Actualism implies that Procrastinate ought to say no, and that he ought to say yes and then write. It is impossible to do both. It seems we have a refutation of Actualism.
However, Actualism does make it possible for agents, including Procrastinate, to do everything they ought. To see this we need to bear two facts in mind: (a) the reason Procrastinate ought to say no is that even were Procrastinate to say yes, he would not write, and (b) the reason Procrastinate ought to say yes and then write is in part that Procrastinate can say yes and then write. Suppose Procrastinate did just this; then it would be false that were he to say yes, he would not write -- the subjunctive conditional would have a true antecedent and a false consequent -- but then it would be false that Procrastinate ought to say no. [...] And he would in this case do all that he ought. [p.242]
That's fine as far as it goes. But so far we have a very limited theory: it tells us only what we objectively ought to do when there are determinate counterfactuals. The next step is to relax one of these assumptions, and so extend the Actualist theory to cover (epistemic or metaphysical) chanciness.
To do that, we consider the case where the only eligible information regarding the future is that it is overwhelmingly likely that Procrastinate will fail to write the review on time if he says 'yes'. In this case, the Actualist (or global consequentialist) will presumably still want to say both that Procrastinate ought to say no and that (though he's unlikely to do so) he ought to say yes and then write. But how can he possibly meet both obligations? Even if he does say yes and then write, the "no"-supporting premise that he's overwhelmingly unlikely to follow through on his commitment remains true. So J&P's response to the objection fails to carry over to this natural extension of their theory.
A better response, I think, is to distinguish acts from act-sequences. The best available act-sequence is to say yes and then write. But that is of little moral significance, since you can't (be guided by reasons to) perform an act-sequence. (Do leave a comment on that linked post if you disagree!)
Procrastinate ought to say 'no'. This is what he is obliged, and has decisive reason, to do. We may add that it would be desirable for him to fail in his obligation by saying 'yes' and then (against all odds) writing the review. That is all that can sensibly be meant by saying that he "ought to say yes and write." (So it probably isn't worth saying in such misleading terms!) The other thing that could be meant by the phrase is that he is obliged both to say yes and to write -- that these two actions are each obligatory actions. But that would be to say something false, since (as already noted) Procrastinate has decisive moral reasons not to say he'll review the paper, since that action is overwhelmingly likely to have bad consequences. There is no further candidate meaning for the phrase that I can see. In particular, there is no further meaning to be given to the claim that one has an obligation to perform an act sequence -- to "say yes and write" -- since, as already noted, sequences aren't things for which there are reasons.