Disquieting news in the Wall Street Journal today; the paper reports today that the official inquiry into the disappearance of missing Malaysia Airlines flight 370 is riven with disagreement:
Ongoing differences of opinion between five teams of experts that include Boeing Co. and the Australian military have led to search vessels being deployed in two different priority search areas. These zones overlap in some places but in others are hundreds of miles apart, highlighting how efforts to solve one of modern aviation’s biggest mysteries remain little more than educated guesswork. Searchers may only be able to scour around 80% of the probable crash sites before government funding runs out.
For its part, the Australian Transport Safety Board (ATSB) issued a response that essentially confirmed the gist of the WSJ article:
[ATSB chief commissioner, Martin] Dolan said that earlier there had been consensus amongst the five groups, based on the data available at the time, but once the data had been refined, “the results from the methodologies did not coincide exactly. There is no disagreement, just the deliberate application of differing analysis models,” said Mr Dolan.
One would like to think that, nine months after the plane went missing, that the experts would have ironed out any loose threads in their understanding of the plane’s final trajectory. Especially given the fact that Inmarsat scientist Chris Ashton told the BBC program Horizon that the company had cracked the nut way back in March, saying: “The graphs matched, the data worked, the calculation was solved.”
But if we take a closer look at the history of the accident investigation, it’s not surprising disagreements exist. For all the confident press statements that the authorities have released, behind the scenes investigators have always struggled to make sense of the data in their possession. It’s not a matter, fundametally, of a difference in opinion between experts; it’s a matter of inconsistencies within the data sets themselves.
Allow me to explain. As we all well know by now, everything we know about the final six hours of MH370 comes from seven electronic handshakes, or “pings,” exchanged between the plane and a geostationary Inmarsat satellite. Each of these pings, in turn, provides two data points. The first, called the Burst Timing Offset, or BTO, is a measure of how far the plane was from the satellite at any given time; this data is well-understood, reliable and accurate, with an uncertainty of only about five miles. The second, called the Burst Frequency Offset, or BFO, measures the wavelength of the signals and both harder to understand and much more inaccurate, with an inherent uncertainty of hundreds of miles.
Taken alone, each of these data sets provides only a rough idea of where the plane was at each moment in time. But the hope has always been that, if combined properly, they would be indicate the plane’s trajectory and final resting place, in the same way that a line of latitude and a line of longitude can be combined to specify an exact spot on the Earth’s surface.
Unforunately, this turns out not to work in practice. No matter how much Inmarsat and the ATSB have tweaked their algorithms, they have been unable to find any routes that provide a satisfying match to both the BFO and the BTO data.
In its June report, the agency tried to make the numbers gybe by running them through three different types of analysis, none of which made sense on its own but which overlapped in a way that suggested they might somehow be right collectively. In a subsequent report released in October, the ATSB tried a new approach, this time creating one set of routes that fit the BFO data and another set of routes that make sense in terms of how a plane might actually be flown. These areas lay near each other hundreds of miles from the previous search area, but did not overlap much; the ATSB resolved the dilemma by calling them one big area and hoping to search as much of it as possible. This solution is illustrated in the picture at top; the “data error optimization” area attempts to minimize BFO (and BTO as well) while the “constrained autopilot dynamics” ignores BFO and tries to match BTO observations with aircraft performance and autopilot functioning.
Obviously, there must be some gap in our understanding of the Inmarsat data and how it relates to the aerodynamic constraints of a real aircraft. Someday we’ll figure out where we’ve gone wrong; the plane must have gone somewhere, and when we find it, the nature of the shortcomings should become clear. At the moment, my suspicion is aimed at the BFO values, because the algorithm used to explain them are so complicated, and the inherent uncertainty is so large.
We’d have to think carefully about throwing out the BFO data entirely, however, because after all, it’s the only reason we believe that MH370 wound up in the southern Indian Ocean.
@Richard Cole: My understand is that her sighting was around 19:30 UTC, which would make it unlikely that she saw MH370. There seems to be confusion about the actual time due to the conversion from local time. I am curious what the basis is for the time in the KML file.
Richard: ignoring my last late night post with its run-on sentences and general incoherence, I feel it would be good to be a bit more explicit.
While it is true that we indeed don’t know what happened aboard the flight, it is perfectly reasonable to engage the more imaginative qualites of the fatty organ between our ears and speculate as to what happened – within reason. Moreover, this practice is perfectly reasonable, given that this is a blog with little formalization other than that it exists and that it seems to have evolved into a high-level forum with organic checks on more hairball, screwy forms of speculation. The blog has Jeff, its barkeep, to keep things reasonably tidy, while we can couple probability with speculation to ensure that we don’t drift off into insanity. It is all good, as they say, and the blog serves an important function: it is a platform for developing ‘pointing instructions’ for anyone in a position of influence reading of our musings to develop specific points of enquiry.
Does this make sense?
Finding ourselves willing to get naked as well as a bit dirty, we can then stipulate a list of ‘highly probables’ that venture a bit beyond the safe and secure domain of the facts.
1. The aircraft was most likely hijacked; the probability that the loss of the aircraft was due to a mechanical failure is significantly lower. At IGARI, we must choose between a hijacking or a mechanical failure; there is no wiggle room here, you must choose.
2. If the aircraft was hijacked, it is less than likely that the destination for the hijacking was the SIO.
3. It is more probable than not that the analysis of Imarsat data set is valid and that the remains of the aircraft are to be found somewhere in the SIO.
