Does the MH370 Captain’s Flight Simulator Prove His Guilt?

Captain Zaharie Ahmad Shah in front of his flight simulator

In the months after the disappearance of MH370, Malaysian police searched for any clues that might suggest that the plane’s captain, Zaharie Ahmad Shah, was the culprit. This would have been the simplest explanation for why the Boeing 777 suddenly went electronically dark and pulled a U-turn forty minutes into its flight, and scarcely a minute after Shah’s voice was heard over the radio calmly telling air traffic controllers “Good night, Malaysia 370.” But to their chagrin, the evidence was slim. Zaharie had left no note. His family and friends had noticed no sign of mental disturbance. There was no evidence of political or religious extremism or of marital discord. He was under no financial pressure. He just didn’t fit the profile of someone who would kill hundreds of innocent people and take his own life in the process.

The police did find,  however, a single piece of evidence pointing at Shah. In his home they found a hard drive that contained a flight simulation program as well as data points created when he saved simulated flights. Five data points recorded on February 2, 2014, were of particular interest. It looked like they came from a single 777 flight that went up the Malacca Strait, passed the tip of Sumatra, then turned south and wound up with zero fuel over the remote southern Indian Ocean. This route so uncannily resembled the flight path deduced from MH370’s radar track and then satcom symbols that it was taken by many as smoking-gun evidence that Shah had practiced absconding with the plane. Some even believe that the flight-sim files could offer clues as to where to find the plane. (Indeed, the discovery of the flight sim files was one of the reasons that the authorities shifted the surface search area in mid-April 2014.)

Last November, the secret Malaysian report detailing these findings were leaked to the public, and last month Australia’s final report on its MH370 investigation, “The Operational Search for MH370,” revealed further details, so now we can more fully examine the data taken from Shah’s flight simulator in hopes of understanding, first, what exactly Shah was doing during that simulated flight, and second, what his motivation might have been for carrying it out.

Here’s what we know.

What Shah was simulating. By examining some of the parameters recovered from the hard drive, we can tell that the five points in question were all created during a flight or sequence of flights. That is to say, either Zaharie could have been saving each file from a single continuous simulation session, or he could have initiated new flights from previously created save points. It’s important to note, however, that the save-points were not made in the course of a single continuously-flown flight, because the fuel levels do not match the distance traveled. It appears, rather, that in between save points Zaharie either manually changed the plane’s location, altered fuel levels, or both.

An especially important point to note is that the save files were not created while Shah was was using the autopilot. None of the save locations is on an airway, nor located between one navigational waypoint and the next. (We have to infer that the autopilot was disengaged, rather than observe it directly from the relevant parameters, because for some reason these parameters were not part of any of the retrieved files–and this appears to me to be something of a mystery, since the Australian report describes four of the data files as “complete.”)

The final two save points deserve special attention. They are located just 2 nautical miles apart in the far southern Indian Ocean. In both data files the plane has zero fuel and zero engine thrust. In the first, the plane is at 37,651 feet and flying at approximately 198 knots indicated airspeed, which is close to the speed recommended in the 777 Flight Crew Operating Manual in the event a plane loses both engines. In the second, the plane is flying much the same way but the altitude has manually adjusted to 4000 feet. In both cases the plane is actually in a climb. The fact that the plane is gaining altitude in both cases is consistent with a pilot who is hand-flying the airplane and so unable to prevent temporary departures from ideal speed and glideslope. In other words, as the plane gets going too fast he pulls the nose up, and if it starts going too slow he puts the nose down. It’s difficult and requires constant attention–the kind of thing that’s fun for a little while as recreation and dreadful if you have to do it for a long time as part of your job.

What Shah’s motivation was. There are many reasons why people carry out simulated flights. When they do, the state of the aircraft at the moment the flight is saved should display certain characteristics that will offer a hint at the user’s motivation. If one wants to hone one’s skill at a particular maneuver, for instance, the saved file should show the plane either carrying that maneuver out or setting up for it. If the goal is to practice for an upcoming real-world flight, one would expect to see the plane flying in a way that conforms with operational practice.

Let’s apply this idea to Shah’s flight-sim files. One theory that has been mooted is that Shah conducted the simulated flight up the Malacca Strait as practice for an upcoming flight to Mecca. Page 99 of Australia’s final report notes: “On the day the simulation was conducted the PIC [Shah] was on a rostered day of leave. The following day the PIC was rostered to fly from Kuala Lumpur to Denpasar, Bali and return the same day. On 4 February 2014 the PIC was rostered to fly from Kuala Lumpur to Jeddah, Saudi Arabia. The first three data points recovered from the simulator were consistent with the route from Kuala Lumpur to Jeddah.”

