I agree with your approach to the question but perhaps if we really take the simulation hypothesis seriously (or at least consider it probable enough to concern us) the first step should be finding a way to tell whether or not we actually live in a simulation. Research in Physics/Astronomy could explicitly look for and device experiments looking to demonstrate systematic inconsistencies in the fabric of our universe that could give a hint on the made up nature of all laws. This in a way is an indirect  answer  to your last question. If effective altruisms is not an ideology just to be followed but a rational enterprise grounded on the actual nature of our universe, then it should also be concerned with improving our understanding of it. Even if this eventually leads to a radical re-think of what effective altruisms should be. 

Thanks for replying to my question. Your argument is certainly valid and an important one. But if we are to take the simulation hypothesis seriously it is only one within a spectrum of possible arguments that depend on the very nature of the simulation. For instance we might find out that our universe has been devised  in such a twisted way that any improvement for its conscious beings corresponds to an unbearable proportional amount of pain for another parallel simulated universe. In such a case would pursuing effective altruism or longtermism still be moral?