Justo wrote:Well, what about this: let

be the singlet state and

, then all you need is

Do you mean the above result is wrong?

It is correct if used as a prediction of the results for each term in my scenario 2, for which the upper bound is 4.

It is wrong if used as a prediction for my scenario 3 or 4, for which the upper bound is 2. I've shown that Bells inequality boils down to scenario 4, if the data can be reordered. But I've also explained why the data can't be reordered. Therefore, there are two strikes against Bells theorem:

* To obtain an upper bound of 2, the data must be reordered then you end up with a 4xN spreadsheet and the correct QM prediction for it does not violate it.

* But it is not possible to reorder anyway, so the inequality can't be derived to begin with.

Finally, the above analysis has already granted the fair sampling assumption. Therefore any other discussion about locality, conspiracy, superderminism, freedom, or other considerations, is irrelevant as they only apply inasmuch as they affect the distributions of lambda.