3 ms·

There are plenty of questions where the most upvoted answers is wrong or out of date. I learned stop looking at the highest voted answer and instead read the an

There are plenty of questions where the most upvoted answers is wrong or out of date. I learned stop looking at the highest voted answer and instead read the answers themselves. Hiding the scores helps with that. I wasn't lead to an answer based on popularity. Maybe that's only true for S.O. because there's rarely more than a few answers. For a sight like HN/Reddit I think it might be ok if there was an upvote, no downvote, but the scores were not visible. You'd still get sorting of comments but you wouldn't get the "think like the rest of us or will downvote you" part of downvotes and you (or I) would not get the addictive dopamine hit from seeing my comment get voted up.

jasode 6y ago
>There are plenty of questions where the most upvoted answers is wrong or out of date. Of course -- because no system is perfect. Your proposed mechanics for designing discussion forums isn't perfect either. It's about tradeoffs. I don't expect perfection with SO's top-voted answers. Instead, I'm more realistic about the probabilities of the top answer being more useful than the bottom-most answer. Relying on probabilities is a much better usage of my limited time than reading every single answer from unvetted users for every single question on SO and mentally sorting them myself. >For a sight like HN/Reddit I think it might be ok if there was an upvote, no downvote, From a statistics perspective, the downvote mechanism helps with the spread[1] of the scores. Instead of just 2 states of 0 or +positive; you get 3 states of -negative, 0, +positive. That way, 0 and negative being won't be in the the same "undifferentiated category". Again, it's about tradeoffs instead of perfection. Your proposal tries to stop "toxic" downvotes at the expense of asking readers to wade through more undifferentiated comments (which is arguably another form of toxicity from the readers' perspective). (Arguably, there's a 4th category on HN which is "flagged" which goes beyond the downvote.) [1] https://en.wikipedia.org/wiki/Statistical_dispersion https://en.wikipedia.org/wiki/Statistical_dispersion

jasode 6y ago

>There are plenty of questions where the most upvoted answers is wrong or out of date. Of course -- because no system is perfect. Your proposed mechanics for designing discussion forums isn't perfect either. It's about tradeoffs. I don't expect perfection with SO's top-voted answers. Instead, I'm more realistic about the probabilities of the top answer being more useful than the bottom-most answer. Relying on probabilities is a much better usage of my limited time than reading every single answer from unvetted users for every single question on SO and mentally sorting them myself. >For a sight like HN/Reddit I think it might be ok if there was an upvote, no downvote, From a statistics perspective, the downvote mechanism helps with the spread[1] of the scores. Instead of just 2 states of 0 or +positive; you get 3 states of -negative, 0, +positive. That way, 0 and negative being won't be in the the same "undifferentiated category". Again, it's about tradeoffs instead of perfection. Your proposal tries to stop "toxic" downvotes at the expense of asking readers to wade through more undifferentiated comments (which is arguably another form of toxicity from the readers' perspective). (Arguably, there's a 4th category on HN which is "flagged" which goes beyond the downvote.) [1] https://en.wikipedia.org/wiki/Statistical_dispersion https://en.wikipedia.org/wiki/Statistical_dispersion