A few things have happened recently in MH370 world that are worth taking note of.
No FMT. The seabed search in the southern Indian Ocean is all over but the shouting, and as a result I see that a consensus is forming that there could have been no “final major turn” into the southern Indian Ocean. Rather, if the plane went south, it must have loitered somewhere beyond the Malacca Strait until after 18.40 before finally flying a straight southerly path from 19:40 onward. This loiter, following a high-speed dash across the Malay Peninsula and up the strait, is quite bizarre, given that no attempt was made by anyone on board the plane to contact the ground, either to ask for help or to negotiate a hostage situation. So the presumption of a loiter doesn’t really shed light on motivation, it does effectively put yet another nail in the coffin of accident/malfunction scenarios.
More of the secret Royal Malaysian Police report released. Mick Rooney, aka @airinvestigate, has released a portion labelled “Folder 6: Audio and Other Records.” The new section contains an expert report analyzing the cockpit/ATC audio up to 17:21, which concludes (with less than 100% confidence) that it was probably Zaharie who uttered the final words “Good night, Malaysia 370.” It also includes ACARS data and the Inmarsat logs which had already been released back in 2014. In perusing the document I was not able to identify anything that would alter our collective understanding of the case, but I hope that others will offer their own assessments. And I applaud Mick for being the only one with the moral backbone to release this information. I am sure that more will follow. UPDATE: The next batch is here: “Folder 5: Aircraft Record and DCA Radar Data.”
Debris trail goes cold. I’ve plotted, above, the number of pieces of debris that have been found each month since MH370 disappeared. After the first piece of debris was found in July, 2015, a smattering of further pieces was found until April, May, and June of this year, when the number spiked and then dropped off again before ceasing altogether. This is a puzzling distribution, since drift models show that the gyres of the southern Indian Ocean act as a great randomizer, taking things around and around and spitting them out after widely varying periods of time. Would expect, therefore, to see the number of pieces found to gradually swell and then fall off again.
There is a complicating factor to this assumption, of course. Even if the pieces do arrive in a certain pattern, overlaid on top of this is the effect of an independent variable: the degree to which people are actively searching for them. It must be noted that a considerable amount of the June spike is attributable to Blaine Alan Gibson’s astonishing haul on the beaches of Madagascar that month. Indeed, Gibson by himself remains responsible for more than half of the 22 pieces of debris found thus far.
Earlier this week, several frustrated family members announced that they would be organizing their own beachcombing expedition, to take place next month. If their efforts prove less fruitful than Blaine Alan Gibson’s, it may raise questions as to what exactly was the secret to Gibson’s success.
@VictorI
Is the PSS model good enough to simulate the RAT turbine and end-of-flight maneuverability in the no-fuel situation? As a non-pilot, I envisioned that could be one goal of the simulator work. Also the sunrise information could be relatively useful as you can see when the airplane is in light, but the ocean surface is still dark, and the lighting transitions.
@VictorI, The practice of reporting out stories as a means of earning a living is nothing to be ashamed of. It’s called journalism, and I think it is a necessary part of a functioning society.
If someone wants to hire you to explain what you know about this case to the wider public, you should do so. You have a right to be compensated for your labor.
@TBill: I don’t know. I certainly wouldn’t assume it was accurate except in a very general way. I would trust the PMDG model much more.
@jeffwise: Again you miss the point. I am a free market capitalist, and I see nothing wrong with getting paid for ones labor. But once you falsely accuse me of having an agenda while conveniently neglecting that you have an interest in promoting a particular scenario for financial and professional reasons, you lose all credibility.
I will not accept money related to this tragedy for the reasons I mentioned. You may choose to continue to derive income from it. That’s your choice. But you should think twice before you make false accusations about others in light the choices you have made.
The work I have done has never been about increasing “reader engagement”. I have written very technical and mundane papers that have appeal to a limited audience, but hopefully provide some insights to those that matter.