The problem with this explanation is that an actual flight to Jeddah would necessarily be carried out on autopilot, rather than by hand as the simulator session was flown. It’s possible that Shah was curious to see what it would be like to try to carry out such a long-haul flight with hand steering. It does not seem that he was trying to create a realistic simulacrum of an upcoming flight, however.

By the same logic, it does not seem likely that Shah was practicing the disappearance flight, either, since that, too, appears to have been carried out on autopilot.

So, then, the heart of the matter: what was Shah trying to experience at the two final save points?

One theory is that he wanted to know what it would be like to point his plane into the remote ocean and just sit and wait for it to run out of fuel. But we know that he didn’t do this, because the distance traveled by the simulated flight doesn’t match the plane’s fuel load and burn rate. He got to these end points by manually moving the plane in map mode, not by laboriously flying there. His motivation must have involved doing something at those spots, rather than the process of getting there.

This also rules out the theory that Shah was exploring putting the plane on an autopilot track to Antarctica.

To me, one plausible explanation is that Shah wanted to practice responding to the loss of both engines. This is something that happens on occasion, not always with disastrous results.  On July 23, 1983, an Air Canada 767 en route from Montreal to Edmonton ran out of fuel at 41,000 feet due to an improperly calculated fuel load. Thanks to their amazing airmanship, the pilots managed to guide the plane to a safe landing onto the only possible landing spot, a disused air base near Gimli, Manitoba, that had been turned into a drag-racing strip. In the aviation world, this legendary feat has been memorialized as the “Gimli Glider.”

The second of these save points reminds me of “The Miracle on the Hudson,” the 2009 incident in which a US Airways A320 hit a flock of geese that destroyed both its engines and then glided to a safe ditching. That descent began at 3,060 feet, an altitude similar to the one selected in the simulator.

If it’s true that Shah was practicing emergency procedures on February 2, rather than planning his demise, it must be acknowledged as a freakish coincidence that the simulated flight’s end so eerily foreshadowed MH370’s presumed end. But there are mitigating factors. For one thing, Shah was a flight-sim enthusiast who flew many kinds of aircraft in many locations under many circumstances. Investigators found data files for more than 600 simulated flights on various hard drives in his home. Given that number, it would frankly be surprising if one or two of them didn’t resemble the accident flight in some way.

Also, bear in mind that Shah’s apparent suicide run into the southern Indian Ocean wasn’t his final simulation. On the same day that he practiced engine-out procedure on the 777, he also flew a historical propellor transport, the DC-3. And three weeks later, he played with a Boeing 737. This is hardly the behavior of a man with a monomaniacal obsession with his upcoming demise.

We’ll probably never know for sure why Shah decided to simulate an engine-out descent over the remote southern Indian Ocean scarcely a month before MH370 disappeared. But if we look at the entirety of the evidence collected by the police–and indeed even if we look only at the evidence contained on Shah’s various hard drives–then the flight sim data comes to seem an unconvincing smoking gun.

65 thoughts on “Does the MH370 Captain’s Flight Simulator Prove His Guilt?”

  1. @Susie Crowe: Add that the radar at the Indian air force base at Car Nicobar seemed to have shut for the night-after all, Indonesia is not expected to launch a sneak attack on India. If it had been functioning we would have a slightly better idea of where MH 370 went. However, interceptors are not stationed there and only a helicopter unit is usually present.

  2. I return to my previous point about thorough analysis of simulator data and not cherry picking information to “prove” a particular theory. Another area of the simulator which needs examination is the hardware.Key mapping in FS9 gives the option to reverse items such as yokes, joysticks & throttle quadrants so how these items were set up in FS9 is important.The data values may also be different depending on associated software. My own CH throttle quadrant had complex software to enable compatability & I’m not sure how it changed the key mapping.
    It’s also been brought to my attention that the coms do have relevance. Both points RMP 4 & 5 have a transponder value of 1200 which idicates 1. That it WASN’T switched off 2. It is the standard conspicuity VFR code for North America suggesting FS9 default.
    The coms frequencies will require hours of work to determine which VOR’s and ATC units were set up for active and standby use at each RMP point.
    IF at each point they match the local situation it’ll be more than an assumption that there was intent to remain visible and communicate.I have much else to do in my life but an interesting project for anybody who wants to go through AIP’s checking nav & coms.

  3. RMP3 (the point near DOTEN at which it was turning) also shows transponder on & squawking 1200. Note that it’s on AFTER turn had commenced not before. The frequencies of RMP3 are particularly relevant because because active & standby nav may show intent for next point as should ATC selections. It’s such a good indicator of intent that I may even go through the Indian AIP myself !

  4. Ever wondered why Putin would MH370 out of all the planes that are flying around?

    Perhaps he wanted to send a message to Obama after the sanctions.

