Been a while, may not be entirely correct: Slashdot's model was perhaps a little overcomplicated, but my favourite feature was the ability to tag up/down votes

Been a while, may not be entirely correct: Slashdot's model was perhaps a little overcomplicated, but my favourite feature was the ability to tag up/down votes with flavours. +1 Informative was different to +1 Funny, and "Factually incorrect" was a different downvote to "Off-topic spam" (whatever they were called). Other quirks off the top of my head: it capped at +5 and ... -1, I think? The score represented a thing closer to the up/down ratio than "Facebook likes". There was a dedicated -1 Overrated moderation for "I don't disagree that it's interesting, just not +5 interesting". Also, logged in users got a fixed number of moderation points at random intervals, and you couldn't moderate in a story that you commented in. I'd like to believe this discouraged "throwing away" points on low-effort joke comments, but I'm not sure the facts of Slashdot comments entirely bears that out. EDIT: And then there was meta-moderation...