@VictorI
Thanks for clarifying. I also follow your work on MH370 a long time now. Never saw a reason to doubt your integrity though.
This mist of secretness surrounding aspects of MH370 is just very frustrating sometimes and feeding distrust I think.
Which when pointed to someone can be disappointing, insulting and not motivating at all ofcourse.
Still believe Jeff and also many people here (and else) sincerely have the ultimate objective to contribute to find answers to the mystery and find the plane.
Dispite the distrust and implicit accusations sometimes that should be overcome in sake of the higher goal imo.
@TBill
A SIM program providing sunrise information would be very useful for determining lighting conditions in the landing/ditching area.
@VictorI, You’ve stated your position. I feel it’s self-evidently inane, and leave it at that.
@VictorI:
I have read Yves Guillaume’s “Flight Dynamics in Microsoft Flight Simulator”. In Appendix 2, it states: “Induced drag coefficient (CDi) is calculated by MSFS almost the same way as found in most text books, but with some simplification. Instead of using the true lift coefficient (CL) at a certain AoA, MSFS looks up the CL value from an assumed linear CL_α slope, called LINEAR CL ALPHA”.
Paragraphs 1.a and 1.b describe the Lift curve construction, defined by the “Zero Lift AoA” and a constant slope called “LINEAR CL ALPHA” for the range of AoA’s of interest.
In your paper you write that you have determined that for the B777 simulator model the Zero Lift AoA is 4.2°. Can you also provide the lift curve slope? That would be much appreciated. I understand that the linear CL-alpha may only be used for calculation of induced drag coëfficient (CDi), but it would be close enough for my interest.
Paragraph 1.d of that Appendix discusses the CL Mach factor (R401) and states that LINEAR CL ALPHA does not change its slope with Mach number. Anything you can provide regarding the magnitude of the CL Mach factor for the B777 model would be much appreciated.
@VictorI. Had you been around I would have directed my review to you. It did in fact entail some effort.
On the rpm, yes my mistake; and obvious at that.
As to the dynamic pressure disconnect, it remains conjectural I see.
The explanation about the autopilot is reasonable though unaddressed in your paper.
Your explanation as to the step climb is hardly ‘simple’ in the sense it was obvious. That climb was mentioned in your paper without explanation. It was unclear whether that was supposed to reflect the actual Z data.
There are other questions I have but I will not be taking up your suggestion that I do my own simulator experimentation.
Were it that simple you would not have needed access to one with extensive modelling experience.
As to your motive of “gaining the experience”, similar to my conjectural “getting the feel of it”, I agree that might emerge as the most likely motive.
Thank you for correcting your paper on one point.
I hope you continue to engage here.
@ROB
Ed Baker shows heavy cloud clover March 8 starting at about 22S (if I am picking off the correct latitude) and suggests the plane stayed a little north of that in the end.
http://mh370apilotperspective.blogspot.com/
@ROB
Flightgear flight simulator will replicate accurate astronomical data: you can set date of flight. For super accurate data taking altitude ASL into consideration, as well as refraction and extinction of bodies according to air pressure and temperature, PyEphem is based on JPL ephemera, and is highly recommended.
So with heavy cloud cover it would be difficult to make a proper controlled ditching with the goal of low amount of debris. Something else must have occurred to mh370.
@VictorI
Once (you) conveniently (neglect) that you have an interest in promoting a particular scenario for financial and professional reasons, you lose all credibility.
To be fair to Jeff, he doesn’t seem at all that bothered about any financial or professional gain. He simply has a theory he passionately believes in. Well, that’s how see it anyway. But nice to see you post here again Victor, welcome back!
@keffertje (and anyone else who doesn’t get annoyed with false leads)
Thanks for the link. I’ve no clue how the MAS in-house tracking system works but the article says when they zoomed in it to the plane over Vietnam it stated it was in Cambodia?!