4. If the aircraft was hijacked and the destination for the hijacking was not the SIO, there is an as of yet unknown intended destination. All aircraft take off or diverted/hijacked with an intended destination – full stop.
5. It is highly probable and even reasonable to state that the Malaysian authorities have been substantially less than transparent/forthcoming regarding radar data, Voice Box recordings, the progress of the official investigation, etc.
6. It is reasonable to state that while the US intelligence apparatus was initially highly concerned about the apparent ‘disappearance’ of MH370 and the obvious potential for this event to have involved islamist terrorism, it is apparently now not so concerned.
7. It is reasonable to state that the US has prioritized both its extensive counter-terrorism program in Malaysia and its macro foreign policy obejctive of containment of the PRC in South East Asia over interest in the loss of MH370, which further indicates that the probable hijacking of MH370 did not involve islamist terrorism (itself a foreign policy macro).
The Malaysian authorities did not hijack this aircraft, but they have, it seems, managed to make it disappear by way of thwarting awareness of what happened and leveraging the focus on the bonafide mystery of the location of the remains of the aircraft. Moreover, they had opportunity (delays in reporting, a classified investigation, political and policy dominance); and motive in avoiding charges of incompetence and perhaps malfeasance that could threaten the power of UMNO in general and Hisammuddin (the PM in waiting) in particular.
To anyone focused only on the location science and finding comfort in XYZ-T, I would suggest that you are being duped, conned, seduced by the elegant over-simplifications of scientism and the Malaysian authorities whom have co-opted its reductionist, overly objective nonsense. The loss of MH370 involves much more than locating the remains of the aircraft, and if you believe the remains of the aircraft will provide all of the answers that you seek, I fear you may be sadly disappointed. For despite the hard work, persistence and cognitive capacity of the ATSB team members, those of the IG, Dr. Ullrich and others, there is a very good chance that the aircraft will not be found in our lifetimes and rather end up the subject of a National Geographic special some 50 years hence. You might also want to be aware of the fact that there has already been considerable wrangling of who will take possession of the voice and flight data recorders if and when they are found, while Malaysia is perfectly entitled to take it all home in its very own little black box under the Chicago Covention (1944).
Somebody hijacked the aircraft at IGARI, and the only reasonable thing to assume from here, given that the flight terminated in the SIO, is that this nefarious deed was itself busted…yes, the hijacking was likely hijacked itself, and thus are left with a story with a double plot. Meanwhile, it seems to me that it is nothing less than madness to deny that the Malaysian authorities are hiding behind the skirts of the bitch godess of the location science, with their hands in the air and their stomachs full of fear at what they know of the hijacking at IGARI ever becoming known – by us.
It’s true what they say: 777s don’t simply ‘disappear.’ From here, I don’t know more what I can say, other than that some parties have indeed been acting nefariously, and if that is a dirty word, than I guess I am filth monger.
Dear Rand, 100% with you. One thing I would add to point # 7, it’s very important to the US that Malaysia sign the upcoming, long negotiated TPP trade agreement. As you say, the US don’t want to get involved in the sad domestic issues of mh370. I wonder how cooperative the Malaysian government is with the US on matters of terrorism, apparently the US were horrified when Yazid Sufaat, a 9/11 facilitator and anthrax scientist was suddenly set free and allowed to run a cafe in a local courthouse until 2013 re-arrest for recruiting for Syrian terrorist groups. The US wants Malaysia close to them, and puts up with endless crap from Najib and cronies, as the alternative is disloyal Umno cuddling into the nearest warm wealthy beds–China and Russia. http://en.m.wikipedia.org/wiki/Trans-Pacific_Partnership
Rand:
“To anyone focused only on the location science and finding comfort in XYZ-T, I would suggest that you are being duped, conned, seduced by the elegant over-simplifications of scientism and the Malaysian authorities whom have co-opted its reductionist, overly objective nonsense.”
That is just gold.
And what follows is not aimed at you, but it needs to said (yet again).
To execute a cover-up, which (which is a conspiracy), you need two or more parties to agree.
Yes, people are being duped (that’s a polite description), and if people leave the number-crunching aside, stand back and LOOK at the bigger picture (horror of horrors), it’s quite obvious that Malaysia could not obfuscate at this level, for this long, without cohorts. People may even learn (though their rigid world view prevents them from grasping such a possibility) that while Malaysia sits smack in the middle of this cover-up, it’s not the party driving the bus. All I’m going to say on that now.
As it pertains to Malaysia’s air defense, we have a structure in the form of the FPDA, wherein for the last 43 years, the commander of the Integrated Air Defence System (IADS) has been an Australian officer, not a Malaysian. In fact, Air Chief Marshal Angus Houston held that role (IN Malaysia) from 1999-2000.
Given the structure of the IADS, there is NO WAY the Australian commander would have been out of the loop about what happened the night MH370 flew dark across peninsular Malaysia. NO WAY. And yet, the bulk of the conversation on this board continues to be about Malaysia’s obfuscation, as though there are no other actors in this story (I won’t even mention the ATSB) who could possibly be a party to the cover-up. IADS is glaringly, and repeatedly, left out of the geopolitical ‘analysis'(to the infinitesimal degree that it happens here) as though its existence (and likely knowledge with respect to MH370) is completely irrelevant.
That’s not only a breathtaking omission from an analytical perspective, but quite frankly, it looks (and smells) very disingenuous.
Perhaps the comments of Australian defense expert Peter LaFranchi are worth a re-read.
http://t.co/kWB1nkvXYl
@VictorI – Yes, there seem to be too many contradictions in boat course/aircraft direction/timing to offer any, even potentially, useful information. Should have checked in more detail.
