Since the correlations in the numerical simulations depend only on common variables, extending the scheme with other arms makes possible to check that all arms pairs are correlated in the same way.

In case of an unlikely doubt, it is computational and can be checked easily with the simulation programs by sharing the variables file by as many participants as desired.

#### Implications for cryptology

This is problematic if you want to use emulation data to encrypt. The common variables file becomes the crypting key, at best dynamic. In the physical reality of quantum cryptology, whether it is hidden variables or quantum properties, a spy will have to alter a flow to intercept it. With bits circulating on a network, this is easy.

How to build variables sequences dynamically? By the way, could not we use a function offering a better gap than the cosĀ² an a better efficiency than 65 to 75% while offering similar properties?

Apparently no ! Without the quantum magic, it will be only a new secret key encryption function.

Perhaps there will be no solution if Bohr’s quantum entanglement is irrelevant in nature or if the detection levels allow parasitism. But there are still tracks to explore with a bit of quantum and hidden variables. It is better to remain cautious and still search before new labs experiments.

in 2018, I can say that the difficulties evocated in this post are no more pertinent … In fact, the study of compatible solutions opens to finite number ( while huge ) classes of solutions depending of a specific correlation evaluation function and of the properties of the outcome polarizer set ( and then of the random function ). Then a dynamic key emerges and it can be protected by a protocol. As a consequence, and sadly, there is no need to join for Alice and Bob if they use our bidirectionnal protocol over any other, on a common network.