From post 18: OK, it is this last statement that seems to contradict what you said in post 14: You say now (post 18) that he is still accelerating at -1g, so when he is 27+ ... I agree with that bit, but the post 14 comment says that this negative (proper -1g) only lasts until he is 27, which indicates that either he stops accelerating altogether at 27, or he accelerates in a non-negative direction, both of which contradict the picture and your post 18. All that is pretty obvious just from the picture. Her age, relative to his CMIF, continues to increase as long as his acceleration is negative (that is, towards the distant inertial object). We don't have a perspective. We don't have a worldline or a current position in any thought experiment. We're the narrator, and it is the pragmatic purpose of the narrator (Mike) which is being served by the narrator's choice of frame. I've said this repeatedly, but y9u keep asking. Acceleration is not required at all. It is after all an abstract choice, a mental task, not a physical one. All one has to do is keep choosing different frames if that's what you want accomplished. Yes, Penrose (another narrator, not anybody actually accelerating) has the same pragmatic purpose as does Green and you. But somebody pacing back and forth typically has no reason to do a continuous frame change like that, as was also illustrated by my real world examples which you continue to ignore. And no, special relativity does not demand that you performs such frame changes. I've given far simpler explanations of the twins thing which do not involve frame rotations at all, nor does it depend on what any of the participants conclude since their role is simply that of a meat-clock. Both TDE and LCE equations have a more general form which you should be using, but don't. The crazy results you get from using the simplified version should clue you in on that, but instead you just dig your hole deeper. In the more general form (but still for 1D motion only along the one axis along which the contraction is being measured, is something like: L = ∫x=0-L L0/(γ(vx)) In other words, if the nonzero velocity of the object is not everywhere identical (which it cannot be for an accelerating object) at the moment in question in the frame of choice, then one has to integrate the contraction over its length. You're not doing that, and thus getting all the results that you cannot bring your self to admit are absurd .