Well, I wouldn't be so bold as to suggest this forum's questioning is a reason why, but just minutes ago NT Times confirmed the computer pre-programming angle:
This NYT article seems to confirm earlier reports that the plane was detected as a primary radar blip passing over the waypoints west of Peninsular Malaysia, not just going in that direction.
If you're cruising out there with lights and TCAS off, would you do SLOP? Or just stay on the centerline, incorrectly using the Big Sky theory?
Years ago aircraft would be scattered a mile or three off the centerline out over the water due to INS drift and lack of precise long range radio navaids. Nowadays with GPS, the planes will appear to be within 10 meters of each other laterally when they pass at (hopefully) different altitudes. MH 370 staying on airways instead of going random route would sure indicate someone was conscious and driving, at least initially. And trying to blend in with the commercial traffic.
The reversal route back past Penang wouldn't normally be loaded for the KUL-PEK flight. However, would it be part of a canned company flight plan stored in the FMS for, say KUL-DEL or KUL-KHI? These canned flight plans seem to go in and out of favor depending on the carrier. Many planes can uplink the route directly to the box. However, internationally, sometime this doesn't work and a canned flight plan, even with some changes, may be less prone to error than typing in the whole thing from scratch.
The report from the New Straits Times about MH 370 descending to 5000 feet doesn't seem to be corroborated elsewhere so far.