I've been thinking about two forms of duplication. One is duplication of the entire universe from beginning to end, as envisioned in Nietzsche's eternal return (cf. Poincare's recurrence theorem on a grand scale). The other is duplication within an eternal (or very long) individual life (goldfish-pool immortality). In both cases, I find myself torn among four different evaluative perspectives.
For color, imagine a god watching our universe from Big Bang to heat death. At the end, this god says, "In total, that was good. Replay!" Or imagine an immortal life in which you loop repeatedly (without remembering) through the same pleasures over and over.
Consider four ways of thinking about the value of duplication:
1. The summative view: Duplicating a good thing doubles the world's goodness, all else being equal; and in particular duplicating the universe doubles the total sum of goodness. There's twice as much total happiness overall, for example. Although Nietzsche rejected the ethics of happiness-summing, something in the general direction of the summative view seems to be implicit in his suggestion that if we knew that the universe repeats infinitely, that would add infinite weight to every decision.
2. The indifference view: Repetition adds no value or disvalue, if it is a true repetition (no memory, no development, no audience-god watching saying "oh, I remember this... here comes the good part!"). You might even think, if the duplication is perfect enough, that there aren't even two metaphysically distinct things (Leibniz's identity of indiscernibles).
3. The diminishing returns view: A second run-through is good, but it doesn't double the goodness of the first run-through. For example, the total subjectively experienced happiness might be double, but there's something special about being the first person on the (or "a"?) moon, which is something that never happens in the second run -- and likewise something special about being the last episode of Seinfeld (or "Seinfeld"?) and about being the only copy of a Van Gogh painting (or a "Van Gogh" painting?), which the first run loses if a second run is added.
4. The precious uniqueness view: Expanding the last thought from the diminishing returns view, one might think that duplication somehow cheapens both runs, and that it's better to do things exactly once and be done.
You might think that this kind of question isn't amenable to rational argumentation -- that there is no discoverable fact of the matter about whether doubling is better. And maybe that's right. But consider this: Universe A is just like our universe. Universe B is just like our universe, but life on Earth never advances past microbial levels of complexity. If you think Universe A is overall better, or more creation-worthy (or, if you're enough of a pessimist, overall worse) than Universe B, then you think there are facts about the relative value of universes -- in which case, plausibly, there should also be some fact about whether a duplicative universe is a lot better, a little better, the same, or worse than a single-run universe. Yes?
There is, I think, at least a chance that this question, or a relative of it, will become a question of practical ethics in the future -- if we ever become "gods" who create universes of genuinely conscious people running inside of simulated environments (as I discuss here and here), or if we ever have the chance to "upload" into paradises of repetitive bliss.