I bet you that the algorithm could be adjusted to consider comments that meet the following criteria: (1) If user has the most upvoted comment in any thread (2

I bet you that the algorithm could be adjusted to consider comments that meet the following criteria: (1) If user has the most upvoted comment in any thread (2) Check if that thread had a lot of flamewar language. If so, return that as the most hateful comment. I'm only half joking about using this approach.