Under those conditions, everyone would pick the utilitarian solution. There lies the problem IMO: psychologists carrying out experiments about "morality" when they have absolutely no theory of morality. Lacking this, they're left with "two of the same bad thing is worse than one of that exact same bad thing," and limit themselves to problems of relative morality, ignoring the fact that you can't do the same bad thing twice.

Ignoring is a mild way to put it, actually, because all of their experiments involve doing multiple additions and subtractions and substitutions. For them, morality is definitely a real number.

I think the complexity of these hypotheticals is entirely a result of a need for varied responses. The real answer is that there's not enough information given to make a determination. In the case of your hypothetical, there's clearly enough information (because one of the premises is that you have all possible information.) The question for the subject becomes easier, and the tough question becomes "Why is this horrible magical creature doing this, and how can we make it stop?"