Yes; there needs to be some highlighting of comments that are a few standard deviations above the average, otherwise they'll get lost in the noise, particularly if they are in child threads.

As it is, I find myself commenting less to replies on my own comments, because I'm not invested in furthering my argument - I can't see how many votes those replies got, so I don't know if I need to keep arguing my case.

(That could be a good thing or a bad thing; a good thing if I'm wrong, a bad thing if I just didn't express myself well to start with.)

Perhaps the length of the comment text can be factored into this heuristic. I mentally correlate lengthy posts as tending to be well thought out and coming from people with experience in the given area.

While a lot of insight can indeed fit into a sentence or two, it still figures to me that those who take the time to write long comments are probably contributing more to the thread than those who leave short replies.

Anyhow, it's just a heuristic that springs to mind.

With lengthier posts, I tend use the score to judge the value of the time to read them. Half a page with a high score may have more value that the linked article, whereas a low score may just be little information presented verbosely.