"You (unnecessarily) assign lambda as the orientation choice for the algebraic basis used, then sum over lambda_k with the requirement on Nature of a 50-50 statistical choice of +1 and -1."
Come on, Rick. The variables are nonlinear input to a continuous function. The experimenter is forbidden to assign initial condition. That's the whole problem with Bell's choice of measure space.
"The math works out to the same conclusion for either orientation choice."
The math does not work, however, for a function continuous from the initial condition. One would get only a flat line for A, flat line for B.
Tom