From page 8 of the Report:
N7E103.5 [point of turnaround] @ 17:19 to ~Penang Island @ N5.3E100.3 is 251 miles
From there northwest to N7E95 [where the plane was headed above Indonesia] is 382 miles.
That’s a combined distance of 633 miles taking 70 minutes @542mph at the point of the 18:29 frequency burst.
It’s hard to believe that the plane could have flown faster and gone any further north or west than ~N7E95 — a point that clears Indonesian radar and gives him a straight shot south.
From page 32 of the report:
N7E95 [point of turn south] to S17.5E92 on the 2141 Arc is 1704 miles over 3:12 @ 532.5mph. Fine
Along that same line from S17.5E92 to S25E91 on the 2241 Arc is 522 miles @ 522mph. Fine
However along that same line to the 2411 Arc @ S37.5E87.5 is 887 miles meaning it would have had to average 591.3mph over the last 1.5 hours. That’s not autopilot and it’s also not likely.
But if he veered eastward toward Perth at about the 2241 Arc he could have hit the center of the search area of the 2411 Arc @ S30E97.5 in 1.5 hours after 527 miles at as slow as 350mph.
Either way it appears that a change of speed and/or direction was made at ~22:41 — coming off of autopilot for those changes to be made at that time.
Thus there had to be somebody alive at that time to make those changes to the autopilot.
Everyone else may have died of hypoxia before that but there was a live person making those course changes in the cockpit @ ~22:41 at the end of the flight.