Controlled Pod Into Terrain episode 6: AA 965

re-uploaded to fix technical issues.

this one features a couple of boomer guys who won't ask for directions because they know where they are, damnit, it's the FMS that's wrong.

as always, Controlled Pod Into Terrain comprises Kyra, the aviation writer who investigates plane crashes, Ariadne, the aviation business expert, and J, who is some kind of engineer.

Episode Transcript

(00:00):
Hello, and welcome to Controlled Pod Interterrain. We are a multimedia podcast about air and

(00:05):
space mishaps, aiming to put them in the broader context of how and why things went wrong.
Now, to introduce myself and my co-hosts, my name is Ariadne. I'm the business and
aviation industry expert. My pronouns are she and her.
Woo.
Ooh, that's new.
It is.
My name is J. I'm the systems and engineering expert, and my pronouns are they and them,

(00:26):
and also a sometimes she/her enjoyer.
And my name is Kyra Dempsey, better known as Admiral Cloudberg, the aviation writer,
and my pronouns are she and her.
Next slide, please.
Today, we're here to talk about this. This used to be a 757, but then some pilots got
lost and then they hit them out. It's a tragic, complicated, and frustrating story, and we're

(00:49):
going to tell it. But first, we have to do some kind of news thing.
Next slide.
One day we'll get music for this.
We are going to get music for this.
Some kind of dum-da-da-da? Whatever. Okay.
So Alaska Airlines has announced a bid to purchase Hawaiian Airlines for $1.4 billion
U.S. dollars. The impetus for this has, quite simply, been Southwest. Southwest has cash

(01:14):
reserves, they have plenty of profitable routes elsewhere, so they can afford to, at cost
or below cost, keep dumping Greyhound busloads full of assholes who saw the White Lotus,
misunderstood everything about it, and went, I would like to also do a colonialism.
Hawaiian did not have these cash reserves, and they were starting to run out of options.
So it became merge or die, and Alaska was probably the least bad option.

(01:38):
Now that merge or die aspect is why this is not like JetBlue and Spirit, at least not
to the U.S. government and the sort of various agencies that are going to have to approve
a merger like this. So Spirit and JetBlue compete with each other in the low cost, or
in Spirit's case, ultra-low cost market. Now these airlines are, for the most part,
not competitors. Culturally, I would say that they do share a common place in the modern

(02:02):
American airline landscape. They are both relatively small and effectively regional
airliners. Both are beloved by their customers.
I mean, calling Alaska Airlines small feels like a bit of a stretch, because they have
over 200 aircraft, not counting even those operated by Horizon and Skywest, and that's
vastly more than the flag carriers of most countries. So it only really looks small next

(02:26):
to other U.S. heavy hitters like Southwest, with its 750 Boeing 737s, just because the
U.S. has such a huge domestic aviation market.
Yeah, that's fair. When you compare them, you have to compare them sort of in the U.S.
by nature to the big three, each of which are in the neighborhood of almost a thousand
aircraft each. Hawaiian, for comparison, has 62 airframes.

(02:48):
Yeah, that's pretty small by U.S. standards.
Yeah. Now, that said, this is not going to be the easiest merger in the world. Alaska
loves to be a single type rating. They love their 737s. If you recall, after the purchase
of Virgin America in 2017, Alaska dumped or returned all of their A320s to the lessor.

(03:08):
Reportedly, a lot of them at a slight loss. Such was their hatred for the superior airliner.
And I'm still mad about that, because I liked flying Virgin America. That airline
had a character.
A Bay Area friend once called Virgin and Alaska the Emily and Zoe Deschanel of airlines.
I don't know why, but the Virgin A321s had the least awful economy seats of any of the

(03:30):
airlines I've tried, and I've tried a lot. I'm old, my back hurts, and they hurt the
least. So I'm a fan of those and less of a fan of Alaska's 737s.
But I mean, let's be real. The thing I really liked about flying Virgin America was the
purple mood lighting.
Yeah, I believe they were unequivocally the most trans airline.

(03:53):
I also liked their safety video that was fairly clearly ripping off John Allison.
Oh, that was a lot of fun. We're gonna live it on up in the sky. I forget how it goes.
So the fleet consolidation plan is gonna be the most interesting thing to watch. Hawaiian
has a mix of A330s. They have a lot of 787s coming on order. Alaska is doubtlessly gonna

(04:19):
try and replace as much of that as they can with 737s. Now, I don't expect them to offer
Honolulu to JFK in the 737, but good Lord, I do expect them to try.
Hawaiian also flies to Japan. Alaska has pretty significant Mexico operations, so international
ops are not going to be a big adjustment for either party.

(04:40):
This is of course- And before Alaska, Alaska-Hawaii Airlines
flies direct from Mexico to Japan. Maybe it doesn't work that way.
This is of course assuming the two airlines have their operations merged, which is not
necessarily a good thing or a cohesive thing.

(05:01):
They say they want to operate them as separate brands and one air operator certificate. So
if they do intend to do that, then they probably need to merge their operations to a significant
extent or else you can't get one air operator certificate.
Yeah, I imagine that sort of administration, maintenance, training, that'll all be consolidated.
I do not think you are likely to fly a United-Hawax-Gawaii plane from Anchorage to Ukiavik, but hey, never

(05:27):
say never.
Yeah, who knows.
Of course, we also have to point out that airline consolidation sucks ass. That's why
flying sucks ass. Thank you.
Having run out of non-contiguous states, will they have to start naming low-cost airlines
after US territories now? I mean, I guess there's probably trademark issues with Virgin

(05:47):
Islands Airways, but I also can't see many people wanting to fly on Air Guam.
I mean, other than people who want to go to Guam, I guess.
Alright, next slide.
A couple weeks ago, the New York Times posted a pretty damning story on the state of air
traffic control in the US right now. It contended that the FAA is basically running the ATC

(06:11):
at its total breaking point. There is a pretty severe shortage of controllers at every level
of the system right now.
Maybe the most damaging part of the whole story is when the New York Times alleges that
99% of facilities are understaffed, to which the FAA rather pissily retorts that actually
only 63% are critically understaffed. And I'm sorry, but this is a real what you should

(06:36):
have said was nothing moment.
Yeah, is that supposed to sound good?
Yeah.
Because it doesn't.
It would not have been a problem for you to say, those numbers aren't accurate, we'll
get you better numbers, or we don't like to discuss specifics. Just, yeah.
Leave them on read.
We're not horrifyingly understaffed, we're only terrifyingly understaffed.

(07:01):
So according to the story, overtime is all but mandatory. Six day weeks with 10 hour
days are apparently not uncommon, as are off-kilter schedules where somebody might get off at
midnight and then have to show up the next day at 10pm. Everyone is tired, everyone is
miserable.
And this is not something you can solve quickly. People you hire today are not going to be

(07:26):
behind a radar display for over a year, and it's not going to be at a busy high density
airspace control for even longer than that because they need literal years of experience
before they're ready to cope with that.
Yeah. And the bottleneck, the training really is the bottleneck because there are way more
people applying to be air traffic controllers than are being accepted as air traffic controllers.

(07:54):
But there are such a limited number of training slots and a lot of people fail out of that
training. So the number of people who want to be controllers who actually get turned
into controllers is really quite abysmally low.
So we are running skeleton crews in airspaces that are busier than they've ever been. We've
discussed a couple of runway incursions here, and while the Houston Hobby incident that

(08:16):
we've discussed a couple of times was definitely the pilot's fault, there have been a few
incidents that were not caught by ATC. The article calls out that there were 503 fuckups
that the FAA qualified as significant, which is 65% higher than it was the previous year.
The article also goes on to say that mental health is starting to become a pretty severe

(08:37):
problem and that controllers don't seek help because they're worried about losing
their jobs and their licenses. And the bullshit system that we have that requires that they
request waivers to get stuck at the FAA Doom headquarters in Oklahoma for months, while
they're not sitting in front of a console making sure that planes don't have a boopsy
doopsy.
And it's weird how all these same things keep coming up when we're talking about the FAA.

(09:01):
Odd that. I mean, presumably look for us to come to you with a story about airline catering
chefs all coming undone because the FAA won't let them seek mental health treatment.
I mean, that's surely how we're going to respond to the epidemic of poop plans.
I mean, you say that as a joke, but having worked in a commercial kitchen, literally

(09:26):
everyone there has mental health issues.
No, and the controllers have spent years trying to raise the alarm to FAA leadership about
the staffing problems, the old and rotting billing that we're shoveling these controllers
into that they have to share their workspace with black mold and rats.
Broken elevators, broken air conditioning, infestation with literal fucking bees were

(09:50):
all claims that were substantiated in the article.
As always, we can blame this as we can blame all things on Reagan, who scared the controllers
into never using their leverage to seize the means of positive airway control.
Nice.
The FAA has no leg to stand on. They know they're in the wrong on this one, so their
defense is always the same. Sight the length of time the US has gone without a fatal accident

(10:13):
as if that means shit about safety on flights today, tomorrow, and so on.
That argument will look great if there's an accident tomorrow, I'm sure.
We don't know how to fix this. Obviously, it's going to take a lot of money, it's
going to take a lot of effort, and it's going to involve the FAA, excepting that it is not
the 70s anymore.
I mean, we in particular obviously don't know how to fix it because we're just a bunch

(10:35):
of queers with a podcast.
Exactly.
Facts. Next slide. So there have been some developments in a baffling, and we think you'll
agree, fucking hysterical evolving story. So due in, let's say large part, to Russia's
illegal, evil, and immoral invasion of Ukraine, Russia has been totally unable to import any

(10:58):
jets or parts due to sanctions.
So they have been importing parts, just not legally.
Yes. It goes further than that. They can't send the ones they have to Lufthansa Technik
for maintenance due to the fact that they will be seized the instant they touch down
in an unfriendly country.
Which is all of them, pretty much.
Which is all of them.
Most of them, yeah.