Anyway, the reason why I asked you for the link, some kid posted this YouTube video in May 2014:
https://www.youtube.com/watch?v=AKdpWFv6S14
To be fair, it does look like a plane, it does match a 777 (I guess), and the ‘ground’ around does look kinda ‘burnt.’ I guess what I’m trying to say is that it doesn’t seem like a satellite capture of a plane overflying jungle… It actually looks like a plane that’s crashed in the jungle.
Anyway, these are the Google Map coordinates:
12.089318
104.151699
Seems like Google ain’t updated its satellite image in 2 years. Would be curious to know if anyone can find a more recent image of that particular spot?
@Sajid UK
I greatly appreciate this open forum provided by Jeff, and he is an excellent moderator with superb writing skills and very interesting posts on MH370.
Having said all that, I also believe that if Jeff were to take advantage of the wealth of information he has gathered through this blog to create a documentary on MH370, I would pay good money to see it.
He has gotten good experience in front the camera, and he also has been an executive producer on GRINGO.
So @Jeff Wise
What do you say? Can we expect an MH370 movie written, directed, and produced by Jeff Wise, perhaps for the 4th anniversary of MH370?
@CliffG
Why didn’t I think of that. Great comment!
@TBill
On your interesting link about the cloud cover.
Looking at that first picture and the indicated crash area on it around ~22S IMO that cloud cover would actualy be a good argument that this area would not have been a chosen crash/ditch area.
The picture shows/suggests skies are getting clear again after ~30S.
Which would make that area a better option to crash or ditch the plane.
It would also fit a more sensible flight path to the south instead of a final leg almost straight to the east as shown in the picture.
I add the picture:
https://www.dropbox.com/s/en341kgk6ym1j39/CloudCover0000zMap.jpg?dl=0
@CliffG
Great idea.
Hopefully by then the true script is known though..
But anyway there would be an abundance of tangilizing twists and scenarios to draw inspiration from showing the effort and proces in finding the plane and unfolding the drama and mystery.
Could make a very interesting movie..
@TBill
To add other arguments against the discribed scenario and crash area around ~22S:
The picture actualy does not show a clear sky area around 22S (zoom in).
Being an IR overlay picture it actualy shows a thicker cloud band there IMO.
And IF there was a clear sky area there, there would have been no way the pilot could have seen or found it there coming that distance from the 6th arc and flying straight to it.
So IMO the scenario makes no sence at all with the arguments it uses.
In fact the used arguments contradict this scenario and make a crash area around ~22S less likely.
RE dynamic pressure errors in the recovered file fragments. I’ve added a column for TBill’s Flight file “Approach to IGARI”: https://www.dropbox.com/s/r8hsaoczn4pjh3v/DynPress.pdf?dl=0
@TBill
To explain my understanding of the IR overlay picture;
In the indicated crash area the whiter higher (and colder) clouds are overlaying an area of lower greyer (and warmer) clouds.
The black areas show clear skies by the warm ocean being visible (and also some coastlines).
@TBill
Still the picture is quite interesting stating it was the cloud situation ~20 minutes before fuel exhaustion.
IMO it could give another indication/motivation why the pilot (if piloted till the end) flew on farther to the south untill he found a necessary clear view on the ocean surface.
Which according to the picture only started to emerge south of an estimated ~28S.
That is an incomprehensible remark.
The Log data on Exhibit 26 show that the Flight Simulator X “playgame” was played many times during December 2013, January 2014, and February 2014. At those occasions the game was variously configured as a Douglas DC3, a Boeing B738, and a Boeing B777. If the playgame in the recovered file fragments had been configured as a DC3, B738, or a steam locomotive, would you suspect that the supposed flight path has any connection to MH370?
