra-productive because this not only 'punishes' them to stay below but actually 'deletes' them thus letting the punishers sink relatively).

I had a look at the karma positive rate of the recent top contributors and it appears that most long-time members have >85% but indeed few have above 95%. I also had a look at some controversial long-time members and those exist. Even one with 70% positive and >7500 upvotes (which translates into nearly 20000 votes total (up and down); karma formula: total votes = karma/(2*p-1) where p = percentage positive).

So obviously even long term members don't 'achieve' 99%. And this doesn't appear to be the goal. Taking Eliezer Yudkowsky as a role model we see a high karma (I'm surprised how he managed to average 100 karma per day!) and a positive rate of 94% which seems a lot. But if you look over his posts you find a surprising number of controversial posts (e.g. the recent April Fools' Day Confession).

I read this to mean that posting controversial topics is encouraged - if it doesn't get our of hand or into Main.

I read this as an example to go forward and work hard controversy (at the risk of failure). This is in the spirit of What have you recently tried, and failed at? and all the posts in the Challenging the Difficult Sequence. You can only learn from this. Hey. Loosing karma is not the end of the world (only if it falls below 0).

Remember the next time when you see someone with a 500 karma but 60% positive it means 1500 upvotes (and 2500 total) and likely contributions that actually advanced something rather than 'only' disseminating information. And they follow the role models too.

And if you are a low positive poster than take consolation from the absolute votes you got.