(11:20):
So rather than just do the right thing, retreat back across the border, maybe throw in Belkarad
as an apology present, Russia has decided to resume production of the Tupolev Tu-204.
And just to clarify, the jet never really went out of production. They had just sort
of slowed the line to about one a year. I think there were some years they didn't produce

(11:41):
any, but it was technically still going. And now they're targeting 10 to 12 or more per
year now. So this is technically a ramp up, which was, but that's really still not a lot
of airplanes.
Okay. So they were keeping the production line going to keep everyone from stealing
everything that wasn't nailed down and to stave off brain drain. But this is for all

(12:01):
intents and purposes. Restarted the production line.
Yeah. I mean, it's also not quite-
It just didn't work.
What?
Trying to stave off brain drain definitely did not work.
So, but also this, okay, this isn't quite exactly news. So they stated back in the first

(12:23):
quarter of 2022, when the sanctions first went in, that they were going to ramp up production
of the Tu-204 slash Tu-14. But what is news is that they're going to introduce cargo and
shortened variants as well, which I'm very interested in.
Sure. Fuck it. Why not? I mean, there has to be a booming market for cargo planes that
could only fly to and from non-sanctioned countries.

(12:46):
I mean, literally, yes. And by the way, their target of 10 to 12 airframes per year is so
abysmally low compared to the needs of Russian airlines, which are estimated to be probably
hundreds of aircraft. So this is really a drop in the bucket. And demand for this plane
is actually going to be massive, despite the fact that it sucks for reasons we'll get into.

(13:07):
So this is a largely, in fact, almost entirely domestic jet, which means that Russia can
source every part it needs, either internally or from its handful of dwindling allies. Obviously,
there always exists the option for black market import parts, but those are outrageously expensive,
and I cannot stress enough how much your Middle Eastern parts broker does not want your fucking

(13:28):
rubles right now. So instead, they're going to build a lot more of this piece of shit.
Now, interestingly, since this jet is from 1989, it would presumably make it one of,
if not the last aircraft the Soviet Union ever designed. It is ostensibly designed to
compete with the 757. All I can say is LOL, LMAO even. Really? Really? So what they're

(13:59):
going to do is they're going to build these, and then they're going to give them to Russian
airlines, and then, question mark, profit? So these planes are terrible. Compared to
modern airplanes, yes. Compared to modern airplanes, they still fly with a flight engineer.

(14:20):
And I swear to God, they have announced that they will go to a two-person flight deck in
2026. At which point we can finally say to Russian commercial aircraft manufacturers,
welcome to the 1980s. Yeah, so we even wanted to superimpose this picture of the cockpit.
I don't know how you would describe that, but I would go with like, Airbus, but over

(14:44):
a bad connection? Like, the displays are way too small. It seems kind of asymmetrical.
The handlebars and the rubber boot on the yoke look terrible, and the shade of blue
they used for the plastic is just kind of sickly? Yeah, it's mental hospital. J
described it in the host chat as looking like a Soviet bedpan, and despite the fact that

(15:08):
none of us have actually seen one of those, we all agree that that's probably about that.
That's probably the plastic factory they've still got left is the one that made Soviet
bedpans. Right, so obviously the biggest problem with this is that they're going to be making
a jet from 1989. Now, that doesn't sound like a problem. The A320 is a couple of years
older, the 737 is decades older, but those have both been extensively updated and retrofitted

(15:33):
over their production lives. But because this is 100% domestic, it is more or less unchanged
from a generation ago. And they had until recently no impetus to actually update this
thing. I mean, they didn't need to. They could just get much better planes from some
other country. Yeah, so they were producing one airframe a year for the Russian Air Force.

(15:55):
And that was like, why would you bother to update it? And you know, the 737 being decades
old is its own problem, which we will discuss extensively in a future episode. And I don't
care what Alaska says, the 737 is a problem. So, so J, I want to dig in a little bit

(16:16):
more. We did the problems with not being able to get parts. So where will that be a problem?
And what's wrong with whatever the Russians can make domestically? I mean, just looking
at this thing, God, it's such a pick me try hard shit tube with two crappy engines. I'm
sorry, I really had to get that out of the way. This thing is just terrible. You see,

(16:40):
the thing is that when this was designed at the end of the Soviet Union, Soviet aircraft
manufacturers did not have the same set of constraints and drives as Western aircraft
manufacturers. They were not in the same business. Fundamentally, they were not in the same business.

(17:02):
And because so much of this thing's design is held over from then and it hasn't been
updated, it is a plane intended to compete in a completely different business, not even
a market. It's just a different thing. Yeah, well, the biggest thing that I want to jump
in and add is that the engines on Russian planes like this are vastly less efficient

(17:26):
than those on modern Western aircraft, which is part of why we keep saying this plane sucks
so much. It guzzles fuel. And that's simply because in the Soviet Union, there was fuel
economy was not a super big concern because fuel was so heavily subsidized by the state.
So as a result of that, there was no real drive to improve fuel efficiency and fuel

(17:49):
efficiency of Russian jet engines still lags behind today. I mean, there's other reasons
for that. Okay, so these guys are sanctioned to hell and back for kicking the shit out
of Ukraine over the last decade, or actually longer than that, but mostly over the last
decade. And they roundly deserve this and more. The C-PIT official policy is that killing

(18:10):
people and taking their shit is unacceptable, you know, just in general. So they can't get
Western stuff anymore. They don't have a whole bunch of the supply chain that's needed to
build modern stuff. For example, they don't have deep sub micron chip fabs, so no advanced
electronics for them. China has them. They have a couple. But for two reasons, China

(18:34):
is not really into helping them. One is that China needs those products at home very badly.
There's a trade war in high technology going on. And China really needs that stuff at home
to be able to compete. The other thing is that China's economy is entirely predicated

(18:56):
on doing trade with the West. And so upsetting the West by giving Russia stuff is not in
China's interest. So they can't do that. And Russia can't afford the imports anyway, because
their economy is in the toilet and no one wants rubles. They don't have the super secret
pixie dust turbine engine alloys that Pratt & Whitney, GE and Rolls are playing with. So

(19:22):
those Kuznetsov turbo fans just aren't as good. Their manufacturing technology isn't
as good. The tolerances on everything they make are 1990s at the very best. And they've
intentionally and deliberately driven a bunch of talent, especially high tech talent out
of the country. So great move, guys, really. This is just joined up thinking at its best.

(19:48):
And having said that, I'm going to be honest, I love Russian homebrew aircraft, not because
they're good, they mostly suck, but because they're different. I mean, in a world of A320s
and 737, seeing a Tupolev Tu-214 would have me pointing with my mouth agape like the soyjack
pre-Gosian may he rip in pieces. So I got to say it, bring it on, do your worst. I want

(20:08):
to see it.
I got to agree, Kyra. I don't know if I would ever fly on one, but I do have respect for
whatever it is. At some point, we'll do an episode about banana Soviet aviation. And
yeah, pour one out for our boy getting domed in a real metal way. The other thing, J,
that I would say about China is that they are trying to get their own aviation sector

(20:30):
off the ground right now with Comac. So I can't imagine that they would be keen to export
parts, sub assemblies, any kind of talent.
Yeah, it's just not happening. It's not happening. And the other thing is, if you've got this
technology for building these large, high efficiency passenger aircraft, then you also

(20:54):
have technology and parts that you need for things like strategic bombers. And obviously,
no one is exporting that to Russia.
Yeah. Okay. Next slide.
Oh boy. So ever since episode two, the three of us have been ironically obsessed with the

(21:15):
BAC 111. So in at least some episodes, we're going to subject you to a thing that we're
calling J's BAC 111 corner, where I tell you the latest ridiculous crap we have learned
about what we have fairly or unfairly decided is the worst and therefore funniest passenger
jet ever made.
I think it's unfairly the worst. It is very fairly the funniest.

(21:39):
Oh yeah, absolutely.
So there was a license built 111. It was licensed built in Romania. It was called the RomBAC
111. They only built the 500 series and the SPAY Mark 512-14 DW engines were produced

(22:02):
under license by Turbomechanica Bucharest.
Was this, are we talking about communist Romania under Ceaușescu that they were doing this?
Or is this after?
This was happening during the fall of Ceaușescu. This was actually happening in 1989 and 1990.

(22:28):
So it wasn't quite that, but still it was an attempt to move in on what they thought
was a burgeoning Eastern Bloc export aircraft market that they could get into. I guess they

(22:50):
looked at the Tupolevs and the Sukhoi and thought, well, these planes are kind of crap.
Even the BAC 111 is better than this. So they tried to pimp it out to these guys in Bucharest
and they did. They built several of these things.

(23:10):
It was the 500, which is that was the ultimate stretched version, right?
Yeah. It was the one that was, and I am not making this up, 22 feet longer than the previous
version.
Yeah. And it was the same one with those same engines with the extremely half-assed water
injection systems that we were talking about in episode two. Remember those? The things

(23:31):
the BAC slapped onto the SPAY 512 with apparently very little thought at all.
And no documentation apparently.
So at some point they must have told someone else how to make them. There must be documentation
that they gave to Turbomechanica because there has to be. There has to be. So I've been trying

(23:54):
to find that obviously because I am pathological and that sound you hear is the plot thickening
and also a pair of SPAYs at 104% N1 obviously.
Obviously. All right. Next slide. The only entry we have for no taps this week is capitalism.

