NotaLOT,
Are you familiar with the concept of secondary areas, as they apply to MOC reduction? ie the protection areas are broken up into primary and secondary areas. The primary area is usually half of the 3SD width at any given point and this area gets full (100%) MOC. The secondary areas then have the MOC reducing linearly to zero at the out edge of the splay.
For protection areas in final approach it is my understanding that, the primary areas roughly correspond to half scale deflection (VOR/RNAV) or +/-5 (NDB). Since the aircraft can legally fly anywhere within these tolerances we have to assume that the aircraft will be anywhere in the primary area. Once the aircraft has exceeded tracking tolerances, then it is assumed the aircraft will conduct a missed approach so that the probability of the aircraft being outside this primary area decreases the further out you get. (Hence the reducing MOC in the secondary area). This is also valid in other phases of flight (enroute, initial approach,etc)
To specifically refer to noise corridors, (this is just my opinion) yes you could reasonably assume the aircraft will fly on track, especially with todays nav and autopilot systems. However, you probably would need to consider aircraft that still use steam driven nav systems or no autopilots and they could be anywhere in that primary area.
Hope this helps
Alpha