He was taking a direct route to Palau Langkawi, a 13,000-foot airstrip with an approach over water and no obstacles.
The captain did not turn back to Kuala Lampur because he knew he had 8,000-foot ridges to cross. He knew the terrain was friendlier toward Langkawi, which also was closer.

Now let me deal with Diego's request for my present view in light of new evidence.

We know there was a last voice transmission that from a pilot's point of view (POV) was entirely normal. The good night is customary on a hand -off to a new ATC control. The good night also indicates STRONGLY to me all was OK on the flight deck. Remember there are many ways a pilot can communicate distress - the hijack code or even a transponder code different by one digit from assigned would alert ATC that something was wrong. Every good pilot knows keying an SOS over the mike is always an option even three short clicks would raise an alert.

So I conclude at that point of voice transmission all was perceived as well on the flight deck by the pilots.

But things could have been in the process of going wrong unknown to the pilots - Evidently the ACARS went inoperative some time before. Disabling the ACARS is not easy as pointed out. This leads me to believe more in an electric or electric fire issue than a manual shutdown. I suggest the pilots were probably not aware it was not transmitting.

The next event is the turn to the SW in what appears direct Langkawi. As I said in the first post the pilot probably had this in his head already. Someone said why didn't he go to KBR on north coast of Malaysia which was closer. That's a 6,000 foot runway and to put that plane down on a 6,000 foot strip at night uncertain of your aircraft's entire systems is not an option. I would expect the pilot would consider ditching before a 6,000 runway if still above maximum landing weight which he likely was. The safest runway in the region to make the approach was certainly Langkawi - no obstacles over water with a long flat approach. In my humble opinion this 18,000 hour pilot knew this instinctively.

Reports of altitude fluctuations. Well given that this was not transponder generated data but primary radar at maybe 200 miles the azimuth readings can be affected by a lot of atmospherics and I would not have high confidence in this being totally reliable. But let's accept for a minute he might have ascended to 45,000 in a last ditch effort to quell a fire by seeking the lowest level of oxygen. It is an acceptable scenario in my opinion. At 45,000 it would be tough to keep this aircraft stable as the flight envelope is very narrow and loss of control in a stall is entirely possible. The aircraft is at the top of its operational ceiling. The reported rapid rates of descent could have been generated by a stall and recovery at 25,000. The pilot may even have been diving the aircraft to extinguish flames. All entirely possible.
But going to 45,000 in a hijack scenario doesn't make any good sense to me.
The question of the time the plane flew on.

On departing Kuala he would have had fuel for Beijing and alternate probably Shanghai and 45 minutes. Say 8 hours. Maybe more. He burned 20-25% in first hour with takeoff, climb to cruise. So when the turn was made towards Langkawi he would have had six hours or more. This correlates nicely with the immarsat data pings being received until fuel exhaustion.

The apparent now known continued flight until TTFE time to fuel exhaustion only actually confirms to me the crew were incapacitated and the flight continued on deep into the south Indian ocean.
There really is no point in speculating further until more evidence surfaces but in the meantime it serves no purpose to malign the pilots who well may have been in an heroic struggle to save this aircraft from a fire or other serious mechanical issue and were overcome.

I hope the investigation team looks at the maintenance records of the front gear tires - cycles, last pressure check and maintenance inspection. Captain or F/O as part of pre-flight looks at tires. Is there any video at the airport to support pre-flight walkaround? Any damage on pushback? A day after I wrote the original post a plane in the U.S. blew a tire in takeoff and the t/o was fortunately aborted with a burning tire.

Hopefully - and I believe now it is a slim hope - the wreckage will be found and the FDR and VDR will be recovered and provide us with insight. Until facts prove otherwise, I would give the Captain the benefit of respect and professional courtesy.

think the guy from New Zealand was wrong, because it's highly unlikely a plane would travel for hours with a fire onboard visible from that far away.