(24:15):
It's our Patreon. We have a Patreon if you want to support the show and Albert Cloudberg's
work. We have Tears Now where you can get access to special benefits and eventually
bonus episodes and movie commentary tracks, which we are hoping to kick off after the
holiday. Even at our entry level, you still get access to our Discord where we hang out,
ship post and talk about airplanes, space planes and pets. Patreon.com slash CPIT. That's

(24:39):
patreon.com slash CPIT. Next slide. We want to give a special shout out to our anonymous
supporter at the $25 fire tetrahedron level. Thank you, donor, for your contribution.
And by fire tetrahedron level, we do mean we will send you an actual physical fire tetrahedron.

(25:01):
And unlike the one in episode three, this one isn't just four triangles.
I mean, it is literally four triangles if you think about it. But this one's real. I
swear, ours is actually 3D. I can confirm that it is very real. I had them made and
I will mail it to you along with a five by seven postcard and a letter signed by the

(25:23):
three of us. Yeah, this very one, in fact, that we have on the screen, I think just went
out in the post. It did. It did. And it was terrifyingly expensive because it turns out
that our anonymous supporter lives very far away.
Yeah, if you wanted to, you could use this tetrahedron for some kind of fire based tabletop

(25:47):
RPG if you wanted. I guess. All right. Next slide. So what even is today's episode about?
Okay, today we're here to talk to you about American Airlines Flight 965, a flight that
crashed in Columbia. So let's talk about Columbia. Next slide.

(26:08):
No, that's the District of Columbia. Next slide, maybe.
No, you see, that's the Space Shuttle Orbiter Columbia.
Next slide. Try again. No, that's Columbia House music.
Next slide. Oh, there it is. Okay. The Nation of Columbia, South America's

(26:34):
continental attachment point. Today we're going to hear it. We're going to talk about
American Airlines 965. This was a flight that was going from Miami, Florida to Calais, Columbia.
It's a routine flight. And while it had some special considerations we will get into, it
was not that out of the ordinary. American Airlines operated this route at this time

(26:54):
using a Boeing 757. There were 163 people on board. Most of the passengers were Colombian
and they were going home for Christmas. The six cabin crew were also all Colombian.
We also have to set a bit of a scene. Columbia has had issues with conflict going back to
the 40s for reasons that are well outside the scope of this podcast. The US war on drugs

(27:17):
escalated that conflict. In the 80s, Ronald Wilson Fuckhead and his hateful dip shit wife
promote Just Say No. But they also have the CIA, allegedly, buy narcotics directly from
Noriega. And these policies poured a lot of gasoline on the fire. And a lot of those Coke
dollars ended up in the hands of anti-government FARC rebels. FARC rebels loved to break shit.

(27:44):
They fucked up the radar at Calais Airport, our destination, a few years before the events
of this episode by blowing it up. So I just want to say blowing up the radar
at your local airport is not in fact praxis. No, it is not. It is not.
So all of the navigation had to be done by the pilots. This meant, crucially, that Calais

(28:06):
Tower only knew of an aircraft's location when it was reported. If a plane was totally
off course, air traffic control would not know it. Go ahead and post it in that fact.
Foreshadowing. Now, while English is the legal language of
aviation, the South American controllers had a very limited grasp of this.

(28:28):
And control isn't a lot of parts of the world know enough English to issue clearances,
and really very little beyond that. And ideally they should know more, but many don't, and
that's just a reality of existing on Earth. Really.
Foreshadowing. Next we have to ask, what is a 757?
My notes here just say we gush about the 757. It's really good, you guys. It's a narrow

(28:55):
body, it's a single aisle, but it's much larger than the 737 and has a lot more room
for activities, and by activities we mean high bypass turbo hands.
It is hugely overpowered for its size. It shares the engines, wings, and the flight
deck, and the type rating with the much larger 767.

(29:15):
The standard engines are the Rolls-Royce RB211, or as you may know them these days, the Trent.
These are also related to the Rolls-Royce Spay engine that we were just talking about.
They're basically a three spool version of the Spay with a much larger fan on the front.
This does mean that our degrees to 111 this episode is one.

(29:38):
Nice. We're not doing this deliberately, it just
happened. It is famously described as a sports car by
its pilots. It can shrug off such hardships as hot, heavy, and high, and all three at
once. While your fat and overbred Bulldog 737 Max
needs seven miles to make VR on a typical summer day in July in Scottsdale, a 757 can

(30:02):
do an unrestricted climb in the same condition. Which really makes the fact that this accident
happened even sadder. I mean, the 757 is, in my view, just the most
beautiful modern airliner, and it breaks my heart that there aren't that many left these
days to say nothing of the tragedy we're about to discuss.
If you come to Hartsfield, there are still very many native species here.

(30:24):
For reasons that will become apparent soon, we need to do an overview of its navigation
systems. Now this particular plane, the one you're
looking at if you have the slides, is November 651 Alpha Alpha.
This aircraft was delivered new in August of 1991, so it was less than five years old
at the time of the accident, and it had pretty new equipment as a consequence.

(30:47):
The main things it did not have are things that we would expect these days.
GPS deeply integrated into the flight management system, RNAV, and ADS-B.
It technically had GPS. I mean, the flight management system used satellite
guidance to determine the position of the plane relative to the stored coordinates of
ground-based navigational aids without actually tuning into them.

(31:07):
But that's not to be confused with the way we think of GPS today, which is like Google
Maps. They didn't have that.
And this plane was also built in 1991, so GPS still had selective availability, and
the only receivers that you could really get were sequential, which meant that they were
both expensive and crap and very, very slow to update.

(31:30):
Over a second per location fix.
In short, if I recall, selective availability was the program that the US had all the way
until 2000 that intentionally made GPS worse so enemies couldn't use it.
Anyway, and because of all that, we need to talk about navigation by ground-based radio
aids.
Next slide.

(31:52):
So your basic radio navigation then consists of NDBs, which are non-directional beacons,
which are just radio beacons that go beep-beep with a particular pattern.
I think it's actually Morse code, so that you can tell which one they are.
There's equipment on the plane called an ADF, which stands for automatic direction finder,

(32:13):
which automatically finds the direction that this non-directional beacon is from where
you are right now.
An ILS localizer is technically an NDB, but most people don't call it an NDB.
They call it the localizer.
Then there's VOR, which stands for VHF Omnidirectional Range, which lets you calculate your bearing

(32:36):
to a beacon at a known position.
By doing this to two or more beacons, you can triangulate your position.
So you can find the intersection of two VOR radials, they're called, and that tells you
where you are and how far you are away from the two, because these two radials will only
cross at one particular location.

(32:59):
It works by either using a rotating antenna, which is the old way of doing it, or by using
an array of antennas that are electronically selected going around in a circle, which is
the newer way of doing it.
This you can see in the top right-hand corner of the slide here.
By measuring the phase difference between two signals being sent out, you can get the

(33:19):
bearing from the VOR beacon to you, the aircraft.
And as I say, by finding the intersection between two VOR radials, you can determine
your location.
Oddly, it doesn't actually measure the range, even though it's called VOR for VHF Omnidirectional
Range, doesn't actually do that.

(33:42):
Then there's TACAN, aka Tactical Air Navigation, which was a military system that planes were
allowed to use.
Measuring it together with VOR gives you VOR TAC, which is a pretty popular way of getting
places.
TACAN's bearing measurement thing is basically exactly like VOR, but it uses UHF instead
of VHF, so it's more accurate.

(34:04):
The distance measuring part of TACAN is basically civilian DME, which works like this.
DME, which is distance measuring equipment, lets you tell how far away you are from a
DME beacon.
It does this by sending a pair of pulses from the aircraft to the beacon.

(34:25):
The beacon sends them back.
You measure the time that the round trip takes, divide by two, divide by the speed of light.
Bish bash bosh, you got your distance measured by this here equipment.
It's very simple.
So you can have VOR DME or VOR TAC or just plain VOR if you're really desperate.
And all three of these systems were commonly co-located.

(34:45):
They still are.
I mean, yeah, but there are a lot fewer of them than there used to be because the FAA
doesn't like that they're expensive to maintain.
Yeah.
Aren't these mostly just being replaced by sort of imaginary GPS waypoints?
Yeah.
Yeah.
Most of them are.
In fact, even these days, even though there's the localizer and the glide slope in an ILS

(35:10):
system, that's a tautology, the essence ILS system, if they have an ILS installed and
it's working, then there's the localizer, which is a non-directional beacon and the
glide slope, which is kind of like a sort of version of VOR that's been turned through

(35:34):
90 degrees.
The thing is that these are very expensive things to maintain.
They have all of these fiddly antennas that need to be kept properly adjusted and all
of this kind of thing.
Next slide, then.
So I used the past tense because GPS is way easier to use, obviously, and it's often preferred

(35:56):
now, especially since there are four, that's four, global navigation systems plus WAS,
which is the Wide Area Augmentation System.
So even if your airport doesn't have an augmentation system of its own to get you that really,
really high GPS accuracy, you still have the satellite-based augmentation system of WAS.

(36:20):
So you can do required navigation performance or RNAV using GPS waypoints, and you can even
do approaches, precision approaches using this now.
So this is very, very reliable, and it's often just better as long as you can still see the
sky.
And if your plane can't see the sky, then I contend that you have bigger problems.

