I've made one important change in how I vote as a result of the hidden score change. I'm now more afraid to downvote. My usual downvote algorithm was: If offtopic/noise and score >-4 then downvote. Now I don't downvote for fear that it already has 100s of downvotes and I'd just be dogpiling.
Maybe I am misinterpreting it, but comments seem to vary in color from black to nearly white. I am not sure how many levels of gradation there are, but it seems that ones that have been down voted are lighter than those that haven't?
If the post is offtopic/noise why is -4 a magic threshold? If that threshold can be justified why not incorporate it into HN as policy? If it's just to protect egoes, then HN can just secretly use something like a log scale or some other relationship when displaying karma to users.
If voting patterns are more independent of a post's current karma, the measurement should be less biased. The same is true for positive karma--I often don't bother vote up comments that already have relatively "high" karma.
Its the functional equivalent of "someone should tell him that his comment is just a little off", and the score: "oh, good, someone has | I guess no ones done it yet, I guess I will".
I would see loosing a record of your evaluation as a loss for the community. You've gone to the effort of evaluating the content, but your feedback is not recorded and is treated the same as if you had not read it at all. Maybe rather than just ^/v voting, capturing that sentiment is better handled by an agree--disagree scale.
