Referencing back to Capt Kremin's post of long ago...
"If a 777 reaches has a route discountinuity in the FMC, it reverts to HDG mode. All the MCP setting are referenced on magnetic unless in the polar regions (80 degrees N or S) or the HDG REF button is pushed. That would be a deliberate act of someone who knows the systems and implications.
In the case of the "Ghost plane" scenario the aircraft, after it turned WNW would have been either tracking to a programmed FMC waypoint or it would have been in a lateral AP mode referenced on magnetic north, HDG or TRK, it does not really matter. If tracking to the FMC waypoint, once it reached it, it would have reverted to HDG."
Some of his earlier calculations based on the likely track have now been superseded by updated information due to continued analysis and therefore revised notions of direction, speed etc. But it does suggest the possibility that the plane could have been on AP for the most part of the southward curved leg of the track to suspected point of impact. That supports the idea that the flight path back to Malacca Strait and then WNW may have been controlled, whereas final leg south not necessarily.
The full post with Isogonal chart is available here:
http://www.pprune.org/rumours-news/5...-lost-377.html