(36:43):
Yeah.
And although it's not used much in certain types of flying anymore, ground-based navigational
aids like VORs and NDBs aren't gone.
It's not like VORs don't exist anymore or aren't being used every single day by lots
of people.
They're just not as popular.
They are very much still taught in private flight school.
Right.
But they're expensive to maintain, and the FAA and their friends in other countries would

(37:08):
like to get rid of them eventually.
Yeah.
And by using GPS for this type of navigation, we mean area navigation or RNAV, which is
GPS-based.
And so using area navigation, you can fly a particular track or a particular approach
to an airport using programmed coordinates that don't correspond to any navigational
aids on the ground, as opposed to our Boeing 757 in 1995, where the GPS isn't good enough

(37:31):
to actually ignore the navigational aids and you have a limited database of possible waypoints
that are literally just ground-based navigational aids because nobody is defining navigation
procedures by made-up GPS coordinates yet.
Next slide.
So the 757 flight management system would let you enter nav points.
It would calculate the bearings to known VRR beacons for those points, and it would calculate

(37:52):
how far you were away.
But it was designed in the late 1980s, so the UX was pretty crap.
The Boeing MFDs were not as good as they could have been, thanks Honeywell.
There are some images here.
Yeah, I sourced a few sort of crap-quality images of what a 1991 757-200 probably would

(38:13):
have had.
Oh yeah, and I see on here the infamous Magenta Line.
Absolutely nothing bad will ever happen to you if you just blindly follow the Magenta
Line.
I don't know if those exact words were ever said, but when this technology was first introduced,
pilots at some airlines were definitely being given training that led to a probably higher
level of confidence in this Magenta Line than it actually deserved because the Magenta Line

(38:36):
just leaves wherever you told it to go to whatever waypoint you selected.
And there isn't like a background map that you can compare it against to make sure that's
actually where you think you are.
Yeah, it's not super great.
Also the image to the right there, the lower right, is how you loaded nav point data into
this thing.
Apparently this needed doing frequently enough that there was a special floppy disk storage

(38:59):
cubby.
And yeah, there's a three and a half inch floppy disk drive with a dot matrix display
above it that says insert disk number one and a hand inserting a disk which is presumably
the Boeing navigation system disk number one.

(39:22):
And apparently this needed doing frequently.
I don't know exactly how long, possibly every time you power cycle the flight deck systems.
I mean if anyone out there knows, I think this image is from a 747-400, but the 757
actually had the same flight management system, so it should be the same.

(39:44):
And the image of the printed circuit board above that is actually the early 1990s.
This one was actually made in 1990, so it's probably the same piece of hardware that was
fitted to this 757.
It was a Honeywell flight deck display controller and it has a 10 megahertz 16-bit processor,

(40:09):
a Zilog Z8002, which is a little bit slower than a Mac SE, which it would have been contemporary
with, and it had less memory.
Of course it was all military grade and rad hardened and stuff because Honeywell, you
know, that's how they roll.
It probably was exactly this.
It was probably built to like a Millstead, probably 1500 or 1700 series.

(40:33):
So, next slide.
So we've got a bit of background, which takes us to our next section, flying to the scene
of the crash.
Next slide.
Okay.
So the date is December 20th, 1995.
There had been a non-fatal runway overrun at JFK Airport earlier that same day, which
fucked traffic around the Eastern seaboard of the US.

(40:53):
So flight 965 from Miami to Cali was delayed by two hours getting out of MIA and it was
dark by the time they left, so the whole flight was at night.
Well, usually the arrival was still at night anyway, but the whole flight was at night.
And let's meet our two pilots, each of whom had plenty of experience, especially on the
type.

(41:14):
So, we will cover crashes where the entire flight deck had less experience than these
guys.
I mean, that said, I wouldn't call their experience levels remarkable, especially not for their
age, but the point is that these guys, you know, they weren't rookies.
So we had Captain Nicholas Tafuri, age 57, former US Air Force pilot with 13,000 flight
hours, including 2,260 on the 757.

(41:36):
And we had First Officer Don Williams, age 39, also former US Air Force with 5,800 flight
hours, including about the same amount on the 757, actually slightly more.
So this was their first flight of the day.
Fatigue was probably not a factor.
Both pilots were former active military duty.

(41:56):
Tafuri flew transports and Williams was an F-4 pilot and instructor.
And in fact, he had been named Air Force instructor of the year for 1985, so he was pretty good.
And both pilots had received exemplary remarks.
They were lauded for their skill and decision making.
They were both in exemplary health.
But Williams had never flown into Kali.
However, Tafuri had done so 13 times.

(42:19):
All of them at night for some reason.
That's just how the schedule worked out, I guess.
So the weather in Kali that night was scattered, excuse me, scattered clouds at 10,000 feet.
The moon was all the way waning with 4% illumination.
Long story short, it was dark outside, dark, dark, starlight.
So for this flight, the captain was pilot monitoring and the first officer was pilot

(42:41):
flying.
American Airlines gave special training to pilots that flew to South America, which we
think speaks for itself.
So here's an extended quotation from the materials.
Quote, flights into Latin America can be more challenging and far more dangerous than domestic
flying or the highly structured North Atlantic European operation.

(43:02):
Some Latin American destinations have multiple hazards to air operations.
At ATC facilities may provide little assistance in avoiding them.
En route and terminal radar coverage may be limited or non-existent.
Mountains, larger and more extensive than anything you've probably ever seen, will
loom up around you during descent and approach and during departure.
Communications, navigation, weather problems, and an air traffic control philosophy peculiar

(43:26):
to Latin America may conspire with disastrous consequences.
This was in the training.
Yes.
Continuing the quote, there are many hazards in this environment, but the greatest danger
is pilot complacency.
From 1979 through 1989, 44 major accidents involving large commercial aircraft occurred
in South America.

(43:47):
Of these 44 accidents, 34 were attributable to pilot error or were pilot preventable with
proper situational awareness.
Foreshadowing.
J, I don't even think this is foreshadowing.
I think a time traveler wrote this.
Yeah.
It's literally possible.
The destination here is Kali again.

(44:07):
Kali airport had a single runway, which was 01 slash 19.
And while the runway is at 3000 feet above sea level, but the mountains around it rose
to as much as 14,000 feet and the minimum sector altitudes in the immediate vicinity
of Kali were in excess of 15,000 feet.
The approach to Kali threaded between these two north to south oriented mountain ranges

(44:33):
that continue more or less straight for hundreds of miles with this big valley in between them.
And that's where Kali is.
And yes, we have drawn that onto the map here.
Flight 965 is going to be coming in from the north and is planning to conduct an ILS approach
to runway 01, which means overflying the airport, turning around and then following the localizer

(44:56):
and glide slope in to the runway from the south.
However, to know where the airplane is as it nears Kali, because there's no radar,
the controller needed the flight crew to report passing a series of ground-based navigational
aids.
Next slide, please.
So once you're in this long valley, there are three important reporting points you need

(45:18):
to overfly and report.
First, the Tuluwa VOR.
Second, the Rozo NDB.
And finally, the Kali VOR, which is also the start of the runway 01 ILS approach.
So for those of you who are watching on YouTube, there is a map here that shows all of these.

(45:38):
If you're not, I'm going to describe in a little more detail where they are.
So the Tuluwa VOR is located at 43 DME, so 43 nautical miles from the distance measuring
equipment co-located with the Kali VOR.
But I should note that the Kali VOR is not actually at the airport, so Tuluwa is only
33 nautical miles away from the actual runway.

(45:59):
The Rozo NDB is at about 12 DME, so only 2.6 nautical miles from the end of the runway,
off the north end.
And so one last time, what Flight 965 needed to do was report Tuluwa, report Rozo, report
Kali, which is 10 nautical miles south of the airport, do a 180 at Kali, and then intercept

(46:23):
the instrument landing system and land.
Right.
So Flight 965 was basically normal for the first couple hours as it neared Columbia.
At some distance, the pilots made contact with Kali approach control, at which point
several exchanges occurred in which small differences in wording began to set in motion
everything that followed.
The next few sections of this story have a lot of back and forth on the radio.

(46:46):
So going forward, when you hear radio dialogue, we'll each be reading a respective section.
So I'm going to be reading Captain Tafuri's lines.
And I will be reading First Officer William's lines.
And I'll be reading Kali's ATC.
Right.
So Captain Tafuri gets in touch with Kali approach control, and he says they are 63

(47:06):
miles from the Kali VOR.
So the controller says...
Roger, clear to Kali VOR, descend and maintain 1,500 feet.
Altimeter 3,002.
No delay.
Expect for approach.
Report to Taluwa VOR.
Now remember, Ramirez, who is the ATC, has no radar.

(47:29):
So he doesn't actually know where these guys are.
All he can do is ask for locations from the crew and then build a mental picture of where
everyone is, although by everyone we mean AA 965, because at the time there weren't
any other planes approaching.
Right.
So Captain Tafuri reads back, quote, okay, understood.
Cleared direct to Kali VOR.

(47:51):
Report Taluwa and Altitude 1,500.
That's 15,000.
3,002.
Is that all correct, sir?
Affirmative.
Thank you.
Now, did you catch that?
It was small, but it was a critical script that started the entire chain of events.
So what ATC is trying to do here is tell the crew that they are cleared all the way to

(48:13):
the Kali VOR.
There is no conflicting traffic anywhere between them and Kali.
He's not telling them to fly in a straight line to the Kali VOR from wherever they are
right now.
But Captain Tafuri reads back, direct to Kali, which is not a word ATC used.
And in a radar environment, it's pretty common to get cleared direct to a fix like that,

(48:35):
like a shortcut.
It's a norm in the US for sure.
But in a non-radar environment, it doesn't work that way.
Even if ATC uses the word direct, which they didn't, you still have to hit every reporting
point, which is why the controller also said report Taluwa.
But the crew interprets this to mean that they can proceed as the crow flies to Kali,
and what Taluwa has to do with anything is unclear.

(48:57):
So now this would have been stupid and a problem, but it would, as you can see from our map,
have been over flat ground in this valley, so nobody would be harmed by going direct
to Kali and skipping Taluwa and Rozo.
However, our second fuckup happens almost instantly after, because Tafuri has previously
entered Taluwa, Rozo, and Kali into the flight management system.

(49:18):
But now, because he's cleared direct to Kali, or he thinks he is, he selects Kali and he
presses the direct to button, which gives him a straight, clofrized line to Kali, and
that also erases Rozo and Taluwa from the navigation display.
GERA We should note that Honeywell actually changed
this logic after the aircraft was built, but the fix wasn't performed retroactively.