An important question that has been ignored sofar is what value the simulation could possibly have had as a preparion for MH370. If it was a single flight, why did he save those particular points in the simulation? Why did he change key flight parameters at each point before restarting the simulation? It is important to note that in all recorded points the Autopilot Master Switch and all autopilot, autothrottle, and FMS functions, except the Yaw Damper, were off. As testing the Autoflight features was not intended, what was the purpose of those simulations? In the context of that question, the accuracy of the simulation representing B777 aerodynamics and performance cannot be ignored.
@Gysbreght:
From the perspective of a deniable disappearance, controlling or predicting the aircraft’s last path to the surface regardless of varying external conditions seems like the main thing to worry about for a perpetrator. I.e. he may have wanted to be able to handle the plane’s performance going down, experiencing changing wind and weather and air conditions, i.e. getting it out of trouble that would rip it apart, or he may have wanted to see if and how a plane without fuel would behave. He may have wanted to try to land on water with the plane, if the sim allows for that. He may have wanted to see the night sky and the rising sun and contour of Australia (if reliable). He may have wanted to save a position from which he could do the above at a later opportunity. He might have wanted to get a feel for the the situation at large, to prepare himself mentally and get a sense of whether there could be practical issues, to soothe nervousity and get ahead through enactment. He may for instance have realised that way (as an example) that the cockpit door could be hard to push open due to rigor mortis state bodies outside the door. That he would get the sun in his eyes, that he would have his back to Diego Garcia and Kaaba (I read someone’s suggestion of that). He may have wanted to go through some scenarios to see the figures that the sim turned out as an aid to memory.
I assume that the sim ghost files are no real witness to all the things he may have tried more on the sim, but did not save — so they are not proof of innocence either. In that case they show most of all that he (we believe) was “there”.
But if anything of the above cogs into something…
@Gysbreght
I probably better not comment on this specialized subject but curious as I am I just wonder and ask anyway.
Could it be all AP-functions, auto-throttle etc. where switched off in this simulations because that configurations best simulated the actual situations that occured in flight MH370 around FMT and the end of the flight?
It seems to me just those crucial stages of the flight were saved.
Stages I can image a pilot would like to go through several times in a simulation because they were the most complicated (manualy) flown ones.
Makes any sence?
@Johan: None of your imaginative suggestions address address the points I mentioned:
Why did he save points during the climb out of Kuala Lumpur, changing flight parameters at each point?
You can’t introduce changing wind and weather and air conditions in this simulator game. A plane without fuel flies no different than a plane with fuel, and this primitive simulator video game is unlikely to reproduce all system implications of engine flame-out accurately. There is no need to fly the simulator from Kuala Lumpur to a location in the SIO if you just want practice a flame-out scenario in that particular locationb.
@Gysbreght
Climb out of Kuala Lumpur was also a crucial stage by any means I would say.
Shortly after reaching cruise altitude it disappeared.
@Gysbreght
In my perception the saved simulation points show the three most crucial and complicated stages that also occured in the actual flight.
@Ge Rijn: “Makes any sence?”
No, not to me anyway. I don’t see anything critical or complicated (manualy) flown in any point except the two with empty tanks.
@Ge Rijn: You asked, “Could it be all AP-functions, auto-throttle etc. where switched off in this simulations because that configurations best simulated the actual situations that occured in flight MH370 around FMT and the end of the flight?”
Previously, in a comment to @David, I said,
“If you bothered to install and test the software, you would understand that there is a large part of the flight file that is missing from the recovered data that is related to variables that PSS tracks, including autopilot settings, roll mode, waypoints, and many others. These values supersede other values in the flight file.”
Without more data from the flight files, we don’t know the state of the autopilot.
@TBill, Ge Rijn
The SIM data appears to back up the theory that the pilot of MH370 planned a flight into the SIO, a flight ending in fuel exhaustion. Victor comments that the more he examines the SIM data, the more incriminating he finds it.
