Anyhow, as Salamanderanagram has shown - you are incorrect in your assertions.
Another simple test is to modify the structure in the image you posted as follows:
Remove the A to E perm, and connect the switch output directly to the + module.
Attach the output of the add to an Expon(F), then connect that to the F input of the sine osc. Remove all connections to the P input.
Notice that the sound is identical, but you have a cleaner structure - no more daft audio rate event stream.
Now, to avoid serious injury, use a spork for the decapitation.