Doesn't the probability basically follow from the uncertainty of your own eigenstate, ie. the system performing the measurement? This contextuality is why deterministic interpretations of QM also entail probabilistic measurements.