You're 40, reminiscing with college friends about a memorable experience from 20 years ago.  One of them confesses that, just 10 years ago now, he broke into your house and kidnapped the person sleeping in your bed to use as an unwitting subject in a secret government fission project.  Lefty and Righty were successfully issued in one of the usual ways; but never woken up.  Righty was promptly destroyed, and Lefty was returned to your bed, still unconscious and none the wiser.  Here are two ways you might reply to your friend's confession.  1) I'm mad that you took such a terrible risk without consulting me, but thank goodness everything turned out okay.  2) I'm sad that my total life expectancy is 30 years shorter than I'd thought, but thanks for creating me.  I find the first reply much more natural than the second.  But I'm also terrified by the prospect of fission.  Leading theories of personal survival have a hard time accommodating both (a) the retrospective intuition that you've survived fission, and (b) the prospective intuition that it's indeterminate whether you'll survive fission.  In this talk, I show that we can coherently combine (a) and (b) by characterizing survival in terms of existence at a time, instead of numerical identity over time.  A simple exdurantist or stage-theoretic model can be used to illustrate consistency.  And taking care to distinguish the object language of the theory itself from the metalanguage we use to model it helps neutralize potential objections, and clarify certain confusions in the traditional literature on personal identity.  Zooming further out, these considerations suggest that Lewis may have been wrong to claim that fission and kindred puzzles of persistence aren't essentially "about identity".  Here and elsewhere in metaphysics, the way forward may involve outgrowing the habit of defining our theoretical options in terms of identity.