There are three basic ways to fly a Boeing 777. The first is traditional hand-flying: manipulating the yoke to change pitch and bank. This is a fun way to fly a plane in a simulator, but in practice, pilots don’t do this in commercial jets at altitude because it is inefficient and requires too much attention. Which leads us to the second way to fly the plane: by using the autopilot to fly from waypoint to waypoint. This is called LNAV, for lateral navigation, and is the way that pilots generally fly commercial routes. The path from airport to airport consists of a series of such waypoints, which the pilots plug into the flight management computer so the plane can fly itself.
The third way is to use the autopilot, but in a different way. By selecting HDG (heading) or TRK (track) hold, pilots can aim the plane in a certain direction and let it fly straight.
You can use any of these in Microsoft Flight Simulator 2004, the program that MH370 captain Zaharie Ahmad Shah use to practice flying a 777 five weeks before the plane disappeared. You can also use a fourth. By opening the map display panel, you can use your mouse to physically drage the plane to a new location.
Looking at the data recovered from Zaharie Shah’s flight simulator, it seems that he used at least three and possibly all four of these techniques during his session that ended with the plane above the southern Indian Ocean with no fuel.
— At point 3N, the plane is close to airway R467, with a heading that is close to what would be required to reach the next waypoint. It’s possible that at this point he was flying the plane using LNAV.
— Between 5N and 10N, the variable which records the maximum G force that the plane has experienced jumps from 1.4 to 2.2, and the minimum G force descreases from 0.6 to 0.1. These are accelerations that will never be experienced by a plane flying on autopilot, so we can infer that Shah was hand-flying the plane.
— At point 10N the plane is located is 13 nautical miles from airway N877 on a heading of 255.5. It’s been suggested that the plane could have reached this point if it were flying along that airway, and then the active waypoint were changed from DOTEN to a point somewhere in the south. However I have run the simulation in a Boeing PSS 777-200LR and found that if that manoever is performed the plane will enter a left bank of 20 degrees (as observed) and achieve a heading of 255.5 about 6 nautical miles from the airway. It seems impossible, then, that the plane could have been flying along airway N877, and since there is no other route it could have been turning away from, I assume that Shah was flying in HDG or TRK mode, which also make turns by putting the plane in a 20-degree bank.
— Between 10N and 45N, the plane moved a distance that is not commensurate with the quantity of fuel burned. The only plausible explanation is that the plane was moved on the simulator’s map display.
The upshot of all of this is that the recovered flight simulator data tells us that Zaharie Shah did not practice flying a 777 to fuel exhaustion in the southern Indian Ocean on autopilot. Instead, he moved the plane manually to the southern Indian Ocean, altered the setting of the fuel load to zero, and then hand-flew a gliding descent.
Given this understanding, the flight simulator session shows significantly less similarity to the accident scenario than was originally believed. Could its uncannily resemblance to MH370’s eventual disappearance be merely a coincidence? A hallmark of a coincidence is that it’s an outlier, with no other similar or corroborating evidence. And that’s exactly the situation here. Nothing else found on Shah’s computer, or uncovered by the Royal Malaysian Police investigation into his personal or professional life, suggests his guilt. As a piece of evidence, the simulator data looks shaky, and all the more so the closer you inspect it.