Statistical foundations of virtual democracy

In this system need to record the voting preference of each participant, then virtual voter will act on behalf of actual voter to vote.

Like a normal vote isn't it going to be hard to get the preference right. We change our vote due to our preference which can change dynamically anytime. Also sometimes people vote in spite of different preferences due to many reason.

I still think it's not an easy problem to be solved by algorithms.

Does this result apply negatively also to concordet methods which are defined to be more error resilient (e.g. minimax like "how many voter should change their mind to get ..."), if not, that would be a plus for minimax and the like.