Mission control would have had the trajectory back up calculated and in hand, but separating from the lower lander stage and engaging the (different) return roc

Mission control would have had the trajectory back up calculated and in hand, but separating from the lower lander stage and engaging the (different) return rocket at low altitude, outside of normal flight parameters, with an ongoing computer problem of (then) unknown origin - I would agree, that sounds still dicier. Plus, a manual landing was practiced in computer-governed simulators; it wouldn't feel the same, but otherwise the task would be quite similar - except for the dust perhaps.