Does the following make any sense?
My thinking was that if two far-away detectors measure an entangled pair of photons (e.g.), each detector will measure both possible results (e.g. up *and* down). Each detector and thereby their environment becomes entangled with that photon. So each detector and it's environment starts a new branch in their respective many-worlds reality. (One side of the branch for “up” and one branch for “down”).
When you later compare the measurements of the detectors, you will find the measurements pair up (for example they are opposite). In the classical interpretation this could be thought of as a “spooky action at a distance” (instantaneous synchronization). But in the many-worlds interpretation only the worlds where the two separate measurements pair up would survive (the worlds where there is no match would cease to exist, as you put it). This would require no instantaneous synchronization, but would appear as such at the moment when the station that is comparing the measurements is becoming entangled with both detectors, e.g. by receiving the measurement outcome information from both detectors. The four “realities” (e.g up-up, up-down, down-up, down-down) meeting at that moment would be reduced to two “realities”, by merging pairs of “compatible realities” (only up-down and down-up “survive”).
INAQP (I’m not a quantum physicist) so I hope all of this makes sense. And I guess I haven’t added much to the parent’s point except adding a (hopefully valid) example.
If any expert reads this, I would love to know where I can read more about these ideas. There would certainly be a term for this already.
I only wonder why this possibility isn’t discussed more often, I seems such an easy way out of the paradox.