On the question of cloud conditions at fuel exhaustion: I don’t think it realistic to assume the pilot could have found out in advance what the weather conditions would be at fuel exhaustion in the SIO on the morning of 8th March. If he had to go looking for clear skies, he would have to meander around and risk being spotted from a ship.
To avoid detection, most of the flight south would need to be under cover of darkness, which clearly rules out the eastern part of the 7th arc, IMO.
Now it just so happens that the great circle path from IGOGU to S37.7, E89.1 for example, on the morning of 8rh March, runs essentially parallel to the sunrise terminator. Can this be purely a coincidence? I will let you decide: imagine you are flying south on this particular path on the Morning of 8th March. Because you are flying parallel to the approaching sunrise terminator, dawn is approaching at practically the same rate as it would be if you were at rest on the surface. 23 minutes before fuel exhaustion, the sunrise terminator would pass underneath you as it progresses from east to west, and it would be light enough for you to spot any shipping in your path (assuming a clear sky or broken cloud) allowing you make a change of course if necessary. When you reach fuel exhaustion at 00:17:30, the Sun will be 4.7deg above the horizon (NOAA Sunrise Calculator) The Meteosat image shows a large patch of broken cloud in this region, at the time in question.
If the flight had been delayed for 15 minutes for example, then the pilot could dump some fuel to bring forward the time of fuel exhaustion by 15 minutes, and still achieve the desired lighting conditions. A flight path that runs parallel to the sunrise terminator, is more easily adjusted in-flight than a path which doesn’t.
@Gysbreght
I take it from you those climb out and FMT points are not complicated flying manualy but IMO it could be crucial that those points were saved in the SIM. Just like the ‘end of flight empty tank’ points.
The coincidence those few saved points reflect/aline with the most crucial 3 stages we all break our heads over for more than 2 1/2 years could be telling IMO.
@All
And as Dennis said the other day, drift models are unreliable.
@VictorI @Gysbreght
But if no one knows the values that supersede the values of the available flight data and SIM data how would someone know the AP, auto-throttle etc. where switched off during those stages or what happened in between
?
@Gysbreght states the AP, auto-throttle etc. were switched off in the saved SIM points.
Can I (we) take this as a fact without knowing the superseeding data we don’t know?
Why is so much missing from the recovered SIM data but only this 3 crucial flight points/stages are saved?
Just staying to the available (SIM) data isn’t it telling all those data are about the most crucial stages of the actual flight?
@VictorI: Each file fragment contains the [Autopilot.0] section that contains the settings of autoflight functionality.
@VictorI: The sections appear in the Flight file in a particular order. The Flight files provided by TBill contain all sections. The only section that defines the state of the autopilot is [autopilot.n].
@Gysbreght: Again, you make claims without knowing all the facts. You should confirm that TBill ran FS9 with the PSS B777-200LR aircraft model. If the PSS values, including those related to autopilot, are missing from the flight files, then he is using a different aircraft model.
@ROB
And as Dennis also said.. unless they indicate a more northern area..
Drifter based drift analizes give a reliable indication of a ~5 degrees area IMO statistically only with as much drifters and time periode possible.
CSIRO, Adrift, Glodal drifter program, use a lot of drifters spanning decades.
Their statistic outcomes are IMO pretty reliable. And they now all point north of the current search area based on the now available confirmed debris finds.
They prove south of ~35S is out of the question by now which is also effectively confirmed by the current search effort.
But your confirmation-bias at 37.7S will stay the same anyway I’m afraid.
Your sunrise-views I appreciated from the start though.
@VictorI, @Gysbreght, @Ge Rijn, I am feeling confused by your conversation and would be grateful if you could help clarify.
The data recovered from Zaharie’s flight simulator, as given in the file “Sim Data.pdf,” contains for each data point a section called “[Autopilot.0]” which contains 21 parameters, including “MasterSwitch,” “WingLeveler,” etc. It appears that the yaw damper was on but all the other autopilot function were off. Do we all agree on that?
