I don't know of one. I doubt that everyone wants the same sort of thing out of such a metric. Just off the top of my head, some possible conflicts:

• Is a post good because it attracts a lot of responses? Then a flamebait post that riles people into an unproductive squabble is a good post.
• Is a post good because it leads to increased readership? Then spamming other forums to promote a post makes it a better post, and posting porn (or something else irrelevant that attracts attention) is really very good.
• Is a post good because a lot of users upvote it? Then people who create sock-puppet accounts to upvote themselves are better posters; as are people who recruit their friends to mass-upvote their posts.
• Is a post good because the moderator approves of it? Then as the forum becomes more popular, if the moderator has no additional time to review posts, a diminishing fraction of posts are good.

The old wiki-oid site Everything2 explicitly assigns "levels" to users, based on how popular their posts are. Users who have proven themselves have the ability to signal-boost posts they like with a super-upvote.

It seems to me that something analogous to PageRank would be an interesting approach: the estimated quality of a post is specifically an estimate of how likely a high-quality forum member is to appreciate that post. Long-term high-quality posters' upvotes should probably count for a lot more than newcomers' votes. And moderators or other central, core-team users should probably be able to manually adjust a poster's quality score to compensate for things like a formerly-good poster going off the deep end, the revelation that someone is a troll or saboteur, or (in the positive direction) someone of known-good offline reputation joining the forum.