If we posit that, for example, for any given choice of strawberry, chocolate, or vanilla ice-cream, the resulting selection that Avalon makes is completely predictable, it seems to me that Avalon's version of free will must involve allowing for some variation in his internal state - i.e. although the selection he actually makes in each instance is completely predictable,
given a slightly different internal state, he would [be physically able to], make a different selection. As he says, he has the
capacity to make any selection.
To me, this approach treats the internal state as effectively private, analytically off-limits, and also appears to boil down to a compatibilist view - free will is exercised if the options are physically available to be selected barring coercion or constraint,
given an appropriate internal state when the decision-making occurs.
Part of the cross-talk here is that one group is taking the wholly analytical approach where the choice is illusory because the single clearly defined internal state evolves predictably as the processing of the input options gives rise to the inevitable output selection, and another group is black-boxing the internal state and saying the choice exists because,
depending on the precise internal state, any selection is possible.
So maybe it's not about whether omniscience changes or forces the selection, but whether or not you want to precisely specify the individual's internal state when discussing this. I'd prefer that we did, however comforting the feeling might be that we 'have a choice' because if the starting point was slightly different the outcome would be different.
How QM indeterminacy can be assimilated into either view, I don't know