(49:42):
So if they'd been flying a newer 757, Taluwa and Rozo would have continued to show up,
but they weren't, so it didn't.
There's no other way that they could have known this, because it's not like the version
of the software on the flight management system is something they check before they fly a

(50:03):
flight.
LUCIE So anyway, now Captain Tafuri is in a situation
where he needs to report passing Taluwa, but it's not on his screen.
And this is going to be a problem, because our third and final factor now comes in about
60 seconds later, when air traffic control tells the pilots that the wind is calm and
asks if they'd like to take a straight in approach to runway 19 from the north instead

(50:24):
of doing the whole overfly and turn around and land from the south thing.
And the pilots say yes, which probably seemed like a perfectly fine shortcut in the moment,
but this was going to really compress the time frame and it's certainly going to exacerbate
everything that was about to happen.
Because remember, these guys were already running two hours behind schedule due to the
delays in the United States, and they wanted to get on the ground.

(50:48):
GERA Also the flight attendants were pushing their
duty time limits.
LUCIE Yeah, and however to do this, to land on runway
19 straight in, they would have to reconfigure the airplane, brief their new approach, including
missed approach procedures, slow down like a parachute, they'd have to do it all in
an impossibly short period of time.
GERA And they were high, so they'd have to descend

(51:08):
like a mannequin thrown out the back of a C-17.
LUCIE They should have absolutely said, unable,
we'll stick to the original plan, thanks.
GERA In fact, the first officer even knew it was
a bad idea, I mean he said they'd have to quote,
grumble to get down, but then he agreed to the plan anyway.
So judgment.
So what they're agreeing to here is a VOR DME approach to runway 19, also called the

(51:32):
ROSA-1 arrival.
So forget everything I told you about the ILS, the instrument landing system they were
using before, because for this approach, they need to hit certain altitudes at certain DME
distances, while tracking the Kali VOR, there's no glide slope.
The initial approach fix, where this approach starts, is the Tuluwa VOR, which we already
talked about, and the ROSO NDB is the final approach fix, that's where you need to be

(51:55):
stabilized by.
So after crossing Tuluwa, they had to be at 5000 feet above sea level by 21 DME, and 3900
feet above sea level, so 900 feet above the airport by a ROSO.
And there is not a lot of time for them to get down that far in the distance they have
remaining, because at this point they're out of 23,000 feet, headed for 20,000.

(52:20):
So Tafuri agrees to this on the radio, he says to the controller that they need a lower
altitude ASAP.
And they're given one.
But in order to do this, these guys are going to need to lose a ton of kinetic energy and
a lot of altitude, so they deploy the speed brakes to lose lift and descend faster.
Go ahead and post it that, you're going to need to remember.

(52:42):
ATC says, Roger, American 965 is cleared to VOR DME approach to runway one niner.
ROSO number one arrival.
Report Tuluwa VOR.
And Tafuri replies, quote, cleared the VOR DME to one nine, ROSO one arrival, will report
the VOR.
Thank you, sir.

(53:03):
To which ATC replies, report Tuluwa VOR.
And Tafuri reads back, report Tuluwa.
And from this exchange, you can sort of tell that the captain is now caught off guard and
confused because he doesn't understand why the Tuluwa callout is happening.
He still thinks he's going direct to Kali.
And importantly, he couldn't have called it out if he wanted to because remember, Tuluwa

(53:25):
has already been erased from the flight management system display.
So Tafuri says to First Officer Williams, quote, I got to give you Tuluwa first of all.
You want to go right to Kali or to Tuluwa?
Because I think he said the ROSO one arrival.
Tafuri says, quote, yeah, he did.

(53:45):
We have time to pull that out.
Foreshadowing.
And no, they don't.
But Tafuri pulls it out anyway.
He looks at his charge and he says, quote, and Tuluwa one ROSO.
There it is.
Yeah, see, that comes off Tuluwa.
So the approach that the pilots probably briefed was the ILS approach to runway zero one.

(54:05):
It's unlikely they've briefed the VOR DME approach to runway 19 because they weren't
planning on doing that.
So Tafuri probably doesn't know off the top of his head that Tuluwa is the initial
approach fix and ROSO is the final approach fix for this new approach that he's doing.
He has to pull out his chart and figure all of that out on the fly as they're descending
like a bat out of hell.
So at this point, for unclear reasons, Tafuri asks air traffic control if he can, quote,

(54:30):
go direct to ROSO and do the ROSO one arrival, which makes no sense to the controller who
thinks that of course you're going to hit ROSO, but you have to hit Tuluwa first.
ROSO is right near the near the runway.
You're not starting an approach from there unless you're in Antonov An-2.
So ROSA isn't where you start an approach.
It's where you finish it.
But if you remember, the controller's English isn't good enough to express all of that.

(54:52):
So he just repeats that they're cleared for the ROSA one arrival and hopes they'll follow
the approach chart, but still wants to do his due diligence.
So he says, report Tuluwa and 21 miles, 5,000 feet.
So Tafuri acknowledges this request, but not the part about Tuluwa, which he's decided
to just ignore.

(55:14):
Now had he taken this command seriously, he would have realized that he couldn't find
Tuluwa on his map.
He's erased it, which might have caused him to announce that he abandoning the approach
or even just taking a step back and thinking, what is he doing for a moment?
But he doesn't do any of that.
So it is dark as hell.
They can't see any lights or anything because they're still too far away from Kali to spot

(55:36):
them.
They need to step back and reassess their approach because this is not really a stabilized
approach anymore.
They're way too high and fast.
The time is way too compressed.
They're getting confused about what the procedure is.
But Tafuri is totally committed.
So he decides to go direct to ROSA anyway.
As an editorial note, don't do this.

(55:56):
Slow procedures, for god's sake.
We don't really know why he decided to do something so cavalier and stupid, especially
considering how experienced he was and how good a pilot he was supposed to be.
But like, just don't do this.
Yeah, so Tafuri decides that he's going to enter the identifying code for ROSA into the

(56:20):
flight management system, or at least he tries to do that.
See, the 757's nav database has the GPS coordinates of every published navigational aid on the
planet, which is one of the reasons American Airlines liked it for South America.
So what Tafuri does is he types R, which is listed on his chart as the identifier for

(56:42):
ROSA, and then he just selects the very first thing on the list.
And normally this would work great because the FMS will bring up a list of beacons with
the identifier R, and it will put the closest one at the top of the list, which should be
ROSA.
But he's going so fast he doesn't stop to realize that the first waypoint on the list
is actually Romeo, a totally separate beacon near Bogota in the completely the wrong direction,

(57:07):
250 kilometers behind them to the northeast.
Romeo, Romeo, wherefore art thou Romeo?
Next slide.
Um, no, Ari, that's your dog Romeo.
Yeah, it is.
Look at him.
Alright, next slide.
Aw, so cute.
Okay, so that's this is the right Romeo.
It's on the top right of this diagram, way, way off somewhere in the distance.

(57:30):
So what Tafuri had done here is actually a pretty easy mistake to make, because again,
on his chart, ROSA was listed with the identifier R, but Romeo also had the identifier R. ROSA
and Romeo both transmitted on 274 kilohertz, and they were both in Colombia and have the
same identifiers.
That means they also have the same International Civil Aviation Organization or ICOW code,

(57:55):
which is the identifier plus a country code.
So the way the designers of the FMS software decided to rectify issues like this was that
they would distinguish the two beacons by giving priority to the more important navigational
aid to and to get the one they deemed less important, you would have to type in the whole
ass name letter by letter.
In this case, Romeo was near a bigger city with a bigger airport, specifically Bogota,

(58:19):
so it gets priority, and to bring up ROSO, you have to type out ROZO.
But Tafuri has no way of knowing this.
And as a side note, the FMS actually comes with lots of prepackaged approaches, so it's
possible to just bring up a list of published approaches to Kali and select the VOR DME
approach to runway 19, and it will plug everything in for you.

(58:41):
But Tafuri is rushing.
He's taking shortcuts.
He just wants to put ROZO on his screen and descend like a bat out of hell toward it.
Had also, if he were preparing this approach in advance with plenty of time, he would probably
have actually read the name of the waypoint that came up and would have kept trying to
typing until he actually got ROZO without even thinking about what he was doing or why,

(59:02):
but he doesn't even look.
He just grabs the first waypoint on the list, he hits execute.
And both pilots should have cross-checked the waypoint and the route, but they're under
immense pressure.
They're now running behind the aircraft in desperate race to get down, so they don't.
They just immediately move to the next thing.
And the autopilot, which is unable to understand that this is stupid, does exactly what it's

(59:24):
told and begins a big sweeping left turn back towards Bogota.
So they lose all reference points on their screen.
There's just a magenta line curving to Romeo on a black background.
And because it's so dark and Columbia doesn't have a lot of lights, they have no horizon
and they don't even recognize that they're banking because they aren't looking at their
attitude indicators.

(59:45):
These dudes are about to be lost as shit.
Next slide.
So these guys are behind the airplane.
They're losing control of the situation, which takes us to our next section.
Next slide.
Okay, so our guys are now flying in the wrong direction.
They're headed east.
And meanwhile, having entered what he thinks is ROZO into the flight management system,

(01:00:05):
Captain Tufuri decides he should probably try to bring up Tuluwa after all, but by this
point they've already passed Tuluwa and in fact they flew right by it shortly before
they turned to the east.
So again, Tufuri is behind the aircraft.
He does not realize how far they have actually gotten.
At this point, air traffic control asks for their distance from Kali and Tufuri replies

(01:00:28):
that they're at 38 DME.
So recall that Tuluwa is at 43 DME, so it's behind them and he should know this, but he
does not appear to realize.
So a few seconds later, sensing confusion, First Officer Williams says,
Uh, where are we?
We're going up to...
Now Tufuri is being asked directly, where in his mental math are we right now?

