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...

Slashdot's method of scoring comments was overly complicated and probably did not produce any better results than reddit-style voting. However, Slashdot's killer feature was that the reader could filter by comment score and thus only read the 'good' comments, and not have to wade through hundreds of replies.

correct, they were better than reddit because they let the user sort based on their preference. slashdot generated a ton of metadata that described their content, and then gave you the power to intelligently utilize that metadata.

I suggest reading the FAQ

https://slashdot.org/faq

especially all the Moderation, metamoderation, and Karma sections!!