Just a meta idea: Treat values −1, 0, 1 as noise. Make a list of your upvoted commends, and another list of your downvoted comments. When you compare the lists, maybe you’ll see some difference.
Hmm. So far, I’m noticing more similarity between the extremes on both sides than I’m noticing with the ‘noise’ in the middle—i.e., it seems like I can roughly predict that a statement will have a large number of votes, but not in which direction.
Just a meta idea: Treat values −1, 0, 1 as noise. Make a list of your upvoted commends, and another list of your downvoted comments. When you compare the lists, maybe you’ll see some difference.
Hmm. So far, I’m noticing more similarity between the extremes on both sides than I’m noticing with the ‘noise’ in the middle—i.e., it seems like I can roughly predict that a statement will have a large number of votes, but not in which direction.