I have long thought that implementing something inspired by a [2nd price auction](http://www.gametheory.net/dictionary/auctions/secondpriceauction.html) for rewards in conjunction with a superlinear rewards curve (to discourage stake-splitting) would go a long way towards solving the valuation problem.

Unlike today's system where a voter can maximize curation rewards by voting at maximum strength, an algorithm inspired by a 2nd price auction could drive high-powered voters to self-limit their own voting strength to accurately match the perceived value that they place on a post.  Otherwise, they'd wind up wasting their voting power.

Among other places, I wrote about that in [Simulating a Steem curation rewards distribution that is modeled after a 2nd price auction](/@remlaps/simulating-a-steem-curation-rewards-distribution-that-is-modeled-after-a-2nd-price-auction)