@VictorI: As you apparently “know all the facts”, it should not be difficult for you to identify (or provide an example *.FLT file) the section missing from the recovered fragments that contains the definition of the autopilot state.
@JeffW: Yes, that is my point.
@jeffwise: When PSS is installed under FS9, there is a separate part of the flight file labeled [Phoenix Simulation Software], which contains many variables which supersede native variables stored in other parts of the flight file. This section was not included in the recovered flight files. Some of the variables in the [Autopilot.0] section might be valid, but I am sure that most are not. The ones that have the potential to be true are AutoThrottleArm and YawDamper.
I don’t know what flight file @Gysbreght is looking at, but my guess he is confused and making statements based on the wrong aircraft model.
@Ge Rijn
I’m gratified that at least you appreciate my sunrise views. It’s reassuring that I’m not lecturing to an entirely empty classroom, which is something Isaac Newton couldn’t always boast. 🙂
@VictorI: Also the msdn.microsoft.com site for Microsoft ESP SDK, the latest branch on the FS tree, describes the Contents of Flight Files and starts with a list of all sections, both in the order they will appear in a saved flight file, and in alphabetical order. in that list section [autopilot.n] appears between sections [switches.n] and [payload.n].
@Gysbreght:
I was just pouring ideas from said perspective, while trying to grapple with it myself. It is a factor that it is hard for me to get a full grip on what the files could be said to depict. But I understand that he would be too experienced and the sim really not that realistic that much would be of any help to him. So what remains? Building an image to himself of going through with it? A kind of “I will do it” desicion while fumbling with the stations of it? And then calculations?
As to the latter and Kuala Lumpur take-off: he did ask for another FL prior to take-off, and he repeated unnecessarily the FL (but was that to the same ATC?), and he didn’t waste any time turning after blackout, and there was that other flight from KLIA to Europe which he “intercepted” coming back over the Strait, so time and fuel seems to have been a factor. Could the shadow files represent (preferrable, tentative) alternatives of, say, only one or two or so parameters? Assume he started from KLIA five times with different FL and fuel loads, just to make sure it would get him where he wanted to go time, and went back and saved the option/s that best met his criteria. Or saved them to get at it again after lunch. Or to remember what parameters the first/last try had, or (if applicable) simply to make it possible to make five comfortable reruns with different numbers. In that case the shadowfiles would indicate that he made a set of successive tries on each position, but they would not necessarily represent the preferred ones, only the last, first, the one before a break or a bit randomly ones that he decided to save since the figures were less familiar to him, was expected to be adjusted slightly depending on how it fell out etc. What was it RMP called it — “possibility of planning”?
@ROB:
Qute. I can see he was in a hurry.
Could he do that (adjust flight path/be spot on the flight path) without destroying the idea of a straight line flight from around FMT (pingwise)? Perhaps.
@VictorI: Please show us how the section(s) [autopilot.n] look like in your version with the autopilot/autothrottle/FMC in the loop.
@VictorI, You wrote, “Some of the variables in the [Autopilot.0] section might be valid, but I am sure that most are not.” What are the clues that indicate to you that the variables in this section are not valid?
Do you feel that the autopilot was on even after fuel exhaustion?
At points 45S1 and 45S2, the headings are 178.22 and 192.99, respectively; neither of these is consistent with a great-circle flight to McMurdo.
@Rob
“Because you are flying parallel to the approaching sunrise terminator, dawn is approaching at practically the same rate as it would be if you were at rest on the surface. 23 minutes before fuel exhaustion, the sunrise terminator would pass underneath you as it progresses from east to west, and it would be light enough for you to spot any shipping in your path (assuming a clear sky or broken cloud) allowing you make a change of course if necessary. When you reach fuel exhaustion at 00:17:30, the Sun will be 4.7deg above the horizon (NOAA Sunrise Calculator).”