The crunchers are engaged in oceanic pin the tail on the donkey, and just spectating on the myriad of wrangles be it details and issues and assumptions and guesses fills me with pessimism. The search went ahead I believe because it was far too early to concede defeat. When experts don’t agree be afraid.
You need some pretty big hairy reasons to hijack a 777 so whoever did so would have made some big hairy preparations. The location science seems to assume they made none at all and simply blundered off into the dark wild with the lights on. I would agree that if it’s down there it was never intended to be – unless it landed. The hijackers got jumped or they have fooled everyone – or most.
Rand, Lucy, Nihonmama, et. al.
Those of us trying hard to understand the data and estimate the highest priority search area are not being “…duped, conned, [or] seduced by the elegant over-simplifications of scientism”. Unlike the conjecture of the conspiracy theorists, our work does not depend much on knowing or guessing what went on in the aircraft that resulted in the path 370 took. It took some path, and we are focused on finding the aircraft, then the black boxes, and only then grading the conspiracy theory papers. Stand by.
“The hijackers got jumped or they have fooled everyone – or most.”
Matty: Dear God, don’t put people in a business they’re not in.
Judging by past comments from more than a few of the crowd here, it appears clear they cannot fathom (again, because of their world view) that there are people out in the world (read: not from Western countries) who have the training and skills to fly a 777, much less plan and execute a sophisticated hijacking of a commercial airliner. Whether it ended up in the SIO (not holding my breath) or somewhere else.
If William of Ockham, god of parsimony, were alive, he wouldn’t have the beginning of a clue.
Airlandseaman:
“we are focused on finding the aircraft, then the black boxes, and only then grading the conspiracy theory papers. Stand by.
And only then grading the conspiracy theory papers.
Yes Michael. Because “we” (read: the IG) are the teachers here.
Well then, let us (the conspiracy theorists), refrain from impudence, repair to our corners and await our grades.
Nihonmama: I have no problem reading all the theories. Some are quite helpful. We all engage in some speculation. I don’t think you will find any IG folks that would disagree with you about the likelihood of some embarrassing mistakes, cover-ups and side deals. Maybe even some intentional nefarious inside deals. Who can say for sure now? But finding the plane and speculating on what happened are not mutually exclusive. Just because some of us are *also* trying to find the plane doesn’t mean we are naive idiots.
Based on all the mathematical BFO analysis that I have read it seems very clear to me ,that what is clear, is that it is not easy to know the exact heading of an aircraft based on only time (us) and frequency (Hz) measurements. Therefore we must assume we don’t know this and consider all the mathematical possibilites of where this plane might have been at 0:19. UTC based on it’s last measured delay time at 0:11UTC, which we know to be fact.
We must discard this “corrected, derived” value for the 7th BTO time delay (18400us), published by Inmarsat, as speculative (not measured) and dismiss it . When you do this the mathematical possibilites of where this plane might have been comes down to a zone of +/- 50nm min of the 6th arc.
This search zone is Infinitely larger than the current search zone and would include all mathematical possible locations on both sides of the southern 6th arc. It shows just how much of a long shot JACC has in finding this plane along the 7th arc, in the current zone, which is just a drop in the bucket.
When you look at the Inmarsat BTO data that way it looks like JACC’s odds of finding the plane is about as good as me winning the Mega Ball Lottery!..lol
@airlandseaman and @nihonmama: In my opinion, we need both analyses and speculation. The reports that the IG have published have attempted to predict paths with the bare minimum of assumptions where evidence is lacking. I think the IG should not speculate in public statements because 1) The science would get put into the same bucket as the speculation and get discounted and 2) It would be practically impossible to get a consensus within the group as to what the likeliest scenario is.
That said, some of us (e.g, Jeff and I) individually have proposed some very speculative events. (Some have tried to discredit the work of the IG because of their disagreement with our individual theories.) And I think all of us subjectively rank various scenarios. This is what Mike means when he refers to “grading”. Nobody is naïve enough to think that their knowledge is absolute. I don’t think Mike meant it in this light. In fact, I see some very strong statements of approval and disapproval of other posts by non-IG members here.
At this point, there is not enough evidence to connect all the dots without some major assumptions. There are a range of paths that satisfy the scant satellite data available, and only by ranking assumptions or with the disclosure of additional data can we propose a particular path or scenario that has high likelihood of being correct.
And the IG does not have a monopoly on understanding the science, either. Significant work has been done by other commenters here, including but not limited to Richard C. and Bobby.
Back to data, folks.
There is a strong reason why the experts cannot agree. It is very difficult to find an autopilot-controlled route that agrees with the BTO and BFO data simultaneously. The closer you look, the harder it gets.
There are 3 clusters of BFO values that ought to be “clean”: the first set from 16:00:30 (after the startup jitter is cleared out) to 16:12; the second set from 16:55 to 16:56:15 when the plane is still climbing but we have an ADS-B report of the climb rate, speed, and heading; and 17:06 to 17:08, when the plane is in level flight and we have another ADS-B report of the plane speed and heading. What I have left out are values from the 30 seconds after the Log-on/Log-off Acknowledge sequence (which we have been told are bad), the time around 16:29 during pushback (when I speculate that the AFC mechanism didn’t know what to do), and during takeoff, when there are no ADS-B data.
After correcting for the frequency-to-frequency BFO offsets, we can make predictions for the BFO at every time and compute residuals for each measured value. Doing so and combining all three sets together, I find that the peak-peak range among the residuals is only 5 hz for 58 measurements of the BFO. The time span covered by these measurements is a little over an hour. Keep in mind that this calculation combines data taken when the plane was at rest with data taken when the plane was moving at 469 knots.
