If in Newcomb's problem you replace Omega with James Randi, suddenly everyone is a one-boxer, as we assume there is some slight of hand involved to make the money appear in the box after we have made the choice. I am starting to wonder if Newcomb's problem is just simple map and territory- do we have sufficient evidence to believe that under any circumstance where someone two-boxes, they will receive less money than a one box? If we table the how it is going on, and focus only on the testable probability of whether Randi/Omega is consistently accurate, we can draw conclusions on whether we live in a universe where one boxing is profitable or not. Eventually, we may even discover the how, and also the source of all the money that Omege/Randi is handing out, and win. Until then, like all other natural laws that we know but don't yet understand, we can still make accurate predictions.
No. I think that is fighting the hypothetical.
More generally, the discipline of decision theory is not about figuring out the right solution to a particular problem - it's about describing the properties of decision methods that reach the right solutions to problems generally.
Newcomb's is an example of a situation where some decision methods (eg CDT) don't make what appears to be the right choice. Either CDT is failing to make the right choice, or we are not correctly understanding what the right choice is. That dilemma motivates decision-theorists, not particular solutions to particular problems.
If it's worth saying, but not worth its own post, even in Discussion, it goes here.