In article , Steve Willner
writes:

In article ,
Phillip Helbig---undress to reply writes:
(To be sure, the finite speed of propagation needs to be taken into
account, at least in a non-static situation,

Are you sure about that? I'm no GR expert, but I thought you can
calculate the potential by considering the positions at a given
instant in a given reference frame. Of course the potential is
frame-dependent, but that's as it ought to be.

I think that's basically equivalent. In practice, I think people use a
Newtonian framework and the finite speed of propagation doesn't matter
since it cancels out at first (or even higher?) order in a full-GR
treatment.

If one thinks of a huge N-body simulation, there would probably be some
"universal frame" (e.g. that of the CMB) and one would use that as a
starting point, not that of some observer moving relative to it.

Think of something like simulating a sky as seen by an observer,
including gravitational lensing. One has to follow a light ray as it is
deflected by an evolving universe on its way to the observer. In this
case, one has to take more than one instant into account.

This is all irrelevant to the OP's question, though.