Giving up spacetime local realism. Re. the ERP argument that predicted variables are elements of reality:---
Finding the state or measurement of a variable is not like pulling a 'magic' existing rabbit from a hat. Such a rabbit must be in the hat prior to extraction. Instead it's more like determining (or predicting) how long the rabbit remains calm or how long it struggles upon removal. The two determinations are mutually exclusive. Like position and momentum. Calm is not struggling. Struggling is not calm. The descriptions of the rabbits behaviour are new observation products. They are not the same as a material existing rabbit.
Experiments could be conducted using pairs of tame rabbits put into two hats and likewise pairs of untamed rabbits likely to struggle. (Ignoring the freeze response . This is analogy, not about real rabbit behaviour.)-------------In a uni-temporal existential reality there is no after extraction state prior to extraction happening. There is not a prior to extraction post extraction observation product in the universe. ( Unlike in the space time continuum model.) -----That does not mean there is no existing rabbit in Object reality. A prediction of what the result would be if measured is not an observation product. The prediction (A would if) depends upon there being an existing rabbit of a like pair. The observation product depends on the extraction and behaviour determination. The prediction is neither existential noumenal rabbit nor observation product phenomenon. It is a maybe, imagined. Not an element of either reality.