No. You have some facts that are correct, individually, but you are drawing nonsensical conclusions.
True, individual muscle fibers are either contracted or relaxed--on or off as you say, but surface electromyography records from far more than a single muscle fiber. So at the population level, measuring a graded response is not only possible, but typical. Furthermore, the signal recorded is roughly linear and proportional to the number of fibers and motor units recruited (let's ignore the differences between type I and type II skeletal muscle fibers for the moment).
Your aside about the sympathetic and parasympathetic nervous system appears completely off topic.