What are the ranges of BFO residuals from our various flight paths for the period 18:39 to 23:15, when the plane is supposed to have been running on autopilot at constant velocity and altitude? Here’s my reading of what people have put out:
sk999, magnetic track route (with recent revisions): 8 hz
Independent Group, spreadsheet 1: 8 hz
Independent Group, spreadsheet 2: 23 hz
Ulich Maimun Saleh route: 16 hz
These ranges have 6-8 data points (depending on how you roll up the C channel communications). None of them is acceptable. Yes, there is always a chance that there is some jitter in the flightpath to the South not represented in the pre-IGARI data, but I think you would have to throw out more than one point to return to an acceptable range.
I am becoming more convinced that the BFO measurements are very good, and impervious to any conspiracy theory.
Airlandseaman – I suppose the only way to crunch the numbers is as thoroughly as possible and some good work has gone in, that much is obvious. Sometimes though the crunchers appear to be chipping away at the bottom of a dark shaft, and need to look up?
I think it’s given that the more time and energy that goes into a set of numbers the greater the tendency to trust them. This thing hinges on debris, be it on the bottom or floating and there should be plenty of both if it goes in close to mach 1. Mundane ocean surveys turn up metal stuff on the seabed all the time even without taking out the magnifying glass as they are doing now. So far nothing to wave the arms about.
A handfull of data beats a conspiracy theory maybe but I’m past the point of remaining comfortable with the data. Technically competent hijackers restore power to the SDU(in all likelihood)and gift a bag of numbers to the search? Crunchers should be squirming over that but on they go. Petty criminals are aware that switching the wrong phone on at the wrong time can get you arrested, or blown to bits in the case of anti-terrorism. I just don’t see why they(crunchers) would not be nervous that the SDU comes to life as the trajectory skews off the page. It could be argued that as far as we can tell the SDU is all good, but there is a seat of the pants element to that. Standing right back, it looks bad – to me.
I agree with @Ken S that there appears to be significant uncertainty of the BTO data.
I agree with @sk99 that the BFO data is good (and there is less than +/- 1 hz uncertainty).
However, based on the above – it appears that the search region needs to be moved approximately 350 nm north.
The speed for this region is in the range of 455 – 465 knots which is considerably less than the current ~ 505 knots.
In order to reduce the size of this proposed search region, the speed and or direction needs to be determined independently.
There are many material reasons to doubt that MH370 actually flew south.
However, if the BFO has not been modified – there is no doubt that MH370 flew south.
However, if hijackers were able to re-program the compensator – it is possible that an ‘aliased’ south route was constructed , and that MH370 actually flew north.
Please see the link below for details.
https://www.dropbox.com/s/8fxuf1a4qwz2nzb/MH%20370%2020141129.pdf?dl=0
On probability, is it suggested that a group of tech savvy hijackers wanted us to know where the plane went? Is there a bit of “see no evil – hear no evil” in the case of the reboot?
Nearly ten months on there is no tangible evidence of a crash, and mounting indication of sophisticated intervention on the plane. So in 2014 a 777 goes missing and speculating on it’s fate makes you a conspiracist? The crunching though is the Rubik’s cube of aviation mysteries, except one was absolutely solvable. It’s sucked a lot of oxygen from a lot of people but if you take policemans’s mindset then you take the data with salt.
“Inmarsat never questioned whether the pings were originating from MH370… any programmable device can be reprogrammed or cloned to mimic the hand-shaking device found on MH370. Such a device could easily be loaded onto a ‘private jet’ and flown along the flight path that investigators believe MH370 traveled. No one – certainly not Inmarsat – could ever tell whether the plane that left those signals behind was MH370 or another aircraft.”
https://twitter.com/nihonmama/status/471958750547292160
Greg – As Chris McLaughlin said at the time: “the data(north vs south)is ambiguous, it’s not 100%.” One was simply a better fit, but still a bit like slipping on the wedding dress 40 years later. And….they could just be wrong? The reboot on the southern route looks like a dogleg. To the north the plane simply carried on it’s way? If both routes ended over remote ocean would there be a lot less confidence re the search location? Are we guided as much by no northern detection as we are by Inmarsat’s algorithm? They only had to make it onto one of those highways in the sky above India and they were gone. Is there a guy sitting there eating Vindaloo at 3.00am, ticking these things off one at a time as they head over? Or is it more like MH17 – bumper to bumper as they sailed across Ukraine in virtual convoy when they got picked off out of nowhere? As I understand it you mainly need to stay on the highway – speed, heading, altitude etc, to be left alone in a lot of places. Interesting remarks in Jeff’s radar post by the RAF pilot who wasn’t challenged while skirting air boundaries in places he shouldn’t have been? Staying on some beaten track discreetly would be far more straight forward undertaking. One thing for sure, a southern resting place is by far the best outcome.
I get a feeling the spin masters used Inmersat to come up with a southern path to cover up MH370 actually crashed near BITOD. The crash site must have been cleaned up right after eyes focused on SIO.
Nihonmama – I’ve seen it argued that if there is an abundance of northern BFO data out there, could you construct southern flight path data from that, and pump it out there at intervals as you jetted away? It’s counter argued that pre-existing data is irrelevant due to temperature/atmospherics etc? On one hand the data is so touchy that only numbers from that evening and timeslot will do, but then they are extrapolated from very little to cover a huge distance/time and it’s all good? The world would grind to a halt without geeks but when they saw this they went into a daze everywhere. Way back in the beginning I thought, albeit intuitively, that this could all be wrong and I’ll keep saying it – I don’t really want to be right.