    “After numerous rounds of, “Did Putin do it?,” Putin himself decided to send Obama a letter in his own handwriting to let him know he was still in the game. Obama opened the letter which appeared to contain a single line of coded message, 370HSSV-0773H. Obama was baffled, so he e-mailed it to John Kerry. Kerry and his aides had no clue either, so they sent it to the FBI. Noone could solve it at the FBI, so it went to the CIA, then to the NSA. With no clue as to its meaning, the FBI finally asked Marine Corps Intelligence for help. Within a few seconds the Marine Corps cabled back with this reply, “Tell Obama he’s holding the message upside down.”

    Taken and modified Laugh Factory

  5. Correction to my last “AFTER turn had commenced not before.” Should just read after turn had commenced. It was squawking 1200 at points RMP1 & 2 as well.

    @CliffG Took me a moment 🙂

  6. Something to brighten up a Monday morning:

    Go to the Movable-Type Scripts online Vincent calculator, scroll down to the Direct Solution box, enter 07 01 25N, 94 36 08E as the start point, bearing 186 48 25.66, distance 5305132.87 metres, and see where it goes to. 0.0012 arc sec is inch at S40.50!

    S40.50, E88.00 is 173NM downrange of fuel exhaustion UTC 00:17:30. The track passes within 5NM of the Bayesian hotspot bullseye.

    But how did he get to our start point at N07 01 25, E94 36 08? Well, he left the N571 12NM right offset at 18:29 and flew directly toward ANOKO. 20 NM before reaching ANOKO, he redirected his aircraft to S40.50, E88.00 manual waypoint. Flying at M0.84, at FL350, he reach the turning point at 18:35.67, the 7.5NM radius turn took about 1.4 minutes, turning through 81.5deg to align with the manual waypoint. As soon as the turn got under way, he began a slowdown to M0.67, to increase his endurance time by a few minutes to make sure he didn’t flameout before sunrise. 23 minutes later, he increased speed to M0.82 and climbed to FL390 for the journey south, touching the 19:41 arc at S01 02 00, E93 39 00.

    This was a procedure he had practiced on the simulator on 2nd Feb.

    After all this time, we find it was remarkably simple how he did it on the night. All he did was wait till he was within 20NM of ANOKO, then turn south toward manual waypoint South 40.50 degrees, East 88 degrees (two coordinates that would take only a moment to enter into the FMC) and then keep going until the fuel ran out. No other waypoints would be implicated. No one would ever be able to work out where he went.

  7. The drift studies are now shown for what they are. Pure fiction! Fabricated in order to get the ATSB off the hook. The ATSB weren’t allowed for one moment to assume that the Captain had deliberate ly set out to evade detection from start to finish. So he wasn’t found in the area they searched around S38. There can now be only one explanation for that failure; he glided as far as he could after the fuel ran out, taking an evil delight in his undoubted flying skills. From 38,000ft o can glide a long way. And if you know how long it takes for the SDU to reboot, which I think our pilot probably did, then you can time a short steep descent to coincide with the logon to mimic a plane descending out of control. That I think was his intention, and he succeeded!

    The pilot wouldn’t know how much fuel would be left for the APU after the main engines flamed out, so to make sure the SDU rebooted when he wanted it to, he simply isolated the LH AC bus for a moment to interrupt power to the SDU.

    Chasing down BFO errors for the journey south is now seen to be a completely pointless task.

  8. @Rob
    I would agree with you that almost all serious path model efforts to date have assumed either: (1) Unintended ghost flight, or in a few cases (2) Intentional ghost flight with passive pilot (after path was set).

    The third idea of (3) intentional flight has been skipped, in part because a decision was made to simplify the problem to assume minimum flight maneuvers.

    I would say Option 1 has failed (ATSB search), Option 2 boils down to McMurdo path and that may be searched if the search is re-started, and Option 3 might explain other end points on Arc7 (eg; 20-35S ), if the search is resumed.

  9. @Ben

    The good news is that Blaine Gibson should feel safer exploring the beaches of WA compared to Madagascar/Africa. I think it highly unlikely he’ll receive any death threats here.

  10. @Ben:
    “…The findings are also part of the basis of the new search to be conducted by the US group Ocean Infinity due to start in January…”

    I wonder if this positive statement from ‘Perth Now’, could be a hint that a deal between Ocean Infinity and the Malayan government has been agreed?


    Thought Blain Gibson had retired from his MH370 projext? I agree that further death treats against Gibson are unlikely and predict he, or one of his disciples, will find some wreakage somewhere along the SW Aus coast. It would seem that the ‘MH370 crashed in the SIO’ meme needs to be preserved at all costs.

  11. @Jeff Wise:

    Yes, the MH370 script does seem to be going off the rails once again. How people fail to see through this constant media smoke-screen and the bizarre contradictory information provided by supposedly authoritative sources is beyond belief.

Comments are closed.