Yes, the complexity is certainly one of the downsides to what I’m proposing, which is one reason why I was curious if people thought the complexity would be manageable. Sounds like you think not?

Just to clarify, my thought is to leave this up to users/admins to choose their own algorithm, which would transparently describe how things are weighted. For me, I would like to weigh factual information most highly, then kindness, with raw popularity at the bottom.

For new users or those who don’t understand the system, it would be fine to have a default sort, maybe configurable by your instance. It could be as simple as just adding up the positive and negative votes, which would make it identical to the current system, or we could just guess at some different weights. Let me people try them out—not everyone will engage but I hope enough would to help iron out the wrinkles and see what works best.