I'm a current B777 pilot.
About a year ago, in response to a Boeing Flight Crew Ops Manual Bulletin (BAB-95) relating to a potential software glitch affecting the selected altitude on the MCP, the operator I fly for changed our procedures to comply. Before the change we used to leave the Altitude increment selector in the 1000s position, subsequent to the change it is now routinely left in the Auto position (i.e 100 foot increments).
Had the flight crew experienced a rapid decompression then following donning of oxygen masks, establishing comms between the two of them, checking for cabin altitude/rate and passenger oxygen as required the next checklist item calls for an emergency descent. In my company this is routinely taught as a double loop starting with selecting a lower altitude on the MCP altitude selector, optionally using HDG/TRK SEL to turn off the airway, pressing the FL CH switch (commence descent at current IAS), deploying the speedbrake and ensuring the thrust levers are closed. Once the descent is established a second sweep tidies up the altitude selected (MSA/10k'), refines the heading, increases speed if appropriate and checks again for speedbrake deployment and thrust lever closure.
If there had been an explosive decompression centred on the E&E bay that knocked out all comms equipment and the crew oxygen could it be that the initial actions of the emergency descent checklist, the first sweep, be accomplished then hypoxia set in before the second sweep could be completed?
Under the old 1000s position of the ALT SEL an anticlockwise spin would reduce the selected altitude by many thousands of feet, an anticlockwise spin of the heading could give a large heading change to the west, the aircraft would descend at current IAS (normally in the cruise at FL350 M.84 equates to 250-260Kts indicated). However if the ALT SEL was in the Auto position a quick spin would change the selected altitude by only several hundreds of feet, possibly by pure chance FL295?
This scenario could explain the lack of comms, the primary radar target tracked on a westerly routing and the fact that no debris has yet been found. At FL295 the pilots would succumb quickly to the effects of hypoxia, the passengers likewise once the passenger oxygen was depleted. The cabin crew may have transferred to portable oxygen and attempted to reach the flight deck (my company trains the cabin crew to do exactly that if the descent has not commenced within 80 seconds). However even had the cabin crew entered the flight deck would (could?) they attempt a radio call, once they removed their mask there wouldn't be much time of useful consciousness to get that call out even assuming they knew how to.
If I had access to satellite imagery I think I might extend the search out in the last presumed direction of flight to the fuel exhaustion point and see if anything was there.
Possible? Plausible?