I can’t model the transition from epsilon votes to all-in votes.
Rick Dudley

“okay, from my subjective view, we are past epsilon”

Yeah, you need an objective measure of this.

Others need to be able to see what you’ve presented as evidence, and, deterministically come to the conclusion that the thing did pass epsilon.

That’s how you get deterministic accountability. If the other person can’t produce the evidence to justify their actions, then they are liable.

Now the trick is to construct that deterministic algorithm in such a way that should there be a fork condition, you can review the evidence from both forks, and know (with proofs!) that some large subgroup (e.g. >1/3) is ultimately liable.

Then you have an algorithm that is deterministically accountable even in the case of forks.

I would love to see Casper develop the epsilon rule so that it does that. When a fork happens, you can figure out who’s responsible.

Anyone can lose a bet. We need to prevent the majority from colluding to take away the voting power of minorities, and get away from it without a hint of evidence.