I would recommend you direct your research toward importing into the @trufflepig analysis system some external theory of quality, possibly from the academic world.

Or maybe the Wikipedia framework could be adapted, why not? Steemit desperately needs a more sophisticated assessment framework, using the rewards is self-referential and ultimately self-consuming. The rewards should be a consequence of intrinsic quality (defined without resorting to rewards) and not the other way around. High rewards are NOT indicative of quality.

This is the most important update to be done to the @trufflepig algorithm. Bidding bots usage and other tricks are only the trees which distract us from noticing the forest