(01:00:50):
So he tries to place them in relation to where he thinks Tuluwa is, which, okay, that's not
a bad idea.
They just need to figure out where actually is it.
So Tufuri selects, in the FMS, he selects the beacon ULQ, which is the identifier for
Tuluwa, and he says, quote, I'm going to give you direct Tuluwa.

(01:01:11):
You got it?
And then he adds, it's on your math, should be.
But by this point, Tuluwa is behind them and to the left, in approximately their 7 or 8
o'clock position, so when Tufuri plugs it in, the beacon itself isn't on the screen
and the projected path to it just shows a turn looping back to the left, which is obviously
accurate.

(01:01:31):
But Tufuri doesn't realize they've turned east toward Bogota, nor does he realize that
they've passed Tuluwa already, so this looks totally wrong.
He thinks that Tuluwa should be in front of them and perhaps slightly to the right.
He just can't square the circle, his mental math is now completely messed up.
So Williams correctly points out that the FMS is showing a left turn to Tuluwa, so maybe

(01:01:54):
they should turn left.
But Tufuri insists something is wrong, so to check whether he has the right beacon,
Tufuri tunes in to the actual radio frequency for the Tuluwa VOR and he listens to the Morse
Code broadcast from that beacon, and it's clearly spelling out U L Q, which is the correct

(01:02:14):
identifier, so he says, quote, okay, I'm getting it, just doesn't look right on mine, I don't
know why.
And by this point they have left the long straight valley, they're flying into the
mountains, the night is pitch black, there are no visual references, no waypoints are
visible on the FMS display.
This is not a situation you want to be in.

(01:02:36):
Next slide please.
So they're high in the mountains, they know this, they also, as we said earlier, are losing
altitude quickly, and with the throttles at idle and the speed breaks out, they really
don't have a lot of time to get themselves out of trouble before they're descending into
the ground.
So Williams seems to be sort of maybe a little bit aware of this, because he again suggests

(01:02:57):
a left turn back to Tuluwa, that'll get them on track, but Tufuri responds, quote, naw,
hell no, let's press on to...
Williams says, press on to where though?
At this point Tufuri suggests a right turn in the general direction of the Kali VOR,
ignoring Tuluwa.
They're able to identify Kali using the FMS, and Williams dutifully instructs the autopilot

(01:03:20):
to turn toward it to the right.
But by now they're not in the valley anymore, they've actually crossed over the high point
of the mountains over here.
Do you want to draw on your screen?
Yeah, well, just where the arrow is.
Yeah, they're over there, and they're now, yes, very high mountains directly in their
path, but they're not aware of this.
Tufuri calls air traffic control, he again gives them his distance in DME, which is for

(01:03:46):
some reason still 38 miles.
Now, the controller later said he was confused by this because they had reported the exact
same distance almost two minutes earlier, which meant they weren't getting closer,
which that doesn't make any sense.
How are they on approach to the airport?
They're supposed to be going direct straight into runway 19, and they're not getting closer

(01:04:07):
in the space of two minutes?
Doesn't make any sense.
But he didn't have enough English vocabulary to really express why this sounded so wrong,
so he just says to them again, you can land on runway 19.
It's not very helpful.
So now our boys are headed towards Kali, which they have, again, they have managed to identify.

(01:04:29):
They're doing so from a point well east of where they think they are, and they're heading
right for some deep shit because there are some very big mountains in the way, as seen
on this map if you're watching on YouTube.
At this point, Tufuri is now pretty freaked out, you can tell from the CVR, because he
says, quote, we're heading in the right direction, you want to shit, you want to take the one
nine yet?

(01:04:50):
Come to the right, come to come right to Kali, Kali for now, okay?
William says, okay, Tufuri continues.
It's that damn Tuluwa I'm not getting for some reason.
See, I can't get okay now.
No, Tuluwa is fucked up.
Okay, yeah, but I can put it in the box if you want it.
I don't want Tuluwa.
Let's just go to the extended center line of which is Roseau.

(01:05:11):
Yeah, Roseau.
Why don't you just go direct to Roseau then?
Okay, let's...
I'm going to put that over for you.
Let's get some altimeters.
We're out of 10 now.
All right.
So you can see from this exchange that Tufuri is still getting fixated on getting Tuluwa
and Roseau to appear where he thinks they should be.

(01:05:32):
He's not thinking about how they're descending into the Andes at 2,700 feet per minute without
knowing where they are.
And First Officer Williams clearly hasn't appreciated the extent to which Tufuri has
lost the plot or else he'd probably say something.
We're a few seconds from disaster, but it's important to note that this whole time the
pilots were extremely aware of their barometric altitude, but not their radio altitude.

(01:06:00):
Pressure altimeters work by telling you how far above sea level you are, but that doesn't
help you determine how far off the ground you are.
For that, you need a radio altimeter.
And radio altimeters work by sending a pencil beam signal directly down perpendicular to
the body.
It doesn't look ahead.
It can only tell you where the terrain below you is right now, not where it's going to

(01:06:24):
be in 10 or 20 seconds.
They may be at 10,000 feet in indicated pressure altitude, but they're only a few hundred feet
above the mountains ahead and that is dropping precipitously.
Yeah, and also it should be pointed out that the radio altimeter doesn't even come alive
and start displaying any information until you're 2,500 feet above the ground, which

(01:06:45):
for most of this entire period they were more than that.
So it's suddenly right at the last minute that it even starts displaying anything at
all.
So at this point Tufuri reports to air traffic control that they're at 9,000 feet and the
controller calls one last time to ask for their DME distance, but he never gets an answer
because at that very moment the ground proximity warning system goes off.

(01:07:18):
This is old school ground proximity warning system, which is based on the radio altimeter.
It's measuring the closure rate with the ground directly beneath the airplane.
So if you're flying toward a mountain, the GPWS is only going to know something is wrong
when the slope of that mountain is directly beneath you.
Because this is the Andes, the mountain is very steep.
They do not have a lot of time to react to this warning, but they do react very quickly.

(01:07:42):
In fact the official report notes that their actions took less than two seconds.
So they go full takeoff go around power.
Williams starts hauling back on the yoke within, again, within two seconds.
The normal reaction time to a GPWS warning with training is considered to be five seconds
or less.
So under two seconds is very good, by the way.

(01:08:03):
Seriously, get a stopwatch and see how long five seconds actually is.
You'll note you'd be hard pressed to remember your own name under the kind of pressure they
were under.
And these guys had to execute a life saving maneuver.
So Williams is giving it everything he has.
He's holding the plane, holding the plane back right at the edge of a stall.
Tofuri is watching their radar altimeter keep ticking down towards zero.

(01:08:25):
They're pulling so hard, they're activating the stick shaker.
They're actually supposed to do that.
The way they extract maximum climb performance during a terrain escape maneuver in a non-fly-by-wire
aircraft is to pitch up to find the stall warning threshold and then ride it.
Yeah, so that's what they're doing.
The stall warning is going off intermittently because Williams is trying to ride the stall

(01:08:46):
threshold to extract maximum performance.
Next slide, please.
But here's the knife in the back, you guys.
The spoilers, the speed brakes, they're still out.
They deployed them, remember, because they were trying to drop as quickly as possible
to do the direct into Cali 1-9.
Just to remind you, these speed brakes work by adding drag and by making the wing produce
less lift, which also causes it to stall at a lower angle of attack.

(01:09:10):
The report actually calls out that had either of them retracted the spoilers within the
first second of GPWS going off, they would have made it over the top of the mountain
just barely.
The failure to raise the speed brakes is, to be honest, very hard to hold against them.
It was not part of their training for the GPWS alerts, so they just missed it.

(01:09:31):
It was not drilled into muscle memory the way other so-called bold print or memory items
are for pilots.
The accident report does conclude, as J said, that if they had stowed the speed brakes
within one second of going toga, which would have been three seconds after they heard the
GPWS go off, they would have cleared the mountain.
Interestingly, that three seconds is enough for two rounds of whoop-whoop pull-up.

(01:09:52):
This is the kind of detailed reporting you only get here on CPIT.
If they had retracted the speed brakes when they pulled up and executed a perfect climbout
on the edge of a stall, they would have cleared the mountain by about 300 feet.
Yeah, and I also want to add a little thing.
I don't think we actually mentioned this in the notes, but a lot of airlines have a
policy where if you have the speed brakes out, you have to keep one hand on the speed

(01:10:12):
brake lever at all times so you do not forget to retract them again.
And American Airlines did not have that policy at that time.
So but who knows, he probably would have taken his hand off of it anyway and forgotten anyway,
but we'll never know.
Anyway, at this point, the pilots, they gave it everything.
The cockpit voice recorder captured Tafuri shouting, pull up baby, and more, more, over

(01:10:33):
and over, but it just wasn't enough.
They were climbing steeply when they hit the mountain ridge about 200 feet below the top,
at which point the airplane broke into several pieces, continued forward, and was catapulted
over the top of the mountain and down the other side, where it came to rest quite far
down the slope amid the trees, totally destroyed.
Next slide.

(01:10:54):
Let's talk about the aftermath.
Next slide.
So pretty much everyone unfortunately died instantly on impact.
The impact initially left only five survivors, one of which whom would later go on to die
in hospital.
So that left a total of 159 people dead and four survivors, which was the worst air disaster
in the history of Columbia and still is.

(01:11:15):
A dog survived.
You cannot forget the dog lived.
Yes, it was found still in its cage in the cargo hold, completely unscathed.
Still a lucky, precious creature.
Yeah.
But this was an extremely energetic crash.
The report noted that both the flight data recorder and this cockpit voice recorder
had been pretty well obliterated, but their data tapes stayed intact.

