Posts from LW 2.0 have a reliably higher score than posts from LW 1.0 because people's vote weight went up. Everyone used to vote 1, but now people can vote between 1 and 15.
At some point it'd be good to backdate people's vote strengths (i.e. change their votes on previous posts to their current weak vote score of 1, 2 or 3) to even things out a little.
posts from LW 1.0 often have a higher score than they deserve because [reason]
Asides from the obvious karma accumulation, I think LW1 used to have a big community that fell apart after... stuff happened. I'm not really clear on the details since I discovered rationality well after LW2 was built, but that's the impression I've gotten from a few halfhearted investigations.
I downvoted this because it's encouraging negativity with no benefit as far as I can tell. A list of underrated posts is theoretically useful because people could follow the links and read them, but what would you do with a list of overrated posts?
If there are strong but not widely publicized criticisms of highly upvoted posts, it might make sense to have those criticisms more widely publicized (so people don't just take those things at face value). But this feels like a special case of your point, where it's really the criticisms that are underrated.
I suspect there are many such criticisms, and that they tend not to get found, especially with older posts. Say that a post from 2017 has a subtle flaw that isn't discovered until 2019. A comment pointing out the flaw probably gets buried under 50 other comments before anyone pays attention. A post directly responding to a 2-year-old post takes hours to write, and probably gets less attention than a new post with equal effort.
I looked at the top 10 posts from 2017; most of them have no criticisms or highly upvoted comments from the last year, and none of them have highly-upvoted criticisms. We've come far enough since 2017 that even if these posts were high-quality at the time, there should still be caveats for today's readers, and I can't find any.
You might be interested in the 2018 Review, which spurred discussion of this sort, both as reviews on the posts and in new posts that were replies.
Thanks. I didn't look closely at the review project before, but it seems to be much of what I need. I wish there were some longer-term review system where we would continue putting LW1 and diaspora era content in a current context.
You would downvote them in order to make the sorted-by-karma archives more useful! (See the tragically underrated "Why Artificial Optimism?")
but I suspect some posts are not nearly as useful or fun to readers in 2020 as their karma makes them appear.
Maybe it's because they're old. (You also might want to treat karma as a noisy signal.)
or a general pattern (e.g. posts from LW 1.0 often have a higher score than they deserve because [reason]).
The older a post is, the more time it's had for people to read and upvote it. (Admittedly, this seems reasonable in that if karma reflect the (positive) impact it's had on people's lives, old posts would still have this advantage.)
ETA:
As for a specific example, I'd say The Sequences, because it doesn't seem like there's been a sequel.
I've noticed the quality of LW posts varies greatly, even among top posts. Some of this is subjective or comes from what I want out of LW, but I suspect some posts are not nearly as useful or fun to readers in 2020 as their karma makes them appear.
Answers should be links to particular posts with more than ~75 karma that the answerer has a relatively unknown strong criticism of, or a general pattern (e.g. posts from LW 1.0 often have a higher score than they deserve because [reason]). Please give a reason why, and don't automatically downvote a post someone else links.