@sk999
You are putting the numbers to what Jeff stated in his original post, that is the “data error optimization” models that attempt to minimize BFO error (and BTO as well) and the “constrained autopilot dynamics” models that ignore BFO and tries to match BTO observations with aircraft performance and autopilot functioning. Models 2-4 in your list are in the second camp, the Inmarsat analysis (and the line I have followed) in the first.
We can argue what are the underlying statistical errors of a large BFO dataset (and I agree they seem low), but when only around six values are available the potential variations (at the 90% or 99% confidence level) become large.
The ATSB/JACC was able to secure the budget for one scan of the larger area using equipment that is being used well inside its capability (my impression) so there no sign of short-cuts. The (first) approved search was only ever going to be one pass of whatever the designated area was going to be. If that is unsuccessful then decisions will be needed, but not yet.
No doubt within the Investigation there are advocates of the two camps who argued that the search should concentrate on one area or the other and will claim success if their particular area is finally ‘correct’. As there will be here.
Of course there are many other people working with models in data error optimization class above, not just Inmarsat and me. As if.
“No matter how much Inmarsat and the ATSB have tweaked their algorithms, they have been unable to find any routes that provide a satisfying match to both the BFO and the BTO data.”
I would take exception to that statement. It is not difficult at all to construct paths that fit the BTO and BFO data precisely. The problem is not with those data, but with the assumptions regarding error models and cockpit activity that are thrown in.
To construct a path that fits the BTO/BFO precisely, two assumptions must be made. The starting location on the 19:41 arc and the value of the fixed frequency bias.
The ATSB’s June report used starting locations on the 1941 arc “which were able to be reached from the last radar fix using reasonable flight speeds”. Consequently, for a groundspeed of about 500 kts the latitude on the 19:41 arc could vary between approximately 15 degrees north and 3 degrees south. The October update is based on the assumption that the turn south occurred before 18:40, close to the extended primary radar trace. That narrows the range of latitudes on the 19:41 arc, to between approximately zero degrees for a turn south at 450 kt at 18:39, and 2.2 degrees south turning at 500 kt at 18:29.
The autopilot does not impose any constraint. Any path can be flown with or without the autopilot. The basic autopilot modes only become constraints if no changes affecting autopilot operation have been programmed in the FMS active flight plan, and it is assumed that there is no cockpit activity.
While I’m ready to accept the possibility that the cockpit became ‘unresponsive’ at some point in time, I don’t think there is a basis for assuming no activity after 18:25 or whatever.
I am happy that my missive surfaced some previously unexpressed tension, but a bit bummed that I somewhat clumsily stepped on some rather big toes.
Airlandseaman: please forgive me, I did not mean to imply that either your thinking or your work or that of the other members of the IG re the location science was in anyway incomplete or lacking. I intended, rather, to surface the shadow of scientism and how it can sometimes ‘flatten’ perspectives that are arguably better served by likewise integrating more subjective/imaginative/creative elements. And, truthfully, I rather view the IG as a mostly selfless, compassionate, high-level group of cognitive hot-shots clearly well versed in their various scientific lines. Yours is also a rather singular group: there is no other embodying a similar level of personal refinement in the search for MH370. Honestly, I don’t really know why members of the IG have not simply been invited down to Oz to better assist with the search; sadly, ours is not necessarily an open, social-value driven world.
Meanwhile, we can witness in Nihonmama’s rather visceral reaction the activation of the meaner aspects of the green meme, while Victorl has attempted to integrate (some would say, transcend) perspectives. The overall process of coalescing disparate world views is interesting – and quite vital to the search/crowd-sourced investigation, I would think.
Returning to my central argument, the objective pursuit of analyses of the data set is clearly central to the work of locating the aircraft. Likewise, a more subjective pursuit of highlighting and testing the level of Truth in terms of disclosure of radar data, voice box recordings, etc. would better inform the search (yes, Gysbreght). Speculation, then, is warranted, as it can serve to highlight the location of the cracks in what is clearly a wall of Malaysian official obfuscation. The motive for the clearly intentional effort to cloak what surrounds the ‘disappearance’ of the aircraft on the part of the Malaysians runs from those overseeing the investigation covering their asses for anything running from incompetence to nefarious malfeasance – and perhaps worse.
Meanwhile, it seems that the term ‘conspiracy’ is another dirty word that by some is too easily (and rather derisively) aggregated together with ‘speculation’ and ‘nefarious.’ I believe its a difference in semantics: as far as I know, conspiracy refers to an unlawful and then secret activity known by more than one person. The hijacking of the aircraft was perhaps the work of more than one person. The effort to obfuscate what is known of the circumstances of the flight on the part of the very people heading up the official investigation most certainly involves a conspiracy, and one would hope that heads would roll for it in the interest of Justice. Here, then, the application of the term is wholly appropriate, and I would challenge anyone to say otherwise.
Finally, I would argue that mine is not a wild conspiracy theory. Rather, it is a hypothetical attempt to reconcile the various elements of the flight, all grounded in an acceptance of the objective science that indicates that the aircraft terminated its flight in the SIO.
Funny, while previously quite skeptical regarding previously models that put the plane in Tibet or Kzakhstan or wherever, I now feel that it would make more sense (and be a hell-of-a-lot easier) if a perp had indeed flown the plane north!
A question: would it be safe to say (Mike) that a B777 is indicative of a commercial airliner that is more easily flown with minimal training than most?