(01:11:37):
The crash was also in a very hard to get to place, requiring either a long trek up a mountain
or by helicopter.
So the crash site wasn't actually found until the next day, and the survivors were there
overnight.
And it's frankly a miracle they survived that long because they all had serious injuries
on a 9,000 foot mountain top in December.

(01:11:59):
And even in the tropics, that's pretty brutal.
I mean, I'll tell you, altitude is altitude.
9,000 feet above sea level is going to be cold no matter where you are on the planet.
Now, not long after the impact, the locals that knew the area were able to get to the
plane and they stole so much shit.
I don't mean they ripped the copper wiring out.
I mean, they stole avionics.

(01:12:19):
They stole mechanical components.
They stole the thrust reversers from the engines.
They did this while the bodies were still in the plane and they were not looking for
souvenirs.
These guys knew what they were doing.
They knew where to get the parts.
They knew how to fence them.
And you will get a single guess as to where all the parts ended up.
It ended up in the most backwater, lawless, hateful place on the planet, the true end

(01:12:43):
state for all aviation and a place where you shouldn't trust a single thing with wings.
Next slide.
The most wretched hive of scum and villainy.
I'm talking of course about Florida.
It's Florida.
Of course it's fucking Florida.
It was never going to be anywhere but fucking Florida.
So they sold the pieces through fences.
A lot of the pieces entered the black market.

(01:13:04):
American Airlines ended up publishing a 14 page small print document listing the serial
numbers of every stolen part.
Despite reaching out to the NTSB, Boeing, and American Airlines, I was unable to find
a copy of this list.
So as with the plans for the SPAY water injection system, if you know a way to get a hold of
them, please get in touch.

(01:13:24):
Next slide.
We had a good bit of discussion in a host chat about what the cockpit would have been
like had they cleared the mountain.
Like do they just kind of tuck their tail behind their legs and go to Bogota because
it's a much better equipped airport?
Do they insist on going to Cali?
And if so, do you circle and go to Tuluwa and start over?
Or do you just like get real annoyed and try a Leroy Jenkins it from Rozo?

(01:13:47):
Colligated explaining their plan to air traffic control who's just exasperated beyond belief.
Hey, Ari, do you know what they could have done though?
Next slide.
Just go around.
Just go around.
Just go around.
Just fucking go around.
Just go around.
Just go around.
Like literally, you can always go around.
We need this on a t-shirt.

(01:14:08):
I'm not kidding.
Just go around.
While this accident is certainly understandable, it is not excusable.
These guys were both extremely experienced pilots.
One had extensive experience on this specific route.
Both of them were very highly rated pilots.
They should not have done this.

(01:14:29):
They should have just gone around.
There are a number of ways we can try to understand how these guys got quote, so screwed up to
quote our captain.
So the first thing I want to focus on is situational awareness.
Obviously these guys did not have it, but what was going on here was a sort of attentional
tunneling.
You have one or both crew members becoming fixated on dealing with a perceived obstacle,

(01:14:52):
which in this case was the FMS not showing the waypoints where they think they are or
want them to be, which they consider to be standing in the way of accomplishing their
goal.
And so the idea that maybe the goal isn't achievable anymore or is never achievable doesn't
cross their minds, and it won't until or unless the perceived obstacle is removed.
So the way I see it is that Captain Tafuri probably thought he could get everything in

(01:15:16):
order if only he was able to get the FMS to do what he wanted it to do, so he just kept
focusing on that at the expense of the actual fast-paced situation that was developing around
him.
Have you ever heard of target fixation?
Yeah, it's similar.
Right.
Not quite the same thing.
So pilots are taught situational awareness strategies, and one of these is that the more

(01:15:38):
complicated a system is, the earlier you should forget about it when a situation starts getting
tricky.
One way I've seen this described is that you don't try to reprogram your nav track to avoid
an impending midair collision.
Obviously, you just grab the yoke and you make an invasive maneuver, right?
So if you're descending toward mountainous terrain and you're fiddling around with the

(01:15:59):
FMS, that should raise all kinds of alarm bells, you don't mess with the FMS in that
situation.
If something seems really off and you have no visual reference and there are mountains
nearby, you climb to the minimum sector altitude and circle until you've got it figured out.
And the same goes for First Officer Williams.
The crash wasn't directly his fault, but the CVR transcript strongly suggests that

(01:16:21):
he had a better understanding of their situation than Tafuri did.
But because Tafuri knew the approach, and Williams didn't, and because Tafuri appeared
to know what he was doing until quite late in the sequence of events, Williams just never
said anything.
And we can't know what he was thinking, but my personal impression is that he sensed something
was wrong but he couldn't articulate what it was exactly, so he just deferred to his

(01:16:45):
captain.
But even a vague statement like, I don't like this, might have altered the course of
events.
I mean, Williams did say, where are we, but that phrasing allowed Tafuri to deflect by
highlighting his futile and misdirected efforts to determine where they were, or rather where
the waypoints were.
It seems like he never thought much about where he was.

(01:17:06):
It was the waypoints that were all screwed up.
It sort of feels like the prototypical boomer dad, right, who insists that, you know, no
problem, it's never his fault.
The airline was delayed, the car broke down.
You know, it's never the, they are never the problem, it is always the system they are
dealing with.
Yeah.
And I don't want to say that Tafuri was in that mindset, we don't know, we weren't in

(01:17:30):
his head.
But it can genuinely be really hard to admit for anybody that I have screwed up, I need
to erase everything and start over, especially when you're the one in command and you're
supposed to know what you're doing.
It's just psychologically difficult.
Yeah, I think the hardest person to admit you screwed up to is sometimes yourself.

(01:17:52):
We on this podcast say that there's no party like a toga party because a toga party goes
missed.
Yep.
You know what else could have prevented this?
You know it, you love it.
It is the official Jack Parsons label CPIT JATO-POP-VP.
Oh yeah.
Yeah, it may be hard to admit that you've made a big mistake and need to go around,

(01:18:13):
but maybe it would be easier if it gave you an excuse to use your airliner's JATO system.
Do we even need to bring up the conspiracy theory that these guys got somehow gas poisoned
and that's why they fucked up?
No, I mean it's a baseless conspiracy with no evidence that basically relies on the NTSB
covering shit up, which the NTSB doesn't do that, I'm sorry, they just don't.

(01:18:33):
The entire premise of this theory is that there's no way two pilots this experienced
would have gotten this lost, but it can happen.
It has happened many times that such experienced pilots have made such basic mistakes.
And there are just quirks of human nature that make us susceptible to biases that make
things like this happen.
You don't need poison gas in the airplane to explain that.

(01:18:57):
So even though that book came out and then it got a lot of attention and now one of the
four survivors of the crash believes that's what happened and is going around spreading
it and I'm just like, I don't know, it's just tragic that...
We have also covered a case of a pilot who had actual poison in his cockpit and still
flew like a fucking boss.

(01:19:17):
That is true, that is literally true.
Next slide.
I want to dig into why they didn't retract the steep breaks.
Well short version is that they just didn't think to do so.
And it's largely because they were not taught to by AA as part of their, as we talked about,
memory item training, which are procedures that are drilled over and over and over until
they become rote muscle memory.
Because if you have like two seconds to act, you're not going to do anything in two seconds

(01:19:41):
that's not a memory item.
It just, you just aren't.
Yeah and I kind of want to blame Boeing for not tying those two systems together.
I mean obviously the Toga button ought to do that.
Again, you know, on Airbus aircraft it does, if you push to Toga thrust the speed brakes
retract automatically, but on Boeing's it doesn't for, I don't know, Boeing reasons

(01:20:02):
TM.
Some pilot going max Toga, he might also want the speed brakes too.
Some asshole engineer in Seattle, probably.
The pilot is always right Ariadne.
Some dipshit in Toulouse on the other hand is like, the pilot he may be on Aselin, so
we are not getting caught this way again.

(01:20:23):
Why not?
Yeah, he programmed the response and then he went on strike.
No, he didn't go on strike, he just went on vacation for a month to Greece.
Okay, so here's the even sillier part, get this, the 757 actually does retract the spoilers
when go around thrust is selected, but only when the plane is on the ground.
So it's, come on, it's right fucking there, you've already built the system to do this,

(01:20:47):
but you just decided not to do it in flight?
Bastards.
So we actually looked this up because this seems like blatantly obvious as a safety feature,
right?
The NTSB thought so too.
There is no situation in which you want to have toga thrust and the speed brakes out
that I can think of.

(01:21:08):
And if for some reason you have managed to magic yourself into a situation like that,
maybe there should be some override or something, but I can't imagine that would ever get used.
The NTSB thought the same as we did because the first two recommendations from this report
were, quote, evaluate the effects of automatically stowing the speed brakes on existing airplanes

(01:21:30):
when high power is commanded and determine the desirability of incorporating automatic
speed brake retraction on these airplanes for wind shear and terrain escape maneuvers
or other situations demanding maximum thrust and climb capability.
Class 2 priority action A9690 and require that newly certified transport category aircraft

(01:21:56):
include automatic speed brake retraction during wind shear and ground proximity warning system
escape maneuvers or other situations demanding maximum thrust and climb capability.
So a class 2 priority action A9691 and the fucking FAA decided not to and the NTSB, oh

(01:22:21):
boy, they were extremely salty about this.
The FAA claimed that microprocessor controlled planes like Airbus Industries and Boeing Company
already had this feature.
They were half right.
Airbus did.
But the fact that an entire 757 full of people just got spread across a mountaintop would

(01:22:44):
seem to imply that Boeing, in fact, very much did fucking not.
An Airbus can't have this happen because of the undeniable fact that it is the best
all-round airliner ever made and the Airbus fly-by-wire system is inherently superior.
Suck it, Boeing.
We can thank Margaret Hamilton for that.

