The only way to make the rational approvability of A* within OEU imply the welfare goodness of W* is (1) to have the objective probabilities reflect full information; and even then only if (2) the universe is deterministic (so that fully informed persons would always know for certain which outcomes will result from which actions).
The move from SEU to OEU permits this convergence by eliminating the possible deviation between the probabilities relevant for the Agent's moral choice and the probabilities relevant for the Patient's rational approval.
More generally, Fumerton's point calls into question all OEU accounts--although there could be such accounts where the deviation from the actor's epistemic situation is less pronounced, for example, where objective probabilities are defined in frequentist fashion and relative to conventional or statistical rather than physical classes.
In sum, although the issues are complex, it does not seem that there is an OEU account of rationality that both (1) is in fact a plausible account of what rational choice consists in, and (2) makes the rational approvability, by some Patient, of some Agent's choice, a morally relevant consideration for that Agent.
SEU, OEU, SEV, OEV, RU, and RV, however, hardly exhaust the set of possible or plausible accounts of rationality.
I argue below against the moral force of rational approvability given OEU.
35) See EELLS, supra note 2, at 31 (suggesting OEU as an alternative account of rationality to SEU).
41) In addition, the "final-preference" gambit is not strengthened by the shift to OEU.