I don't think that time shifting the gravity vector has a measurable effect on the orbit. Indeed, the vector is 1au*c late in real life, but we didn't know that for sure until the recent neutron star collision observation.