Ninhonmama: love your work (baby) to ferret out a big crack in the official communications and reports and the associated timelines; great stuff.
Greg Yorke
Posted November 30, 2014 at 12:05 AM
Hi Greg ~
Since you are using the following information in your analysis, please be aware that ATSB report AE-2014-054, page 31, describes “sister ship flights departing Kuala Lumpur on the same day as MH370” but neither MH021 (Figure 30) or MH009 (Figure 31, a flight number that no longer existed in 2014) ever departed from KUL; flight MH021 currently originates from CDG (Paris) and flight MH009 historically originated from ZRH (Zurich), both having KUL as destination. Both Inmarsat and the ATSB have been notified of this discrepancy.
~LG~
Gysbreght: your perspective on fits for the BTO/BFO data is rather disheartening, to be honest, as your clarification regarding the assumptions inherent to the ATSB analysis has a ring of reality to it.
Question: in your view, what would be the km2 size of a conservatively constrained search area that disregards resourcing issues? And the time involved in searching such an area at the present rate of approximately 240 km2? I guess that is a rather rhetorical question, given that we could do the simple math ourselves, but it would perhaps be better were you to state your own conclusion. Thanks.
Richard,
The core of analysis of data from an experimental system is characterization of the errors. I haven’t seen anyone other than you even attempt to do so for the BFOs, even though we have hundreds of pre-IGARI values plus the MH21 data to work with. The BFO is not magic – it is governed by physics. It appears to be quite well-behaved, with outliers arising from causes that we can identify. There may be other causes, but don’t have access to the Inmarsat database to search further. However, If I were about to spend tens of millions of dollars on a difficult underwater search and someone chose to ignore the BFO values when choosing a search area, I would want to know why – and simply saying “I don’t trust them” would not be an acceptable answer.
@ Rand,
I don’t have a conclusion, just pointing out another way of looking at the problem. What exactly do you find ‘disheartening’ in my ‘perspectve’?
Gysbreght: disheartening in how your perspective implies a ‘trailing’ lack of confidence in the search parameters. My request for your conclusion was in reference to your opinion as to the generally more conservative size of the search area were a more conservative approach in terms of the assumptions adopted.
sk999:
Speaking for myself, and not the IG….The IG has spent considerable time analyzing, debating and characterizing the BFO and BTO errors. The total instantaneous BFO error includes (1) bias calibration error, (2) bias drift error, (3) EAFC model error (several components to this), (4) ±0.5 Hz quantization error, (5) AES TX Offset model error (for horizontal flight), (6) AES TX Offset model error due to vertical a/c motion. I have been working with ATSB and equipment vendors to better characterize, and potentially reduce some of these errors.
The calibration bias error is on the order of ±3 Hz. ATSB stated 145-155 Hz. I think most independent observers would agree the best estimate is somewhere in the 150-155 Hz range. The drift should be virtually zero, provided the AES OCTCXO is at equilibrium. Observations shortly after 1825 may have some drift error, if the pre 1825 outage was due to AES power off for an extended period. EAFC errors are probably <2 Hz and change smoothly and gradually over the flight, not like random noise. Though unconfirmed at this point, the BFO observations suggest that the AES TX Offset algorithm may be the cause of up to ±3 Hz P-P random noise observed on the BFO observations. This is clearly evident by differentiating the BFO values wrt time and filtering for changes that occur in <200 msec. https://db.tt/kzVsLeqf Of course, any vertical motion is not compensated by the AES, and thus vertical motion causes very large BFO changes compared to all the other factors discussed here. This is not really an error, but it must be taken into account to avoid misinterpreting a large change in BFO (such as 001929 and 001937) as an error or horizontal speed changes.
A paper discussing all this has been in process for many weeks, but the jury is still out on some of the details, so delays continue while we work the problems. Hopefully this short blog will provide some insights and assurance the BFO errors are not being ignored.
The only facts we have post 18:22 are contained in the Inmarsat communication log. The logged BTO’s and BFO’s can be used to determine a probable path of the airplane. That path is not the solution of the problem, because there are many sources of possible errors that can affect the calculated path.
However, all those errors are plus or minus variations around the probable value. In other words, those errors, as far as they can be quantified, define the width of the probability distribution around the most probable path, but do not change that path itself.
So those errors can be used to argue, for example, that the IG proposition is not in conflict with the data, they do not make that proposition more probable than others.
Rand – My understanding of Mike’s response to Jeff Wises’ post regarding the activities in the EE Bay was the SATCOM could be turned off from the flight deck but that would require disabling just about all onboard electrical equipment. Is that correct? If so, that explanation for the shut down of the SATCOM is improbable? If so, I agree with your scenario and the hijackers were in the EE Bay at takeoff and not a passenger or part of the flight crew.
As mentioned before, the hijackers of Ethiopian Airline Flight 961 wanted to go to Australia and did not believe the pilot when he told them there wasn’t enough fuel to go that far. The plane crashed after fuel exhaustion.
Richard, Victor, Niels – Regarding Kate Tee “not being sure of what she saw” sounds like an accurate description of MH370.
A few months back, in the cruisersforum.com regarding MH370, I questioned the time stamps in the spreadsheet track of the Aaza Dana (Saucy Sailoress – Kate Tee’s boat) because their GPS displayed “local” time not UT. In addition, they were coming from the west and had not yet crossed into the Malaysian time zone. My question from the cruiserforum thread:
“Have you seen that MAP Waves posted two spreadsheets with times and locations? Here is the issue about them:
Row 9106 of the table on sheet 2 shows a UTC time of 18:20:45 and a heading of 353° true as you have shown in your Post #394, above. If “local” time is UTC +7, this should be 1:20:45 AM local time. However, the table on sheet 1 does not list any data for 1:20:45 AM.
