B About a Demon

Include stuff about arbitrarily accurate Note that known perfect accuracy leads to challenges, one’s we’ll come back to in…

Also note that we’re assuming the demon predicts the strategy a player chooses. That could include randomisation.

Note that I’m simply assuming expected utility theory is right for non-demonic problems.

But that does imply an assumption that there is some reason for that. And I think that’s got to include something like a Sure Thing Principle.

But I am including the open box Newcomb Problem as demonic.

And maybe a riff on why it’s Player/Chooser/Agent whatever