I think flagging for moderator attention is far superior to any automated response. I really like JoshuaZ’s opposite approach, and I have no problem with someone repeatedly encountering a troll, and then going back and downvoting other trolling comments selectively. What I have a problem with is retributive downvoting without regard to the content of the comments being downvoted.
I’d be pretty happy to leave appropriate responses to moderator discussion. My suggestion would be something along the lines of undoing all downvotes that person made within the recent past (past week? moderator discretion seems appropriate), and removing their ability to downvote for some time into the future (a week or so?), long enough for them to cool off consider things from a different perspective. I would hope that the moderators would send a message explaining their action in such cases, even if it was only a description of what behavior was problematic and a boilerplate explanation of how mass downvotes hurt the site.
How do you deal with the induced increase in censorship-trolling noise?
I’m not sure what you mean by this.
With political differences among moderators? With increased moderator fatigue?
I’m hoping that this is a relatively rare occurrence. I’m also hoping that the moderators have handled political differences before, and I don’t see this as more likely to cause political conflict than other moderator actions. That’s part of why I proposed undoing all recent downvotes, not a selected set. Undoing only some seems like too much work, and also too much discretion.
I’d love to see data on downvote patterns; if the problem is far more common than I think it is, I’d be inclined to agree with you.
No, the correct solution space is the realm of automated applied psychology.
Maybe, but I don’t find that to be obvious. I worry that any system with a threshold (too many downvotes triggers an automated response, or something similar) is likely to produce weird distorted incentives. Charging karma for downvotes doesn’t have this problem, but I’m not sure whether I like it or not for other reasons (likely to discourage retribution less than it discourages normal downvoting, see Kindly’s comment).
I’d rather look at individual proposals from both solution spaces before concluding that one is the correct space.
I think flagging for moderator attention is far superior to any automated response. I really like JoshuaZ’s opposite approach, and I have no problem with someone repeatedly encountering a troll, and then going back and downvoting other trolling comments selectively. What I have a problem with is retributive downvoting without regard to the content of the comments being downvoted.
I’d be pretty happy to leave appropriate responses to moderator discussion. My suggestion would be something along the lines of undoing all downvotes that person made within the recent past (past week? moderator discretion seems appropriate), and removing their ability to downvote for some time into the future (a week or so?), long enough for them to cool off consider things from a different perspective. I would hope that the moderators would send a message explaining their action in such cases, even if it was only a description of what behavior was problematic and a boilerplate explanation of how mass downvotes hurt the site.
How do you deal with the induced increase in censorship-trolling noise? With political differences among moderators? With increased moderator fatigue?
No, the correct solution space is the realm of automated applied psychology.
I’m not sure what you mean by this.
I’m hoping that this is a relatively rare occurrence. I’m also hoping that the moderators have handled political differences before, and I don’t see this as more likely to cause political conflict than other moderator actions. That’s part of why I proposed undoing all recent downvotes, not a selected set. Undoing only some seems like too much work, and also too much discretion.
I’d love to see data on downvote patterns; if the problem is far more common than I think it is, I’d be inclined to agree with you.
Maybe, but I don’t find that to be obvious. I worry that any system with a threshold (too many downvotes triggers an automated response, or something similar) is likely to produce weird distorted incentives. Charging karma for downvotes doesn’t have this problem, but I’m not sure whether I like it or not for other reasons (likely to discourage retribution less than it discourages normal downvoting, see Kindly’s comment).
I’d rather look at individual proposals from both solution spaces before concluding that one is the correct space.