I agree that pools are the biggest problem.

They logically result from the winner take all algo.

I postulate that in a theoretically ideal POW system, every single participant that contributes cycles/work towards securing the network would be rewarded in direct proportion to their contribution.  ie, even if they only calculate 1 hash during a given block and do not match target difficulty.

How to achieve that?  I have no idea.  haha.  But I'd like to see more discussion/brainpower/research on it.


First question:  Is the above scenario ideal/desirable for maximizing distribution of work across non-colluding parties?   Or the incentives are flawed somehow?

Followup question:  What technical obstacles prevent it, and how might they be overcome?   (assuming a brand-new system with no legacy code/chain).







I will leave these open-ended for now...