The concept of randomness carries with it a puzzling paradox. For example, whatever test you develop to test for randomness, then one can argue that if it obeyed the test, it can't be random if it conformed to a defined pattern, conclusion being that the validity of any such test is disproved by reductio ad absurdum. Localised shuffling cannot be even pseudo-random because a minimum definition for randomness has to include that all candidates have an equal chance of selection. Obviously pseudo-random doesn't quite do that, although in a way it does if you can manage to hide some of the calculation details from yourself, so that you can't make skewed predictions like the result being in a certain range.