(01:23:04):
And fly-by-wire systems never go wrong, especially not in particularly complex and complicated
ways that require a very soon-to-come episode to entirely explore.
Thank you for that.
Another significant problem that led to this accident was the English proficiency of air
traffic control.
That was a pretty serious problem in Latin America at the time.

(01:23:25):
And according to a pilot I've spoken with, this has not improved much since.
So that's nice.
Next slide.
Can we talk about Thai Airways Flight 311?
Yeah, I think we should.
So this flight is a crash that happened a few years before Flight 965 in July of 1992.
Basically, an A310 was on approach to Kathmandu into runway 20, which like Flight 965 would

(01:23:49):
require a downwind approach, the turnaround.
And the controller instead cleared them straight into an approach.
And like Flight 965, there are pretty serious language barriers between the plane and air
traffic control that led to difficulty understanding where each other was.
This was also a non-radar environment.
The flight crew got behind the airplane while dealing with a minor mechanical fault.

(01:24:09):
They got lost in their mental picture of the area.
The captain tried to figure out where they were using his FMS.
And the ground proximity warning system went off.
And in this case, the captain also pissed away valuable time for evasive maneuvering
by trying to keep going and ignoring the GPWS, believing it to be a false alarm because his
mental map had him somewhere else entirely.
But it probably wouldn't have made any difference because a few seconds later, they slammed

(01:24:33):
into the Himalayas at 400 kilometers per hour and killed 113 people.
So why do we bring this up?
Because much like American 965, a flight deck full of qualified pilots got behind the airplane
and with no technical issues, got lost enough to get killed.
Yeah, or they had one technical issue, but it went away.
So also, the pilots of Flight 965 almost definitely participated in a training module focusing

(01:24:57):
on Thai Airways Flight 311 just a few months before the crash.
And the point of that module was to highlight how focusing on the FMS in a fluid situation
narrowed the crew's attention while they could have clarified that they were going
the wrong way had they simply taken the time to look at an analog compass.
But the point clearly did not get drilled into their heads hard enough.

(01:25:18):
And that's not to say that they didn't understand the point.
If you had asked them, I think they probably would have been able to articulate it to you
quite quite well.
It's that when they found themselves in a similar situation, they didn't have the self
awareness needed to recognize that.
And that's something everyone, I think, should strive to develop.
Next slide, please.
Obviously, there were quite a few changes in the aftermath of the accident.

(01:25:39):
Some of them were procedural, like raising the stereo cockpit threshold in Latin America
and aerospace to 25,000 feet.
Now, I didn't include any of this dialogue in the episode because it's not really relevant
to it.
But there's a huge amount of cross talk about things that are not the mountains when their
conversations should have been the mountains.

(01:26:00):
Mostly they were sort of bitching about crew rest regulations in particular, airline management
in general.
This is on brand.
I'm assured by pilot friends that complaining about the airline is the number one topic
of conversation on the flight deck.
These stereo cockpit regulations matter and we will discuss why in a later episode.
There was also an overhaul in CRM training after this crash as well.

(01:26:24):
Discussing this crash with an airline captain, they described it as a bit of a wake up for
the industry in the same way Colgan would be to an even more major degree later in the
2000s.
The technical overhaul was significant and it has all but eliminated CFET as a cause
of fatal accidents.
In particular, this crash cemented the adoption of eGPWS and I'm going to talk about it because
I want to talk about the Honeywell engineer Don Bateman who built it.

(01:26:49):
So this guy helped develop the original GPWS and then sometime after the fall of the USSR,
he figured out that the Soviets had this comprehensive global terrain database that they developed
to help guide cruise missiles.
And Bateman realised that when combined with the recently made public GPS, these could

(01:27:09):
be extremely useful in saving lives.
So he developed eGPWS that would compare a plane's heading, speed and location and calculate
potential impacts more than a minute in advance.
This is different from the GPWS that Tafuri and Williams had in a couple of significant
ways.
That was based around a radar altimeter that could only look at the ground below the plane

(01:27:34):
and note whether it was rising faster than the plane was ascending.
Given a hillside steep enough, like this one, you might not get enough warning time.
It also included this nice moving map that you can see above on the flight management
displays.
We're all familiar with this now as this sort of thing became the basis for all kinds of

(01:27:55):
GPS based navigations.
The original prototypes used this old Soviet mapping data which they could get hold of
relatively easily after the fall of the Soviet Union.
I wonder how this went down.
They bought this off of God only knows who during the collapse of the USSR.

(01:28:16):
Who was that person who sold the terrain database for cruise missiles?
I'm imagining just like old Soviet hard drives being sold out of the back of a Lada and he
just hands him like a sack full of Deutschmarks.
Well I mean of course the thing is that the US Department of Defense or more specifically

(01:28:36):
the National Reconnaissance Office obviously had the exact same database but they weren't
going to share it with, well at least not that part of Honeywell.
The original prototypes used this old Soviet mapping data but nowadays of course the system
has long since switched to different mapping data.
Basically the same mapping databases that we used to generate the relief maps that we

(01:29:01):
had earlier in this episode in the slideshow.
The tech expanded, gained a lot of popularity and by 2002 anything commercial with more
than six seats was required to have one.
So ever since then controlled flight into terrain crashes really only happen when the
pilots ignore the blaring on proximity warnings for like 45 straight seconds and the people

(01:29:24):
I'm talking about know who they are or rather they would but they don't because they're
dead.
I'm sorry.
If you ignore the, if you have a controlled flight into terrain after with a properly
working EGPWS that is, that sounded a warning within the proper envelope and you still crash

(01:29:46):
you are an idiot.
Obviously there are situations where it may not sound properly but for the most part this
has greatly reduced the amount of C-Fit accidents involving actually competent crews.
I have to ask actually if you get a GPWS warning of either kind and you hit the toga button

(01:30:07):
and then one of your engines explodes and you crash as a result does that still count
as C-Fit?
No, it would probably be something else.
I mean you're still supposed to be able to, yeah I mean that would be incredibly unlucky
if that happened.
It would be very unlucky and I think it's an entry that needs to be on your C-Fit alignment

(01:30:29):
chart.
Maybe I should bring up that I forgot the exact flight number but back in the 90s there
was a United Boeing 747 that almost C-Fitted into Mount San Bruno after takeoff from SFO
because an engine failed and the pilots did not properly monitor their client performance
as they turned to the right to make a missed approach and they came probably within at

(01:30:52):
least within a couple hundred feet of the houses on the shoulder of Mount San Bruno.
United Airlines flight 863 in June of 1998.
Oh it was with the real good looking gray paint job.
Oh god those 747s looked good.
Yeah so that's kind of similar to what you're thinking of J except it didn't actually
crash but that was before the enhanced ground proximity warning system anyway so they didn't

(01:31:16):
they did not get the kinds of advanced warning that we would get now.
Okay next slide.
Finally we can close this section by noting that this is a wrap that Americans still runs
using a 737 Max 8 which is great because this is a very high performance airliner with plenty
of excess power to climb if it needs to get out of trouble.
And famously the 737 Max 8 has never had issues with computer fuckery leading to it flying

(01:31:39):
into terrain.
Yep definitely not.
Next slide.
So what the fuck did we learn?
Since we kind of covered what changes were made in the industry in the aftermath of the
accident I want to talk about what the three of us took away from this.
Fundamentally this was a failure of humor and judgment and it was a failure in a method
and environment that did not permit any sort of mistakes and all of their state of the
art tech was not able to intervene in a way to prevent this.

(01:32:03):
Two group guys relied on their own training and instincts but then they developed confirmation
bias and didn't challenge their own or each other's assumptions.
This obviously is a reason that career resource management is taught over and over and over
and over again to modern pilots.
There's not two of them so one can keep the other entertained.
There's two of them so they could constantly be troubleshooting and providing alternate

(01:32:23):
perspectives.
There is an argument to be made.
I think that the relative experience and training levels of our pilots actually worsened this
confirmation bias.
For Tafuri especially who had done this exact approach 13 times his comfort level prevented
him from accurately taking stock of the situation.
These guys were on the forefront of new technology but they did not understand it on the necessary

(01:32:47):
level.
They had been taught to follow the line on the navigation screen and to trust that their
systems would get where they needed to go and that their safety systems would keep them
out of trouble and neither of those things happened.
I wanted to put this discussion here instead of in the previous section because there's
a lot of elements from this that could be applied to every aspect of life.
A lot of motorcyclists are hit by people who think they know the rap home and may be looking

(01:33:08):
out the windshield but aren't processing what they're seeing.
Recently we've seen a metric ton of people in Teslas and similar cars get in accidents
because they trusted their onboard safety system to prevent accidents and weren't
ready to respond when it was necessary.
And an even funnier version, the fact that you can read story after story after story
about tourists driving into lakes and rivers because GPS told them to and they never took

(01:33:31):
a beat to look out the damn window.
And on a high enough level that's what this is about.
These guys could have pretty much until the last 30 seconds announced that they were getting
a bad feeling about this, gone missed, and climbed out of the valley to a safe altitude
to regroup and try again.
But instead they pushed a bad position and it cost them and 157 other people their lives.

(01:33:51):
They didn't aviate, navigate, communicate.
People talk a lot about how you aviate and navigate before you communicate but aviating
before you navigate is important too.
Love to say.
Very well said and also this is fairly typical behaviour for two guys in a car who are lost.

(01:34:11):
They're not going to stop and ask for directions.
Yeah, that's kind of real.
Yeah.
Alright, next slide.
Alright, thank you everyone for joining us.
Our next episode will be on Malaysia Air 370.
Yep, see you next time.
Bye bye.
See ya.
Bye bye.