First, there's no preferred reference frame. There's no frame you can point to as having been "always at rest". The very word relativity means that all frames are equally valid. If you accelerate from one frame into another, you can't say that the new frame is "tainted" or less valid somehow. If your physics doesn't work equally well in all frames, it's wrong.
Second, you don't always know whether it's true that any given object has undergone accelerations, or what those accelerations might have been. I said "spaceship", but it could just as well have been an asteroid or subatomic thingumbob that has been at a constant velocity since the moment of creation. It wouldn't change the problem.
Third, it doesn't matter if the ships "have undergone" accelerations. The spaceships aren't accelerating in the problem that I posed. Everybody maintains constant velocity throughout the entire problem, from before the beginning until after the end, and still the counters don't agree, and that disagreement is quantitatively calculable based only on the velocities and distances.
I think I've introduced some confusion here. If the problem is looked at as assuming the frames of the spaceship and the Earth are symmetric, each inertial observer concludes the other has aged less. The only way to identify the correct frame for time dilation is to note which one has accelerated, or is moving with respect to the turnaround point.
The Earth observer sees the distant turnaround point as fixed and at a greater distance than the observer on the spaceship, after the spaceship hits cruising velocity. In your examples, the spaceship was always the one moving and frame symmetry was not mentioned.