There are multiple formulations. In the most typical formulation, from (afaik) Gary Drescher’s book Good and Real, you only have to precommit to one-boxing if you do see the million.
I think it varies somewhat. Normally, you do have to one box, but I like the problem more when there’s some probability of two boxing and still getting the million. That way, EDT tells you to two box, rather than just getting an undefined expected utility given that you decide to two box and crashing completely.
In transparent box Newcomb’s problem, in order to get the $1M, do you have to (precommit to) one box even if you see that there is nothing in box A?
There are multiple formulations. In the most typical formulation, from (afaik) Gary Drescher’s book Good and Real, you only have to precommit to one-boxing if you do see the million.
I think that problem is more commonly referred to as Parfit’s hitchhiker.
I think it varies somewhat. Normally, you do have to one box, but I like the problem more when there’s some probability of two boxing and still getting the million. That way, EDT tells you to two box, rather than just getting an undefined expected utility given that you decide to two box and crashing completely.