We are in both. Both universes are equally real. Each 'copy' of you knows it's in the universe where the result matches the observation.
If in my universe I could always predict the correct results, then we would just have determinism, and I could predict exactly when an atom would decay. There would be no need for statistics.
Some high level background that might help.
https://www.youtube.com/watch?v=z-syaCoqkZA
The 2022 Nobel Prize in Physics was for experimental verification of Bell's theorem. The experiment did not rule out superdeterministic local hidden variables; superdeterministic local hidden variables does not violate Bell's theorem, since Bell's theorem assumes "free will" (that there is no correlation between arbitrary choices made by an experimenter and the state of the system being measured), but superdeterminism is the denial of that assumption.
And for many worlds. Doesn't it punt randomness into other universes, but doesn't help us solve for results in our own individual universe. Since we can't measure what happens in the other universe we don't really know. If there were two results, and one is in one universe, and one in our universe, sure we determinedly know both results. But we don't know which universe we are in, so instead of a random result, now we have 2 answers and 2 universes, but now randomly don't know where we are?