I would also like to state explicitly:

With PPQ=4 and schedhz=16, as you suggest, the scheduler quantum
(really, the speed at which a process falls from one effective
priority to the next) is really 1/4s.  I'm not sure where this was
ever documented, but it is the historical behavior with stathz=64.  If
this is to be our Official Scheduler Quantum, it should really be
documented somewhere.