Please lay down what the advantage for a suicidal pilot would be to crash in the dark night, in the twilight or in the daylight. He is intending to die anyway. But as you mention it, lets talk about some points you raise.
– not to be seen from ship
Crews of merchant ships do not monitor the airspace above them, they are interested in water traffic. If they would observe an aircraft flying enroute, they would take no notice of it, would not record its position or heading and would not report it to anybody. Military ships will monitor air traffic, but they do it with radar and are not dependant on good visibility. You cannnot hide from primary military ship radar by flying in the dark.
If you are talking about not to be seen by ships when coming in for the crash, then the twilight zone would not help, as aircraft above the ground level are exposed to the sun and thus highlighted well ahead of the respective ground zero position, depending on altitude. This is especially true if the observers position is in the dark and the aircraft is illuminated by the rising sun. You can test that yourself on any given day with clear weather and during the times of sunset or sunrise. Aircraft and their contrails can be easily spotted.
– to have enough light for a ditching
Descending from higher altitude, thus already in daylight like brightness into the twilight zone creates problems with visibility. As said before, the best way from a visibility standpoint would be to approach with the sun in your back ( from the east heading west) that way the pilot is not blinded by the sun, but the landing area is well lighted. For that the pilot of MH370 would have had to fly to the east first and then reverse course to the west. Not a lighthearted task without engine power.
-to check, wether the landing area was clear of ships.
Absent from the fact that you need a clear sky with no clouds it is questionable to see ships from higher altitudes except in bright daylight. What you see first is the water contrail aft of the boat. But what would the information be usefull for? With two dead engines and your assumed importance of flying along the twilight zone the maneuvering options would be limited to stay hidden from such ship traffic.
– Shah trained to get a feeling for the lighting conditions
That is an absurd assumption. Over his flying carreer he had done and had overseen many approaches in different light conditions icluding the worst ones, approaching on top in bright sun shine and descending into pitch black foggy sunset darkness in the last minute with visibility of 300 meters and less. Shah would not have needed to train or observe any of these multiple possible conditions for a ditch/ crash in the SIO. He had the expierience already.
– moonless night was important
Again a trivial and unimportant point. An observer on ground has no advantage at all in detecting an aircraft at altitude with its position lights off with a full moon against a moonless night. The aircraft itself does not reflect enough moonlight to make a difference. On the opposite bright moonlight is influencing the night sight capability of the human eye, thus less bright objects are harder to detect with a bright moon. This again you can test yourself by observing the nightsky on a moonless night where you will be able to watch satellites pass in orbit with the pure eye alone. Not so on a moonlit night, you will miss many stars.
The same is true if you move the eye of the observer from the ground into some cockpit.
I see it coming, now somebody will say Shah choose a moonless night to see other traffic better, as he was flying dark.
It is also wishtinking to assume that pilots are visually checking the aispace ahead or around their flightpath when flying under instrument flying rules under a current IFR flightplan and under positive radar control. The flightpath is deconflicted by ATC and a visual check would only be done when ATC gives a headsup of some closer traffic. This is even more true when flying during hours of darkness.
@Johan
Yes, he could adjust, and keep a straight course. Keeping to a straight line after FMT is obviously the most efficient way of travelling as far as possible into the remote SIO. But having said that, attaining the maximum distance possible on the available fuel, doesn’t appear to be the primary aim. It was more important to synchronize with required sun angle, and remain as long as possible in darkness. I originally thought that if he took off late, he would speed up from M0.81 to M0.84 after FMT, to make up the time, but I now realized it would be easier for him just to dump some fuel, instead. It didn’t matter how far along the line he got before exhaustion, it was more important to synchronize exhaustion with sun angle, and this was quite easy to do because sunrise occurred at the same time at all points along the line, at least during the last few hours of flight.
The path ticks all the boxes as far as the BTO/BFO are concerned. The path is almost coincident with the DSTG Bayesian most probable path, which reassuring, to put it mildly.