Row 9118 of the table on sheet 2 shows a UTC time of 19:20:45 and a heading of 88° true and correlates to 2:20:45 “local” time on sheet 1.
This could mean that the conversion between UTC and “local” time between sheets 1 and 2 was not calculated properly. It appears that MH370 was no where near the boat at 19:20 UTC. The question is whether boat was headed northerly or easterly sometime between 18:20 and 18:40 UTC?”
Stew Stoddard posted a map showing the track of the Aaza Dana with the “Gybe” (the maneuver that happened around the time of the sighting) starting at 19:10 UTC and stated the second spreadsheet that shows the Gybe occurring at 18:10 UTC was incorrect. Note, IF MH370 did fly near IGOGU before heading south (Dr. Ulich evaluated this possibility but found it did not match the BFOs as well as his early southern does) and the times on the second spreadsheet are not incorrect, the Aaza Dana started its 88deg easterly track at 18:20 UT and MH370 would have crossed astern of and from port to starboard of the Aaza Dana. Also, MH370 would have had to have been much lower than 35,000ft to fit her description. Note, in order for MH370 to have flown low and fast in this area it would have had to have climbed to around FL400 for the balance of the flight to meet the 6.13mt/hr average burn rate.
New questions: Payne Stewart’s ghost flight had been cleared to FL390 but ended up at FL489. Why does everyone use a constant FL350 for MH370 at a near constant TAS? MRC (british for Maximum Cruising Range) would have the flight increasing altitude and speed as fuel is burned. I believe Gysbreght said that this combination is possible.
Rand, you really add nothing to this discussion. You just post long-winded speculations and rhetorical/meaningless questions. Please try and take it down a notch.
Rand – Agreed, north makes more sense but these numbers really did become the master. What if the measured BFO were modeled to begin with, dispatched by the plane at intervals and weren’t even real? You would have two models in rough agreement – Inmarsat and the hijackers, and to my less than expert eyes that’s how it looks. The southern leg reboot is a vicious instantaneous dogleg while the reboot on the northern scenario is a smooth flight. The southern route is where a tech savvy, in control group of people suddenly steer the plane to a watery fate. The northern routs has logic on it’s side. What the hijackers primarily would need to do is make sure that the BFO data did not resemble a northern route, and roughly resembled a southern one. Crunchers might read that and wince and say Matty’s at it again, but the location science does seem to assume that the hijackers behaved impeccably throughout and treated the electronics with total respect.
Comes down to what is possible in this day? The crunching became an intense hobby horse, even obsession for a small army of modelers/math experts out there but none of it is viable if you conclude that some spanners got thrown in. I reckon there are spanners in there. As long as they have building satellites the respective intel/spy agencies have been snooping/studying them. Picture whole departments dedicated to that. As a former soldier it definitely looks to me that our experts here have a civilian mindset, and they might say I have a twisted one, but if it’s state terror then it could have been executed from the absolute fringe of what is do-able electronically today.
Jay: it is true that my posts often could use a bit of pruning; Otherwise your snide, derisive, short-form remark is little more than an ad hominem poke not requiring further consideration. At least yo mama apparently taught you how to be reasonably polite in utilizing the term ‘please.’
Matty:
“So in 2014 a 777 goes missing and speculating on it’s fate makes you a conspiracist?”
Isn’t THAT something?
“The world would grind to a halt without geeks but when they saw this they went into a daze everywhere. Way back in the beginning I thought, albeit intuitively, that this could all be wrong and I’ll keep saying it – I don’t really want to be right.”
Impossible to concur more. And hold that thought…
Rand:
“it seems that the term ‘conspiracy’ is another dirty word that by some is too easily (and rather derisively) aggregated together with ‘speculation’ and ‘nefarious.”
+100.
And here’s a dirty little secret: we live in a dirty little world where conspiracies (large and small) — and nefarious (oh no, not that word again) events — occur everyday.
Who knew? Please, somebody tell them…
And since you called me ‘baby’, baby, you get another song:
https://www.youtube.com/watch?v=f4QAqEOq9WI
You two know where you can find me. I’ll be in the conspiracy lounge. The one sitting way in the back, in the dark, sipping a cabernet and eating white cheddar popcorn while watching the people go by.
@sk999
I think that, in the end, the BFO statistical analysis did not give sufficient (statistically) significant constraints so perhaps that’s why it is not described in the reports. The usual thresholds for significance are 99% or even 99.7% (3-sigma), that is statistical fluctuations at the level of 1 part in 100 (or 3 parts in 1000 at 3-sigma) are ruled out (I mention this only for the record).
The problem here is that there are only five BFO data points. If I use an error (normal distribution) of standard deviation 2.4Hz, roughly what is implied for the Amsterdam flight in the Inmarsat paper, then with the 99% criterion the maximum error on one data point out of a set of five will be 6.8Hz and the maximum r.m.s. error contribution 4.1Hz. Also, the maximum error on the mean is 5.6Hz (peak to peak), which would look like an error on the bias term. This is before systematic effects like the bias term gross calibration error. So the averaging effect of five points is limited.
Therefore at a justifiable level of significance a large track uncertainty is induced just by statistical effects and the BFO doesn’t give a definitive constraint (in terms of position along the 7th arc). Of course the 18:40UT southern course requirement does impose an additional constraint.
