This totally ignores effects from the long tail, black swans, power users, and criticality dynamics. Downvotes are way more "potent" than upvotes. If I'm liste

This totally ignores effects from the long tail, black swans, power users, and criticality dynamics. Downvotes are way more "potent" than upvotes. If I'm listening to music that I like, I'd say I probably log over 100x upvotes than downvotes. But when I downvote, I usually want to nuke the site from orbit. What is the ratio? What ratio of users use downvote? What ratio is that compared to users that interact in any way? How are you controlling for sampling bias? Berkson's paradox? How much utility are the users that downvote obtaining, vs the disutility of being unable to?