Newcomb’s Problem

This is a question about “Newcomb’s Problem”.
Two boxes, one transparent and visibly containing $1,000 (A), and one opaque (B) are placed before an agent X at time t.  The contents of A are supposed fixed, the contents of B known by the agent to have been determined by the prior action of a highly accurate Predictor that has placed $1,000,000 in box B if it predicted that X will select only box B and has placed nothing in box B if it predicted that X will select both boxes.
A)Assume (for the sake of argument) a universal acceptance of the in fact  completely discredited hypothesis that the strong statistical correlation between smoking and a host of serious diseases including lung cancer is accounted for by a genetic factor that is the common cause of both.  In your view, can someone who believes that these statistics should not in themselves present a deterrent to smoking reconcile this view with an advocacy of the “one-box” solution to Newcomb’s Problem?
B)Suppose that the back of box B is transparent, and that a completely trustworthy and reliable friend of X is able to see whether $1,000,000 is in box B.  Assume that the Predictor can predict what if anything the friend will say and what X will hear, and that it has factored this into its prior analysis and decision.  If the friend were able to communicate with X, does it matter whether i)the friend simply recommends a selection (“Take both boxes!”) or ii)reveals the actual contents of box B to X (e.g., “Box B is empty”)?  Why or why not?  If the friend were able to announce out loud the contents of box B, would it be advantageous for X to place himself in a situation in which he is unable to make out what his friend says?  Explain.  

probability