Hi. Please could someone explain the following. I don't understand why the low probability experiment gives 50% spin up, leading the observer to conclude that the Born rule is incorrect. Even in a frequentist interpretation.

It is particularly interesting for Everett’s Many Worlds Interpretation (MWI), where we will argue that it eliminates the perplexing feature that, loosely speaking, some observers are more equal than others. To wit, suppose a spin measurement that should yield “up” with probability p = 0.5 is repeated N = 10 times. According to the MWI, the final wavefunction has 2 N = 1024 terms, each corresponding to an equally real observer, most of whom have measured a random-looking sequence of ups and downs. This suggests that quantum probabilities can be given a simple frequentist interpretation. However, for an unequal probability case such as p = 0.001, the final wavefunction still has 2 N terms corresponding to real observers, but now most of them have measured approximately 50% spin up and concluded that the Born rule is incorrect. (We are supposed to believe that everything is still somehow consistent because the observers with a smaller wave function amplitude are somehow “less real”.)
From https://arxiv.org/abs/1008.1066