I am worried about how easy this makes it for one individual to come along after people have stopped paying attention to the thread and massively distort the voting. Like, the combination of a strong upvote and a strong downvote could easily outweigh many people acting in good faith.
Hot take: that isn’t something people couldn’t have done before, if-and-only-if they had a very strong upvote/downvote. I don’t think adding gradations much increases the affordance/incentive for people to do it more than weighted karma did originally. As before, karma is a certain measure of trust.
The problem is that one strong vote can beat many average votes, even from equally high karma users. This kind of outvoting couldn’t have happened before.
That does seem like a potential concern, but I think that’s something I’d file under the “maybe put limits on strong upvotes if this seems to be a problem.”
It doesn’t seem intrinsically bad if someone comes along and strong-downvotes a bunch of stuff after the fact – that may in fact be the system working as intended, especially if the upvotes were small-upvotes by people giving a mild “ah, this seems good”, and the strong-upvoter said “actually, I think people are missing a subtle way in which these comments were wrong, or were probably all groupthink/applause-lighting, and yes these comments should not be as highly upvoted.”
My threat-model mostly fires if someone was doing that all the time or something.
A simple thing you guys could do is freeze the scores on all the LW 1.0 content that got imported and used the old voting scheme. Otherwise the signal from those old votes may gradually get degraded by opinionated archive readers.
I am worried about how easy this makes it for one individual to come along after people have stopped paying attention to the thread and massively distort the voting. Like, the combination of a strong upvote and a strong downvote could easily outweigh many people acting in good faith.
Hot take: that isn’t something people couldn’t have done before, if-and-only-if they had a very strong upvote/downvote. I don’t think adding gradations much increases the affordance/incentive for people to do it more than weighted karma did originally. As before, karma is a certain measure of trust.
The problem is that one strong vote can beat many average votes, even from equally high karma users. This kind of outvoting couldn’t have happened before.
That does seem like a potential concern, but I think that’s something I’d file under the “maybe put limits on strong upvotes if this seems to be a problem.”
It doesn’t seem intrinsically bad if someone comes along and strong-downvotes a bunch of stuff after the fact – that may in fact be the system working as intended, especially if the upvotes were small-upvotes by people giving a mild “ah, this seems good”, and the strong-upvoter said “actually, I think people are missing a subtle way in which these comments were wrong, or were probably all groupthink/applause-lighting, and yes these comments should not be as highly upvoted.”
My threat-model mostly fires if someone was doing that all the time or something.
A simple thing you guys could do is freeze the scores on all the LW 1.0 content that got imported and used the old voting scheme. Otherwise the signal from those old votes may gradually get degraded by opinionated archive readers.