Ah, sorry for being unclear. We show the correct net change for each comment/post, until that net-change becomes zero or negative at which point we don’t show it. The total net sum above your karma notifications is the total sum of changes to each individual comment/post that are net-positive.
I’m not sure I get the model of the value of allowing downvotes but hiding them from view. The only reason for karma (afaik) is feedback on whether a small self-selected portion of readers wants more or less of that kind of post/comment. Downplaying the “less, please” part of the signal seems to break a useful quality feedback mechanism.
Basically, based on my and others experience, most people have a bias to weigh negative social feedback much more strongly than positive social feedback. This means in order to get a symmetric response of updating equally much on an upvote as you would on a downvote (such that an upvote plus a downvote on two consecutive days leaves your behavior roughly the same) we have to reduce the salience of negative feedback quite a bit. Different platforms handle this differently. HN for example doesn’t allow the vast majority of users to downvote at all. StackOverflow requires a much higher karma total to downvote than it does to upvote.
My model is that people will notice quite reliably if their stuff gets downvoted, since people check the score of their old content quite frequently. As such, I think in terms of salience we get closer to the right balance if we use karma notifications to highlight positive changes and leave the discovery of negative changes as it was.
Ok, so the mismatch will be that my daily total may be higher than the change to my karma total. No worries. In fact, points don’t matter regardless, so please don’t take my comments as particularly strong recommendations.
I am curious about the tweaks to make the downvotes less visible rather than more fundamental recognition of the assymetry—by making downvotes smaller than the upvotes throughout the system.
Ah, sorry for being unclear. We show the correct net change for each comment/post, until that net-change becomes zero or negative at which point we don’t show it. The total net sum above your karma notifications is the total sum of changes to each individual comment/post that are net-positive.
Basically, based on my and others experience, most people have a bias to weigh negative social feedback much more strongly than positive social feedback. This means in order to get a symmetric response of updating equally much on an upvote as you would on a downvote (such that an upvote plus a downvote on two consecutive days leaves your behavior roughly the same) we have to reduce the salience of negative feedback quite a bit. Different platforms handle this differently. HN for example doesn’t allow the vast majority of users to downvote at all. StackOverflow requires a much higher karma total to downvote than it does to upvote.
My model is that people will notice quite reliably if their stuff gets downvoted, since people check the score of their old content quite frequently. As such, I think in terms of salience we get closer to the right balance if we use karma notifications to highlight positive changes and leave the discovery of negative changes as it was.
I’ve stopped doing this since karma notifications came online.
Elizabeth’s point seems important. What if there was a weekly notification for “here’s the downvotes you received in the past week”?
Ok, so the mismatch will be that my daily total may be higher than the change to my karma total. No worries. In fact, points don’t matter regardless, so please don’t take my comments as particularly strong recommendations.
I am curious about the tweaks to make the downvotes less visible rather than more fundamental recognition of the assymetry—by making downvotes smaller than the upvotes throughout the system.