Would it be statistical heresy to say that a path with an r.m.s. error of 10 Hz has a lower probability of being correct than a path of zero error?
The ATSB Update says:
“For the data error optimisation method, the paths were weighted according to the root-mean-square (RMS) error of the BFO values at each arc crossing. The best 100 paths selected were coloured by probability at 0011.”
Why not take “the best”, the path with zero RMS error?
Clearly the higher probability area is only part of that larger search area. But using a low level of significance for setting the search area would have made it a bet.
(a path with zero error would suggest a problem with the model)
@ Richard Cole,
“(a path with zero error would suggest a problem with the model)”
Please explain. What problem, I don’t see it. I have shown that constructing a path with zero errors is quite straight forward.
I think you and I have been here before…
If the data has random errors, which this data clearly has, then even if we knew the absolutely correct path from other knowledge the measured values would be scattered around that path. The rms error would be roughly the standard deviation of the error (the 99% limits on the rms error in this case are 0.8-4.1Hz, same assumptions as before). If the error is less, then the model is over-fitting the data – it is fitting the noise. Bobby Ulich gave an introduction to this some threads ago.
@ Richard Cole,
Sure, a zero-error path fits noisy data. That’s not over-fitting, it’s the best you can do with the available data. Adding noise like Dr. Ulich is proposing only makes the data more noisy.
@ Richard Cole (cont’d)
If you don’t like the small variations in heading and speed that result from the zero-error path, you can smooth them out and attribute these varations to noise, wind, magnetic variations, etc.
Nihonmama – If the plane isn’t in the box we clear the table and all sit there and look at each other basically. I don’t believe the numbers could yield anymore than they have but round and round it goes. If it goes round one more time they’ll ring the lap bell.
While the hand wringing over BTO and BFO is interesting (and challenging) in my pre-retirement working life we often referred to this behavior as the “talking dog effect”. That is the amazement that the dog could talk overwhelmed the sensible question of whether or not the dog was saying something intelligent.
Modeling the flight behavior of MH370 using just BTO and BFO and autopilot assumptions implicitly ignores the fact that the aircraft was not seen by the radar on the Northern tip of Sumatra when it should have been easily seen. It ignores the fact that no debris has been found. Lastly, it ignores completely the notion of motive or, in the case of a failure of some kind, causality.
While the math is indeed interesting and beautiful it does suffer from a serious negative attribute – the airplane is probably not where those analytics suggest it is. The reality is that if the plane were being actively flown by a pilot, it could lie just about anywhere on the final ping arc in the Southern hemisphere.
My own take, expressed elsewhere, is that a depressed (over marital difficulties) and angry (over Malay political issues) Captain Shah wanted to make a statement by landing the aircraft on Christmas Island, haven to refugees, and having a current population that is 70% Chinese – mostly refugees. There was no intention to harm anyone. Only a desire to show the world, and his wife, that he was more than just a timid little aviation geek.
His flight simulator played into this scenario nicely. He could model a difficult landing virtually out of fuel. Unfortunately the fuel consumption simulation was flawed (because he was flying low to avoid radar), and he ended up dropping the plane into the ocean South and bit East of Christmas Island (which happens to lie on the final ping ring). As chronicled elsewhere, it would be easy for Shah to lock the FO out of the cockpit, and simply tell the passengers he was diverting the flight, and that they would be able to carry on unharmed if they behaved themselves.
This location has not been extensively searched for debris, and there is a current there that could carry any debris into the Indian Ocean toward the Southern tip of Africa.
The satellite data is really the only data we have after the last radar blip at 18:22 UTC. I have seen no better interpretation of that data than an end point in the SIO in the general area of the current search zone.
However, as time goes on with no debris found, the probability of the end point in the current search zone decreases. It becomes increasingly necessary to re-examine the BTO and BFO models and question every assumption therein. I don’t mean refinements of the model that shift the predicted end point by 10 km. We have to ask ourselves if we have made a more fundamental mistake. I see no evidence of that, and indeed the satellite data combined with the range estimates, autopilot modes, and operational characteristics of the plane all seem to be more or less self-consistent with the current search zone. However, with no debris found,we have to be open to the possibility of a fundamental error in our models and consider all options.
I am not suggesting we abandon the science as some might propose. Rather, we have to scientifically question the models and assumptions.
I don’t think Bobby Ulich was suggesting adding noise, he was examining the process by creating a noisy dataset and then fitting it.
The aim of modelling is to separate the underlying trend from the noise contribution to the data, so fitting to the noise is absolutely not the best one can do. The relevant models are those that have some chance of being correct and the modelling aims to judge that probability. A model that fits exactly to the data has by definition zero chance of being correct, so isn’t a good place to start.
OK Richard, let’s agree to differ and leave it at that.
If I may paraphrase your last sentence: “A model that fits exactly to the data has by definition zero chance of being correct, so isn’t a good place to start”,
IMHO a model that fits the data has the highest probability of being close to the actual path, and therefore is a good place to start.
A model that ignores the available data has a much lower probability of being correct.
I have a great deal of confidence in the BTO values. The speed of light has been constant over the last century or so of measurements. The BFO is another story entirely. Having spent a lifetime working with oscillators ranging from inexpensive TCXO’s you could buy for 7 yen or so to high end atomic clocks I can say that great caution is warranted. Using the BFO values is certainly warranted in a Kalman sense – it is data, and it does contain information. However, the weight one gives to this data needs to be considered carefully.
No one seems to be terribly troubled by the autopilot assumption, but it is after all an assumption. Things get confusing pretty rapidly if this assumption is not made.