The greatest irony is that after that dozen back and forth set of emails where I was completely unable to communicate to Peter R the ability of miners to completely remove the communication delay from the time of mining, amid his irrelevant poetic waxing about information theory... He went on to submit a paper to Ledger describing the idea, without attribution... which incorporated most of the previous elements I'd already described in public-- save one main difference: it screwed up the design so that participants would take orphan risk for new transactions by only adding them to the pre-consensus by attempting to put them in a Bitcoin block.

Of course, rational miners could easily overlay on top of it the more efficient scheme to commit to and pre-forward new transactions _before_ actually including them; so it's no great harm... the only real effect of this design flaw was that making it seemingly serves the desperate attempt at perpetuating some residue of the "natural fee market" myth.

We probably have spent too much time here thinking about these fancy schemes: The mechanism miners have historically used to mitigate orphaning is to centralize around larger pools. It's easy, has no startup cost, no development risks, and can be incrementally-- even accidentally-- deployed.

In my view these more efficient schemes are important not because they show that "natural fee market" fails the game theory test, but because they're useful tools to cut latency (thus progress) out of mining, and thus have a positive use for making mining more fair and less centralized. The fact that they also blow up any remaining claim that orphaning risk will act as a natural, if too limited to be at all effectual, control on block size is regrettable but unavoidable.