For those of you who don't read NTSB reports as a hobby, know that Preliminary reports generally do not assign probable cause or list contributory findings. They're usually just full of basic, factual information like the date, time, location, weather, info about the aircraft and pilot, and so on. So don't go into the PDF thinking you're going to learn the cause of the crash.
While this is generally true and having read this report I do not believe it to contain any new material information related to the probable cause, it continues to paint the picture of the captain making a pretty big series of mistakes.
It really sucks that the FO could not intervene further as he clearly identified the problem and articulated the solution.
I certainly doubt there will be a satisfying ending to this one.
> It really sucks that the FO could not intervene further as he clearly identified the problem and articulated the solution.
First officer could have easily intervened, all they needed to do was call for a go-around. If the captain wouldn't initiate a go-around(AFAIU typical airline policy from my understanding is that if either pilot calls for a go-around then they must initiate a go-around) after being called for by the FO they should have initiated one themselves. The events from the CVR show only the captain ever called for a go-around after it was too late.
Even if they had died, it would still be relevant. It may point to deeper systemic issues at the company or could lead to identify additional ways to train aircrews to avoid this situation in the future.
It's always relevant but the NTSB's recommendations sometimes fall on deaf ears in a way that endangers future safety.
The fundamental question is why they pressed on and what we need to look for in other airlines to avoid this. Why was the FO so unassertive? Was this a common pattern at that airline? Was there pressure from the airline to minimize go-arounds?
Is this an issue with the re-training of rotor to fixed wing pilots?
> Was there pressure from the airline to minimize go-arounds?
The thing I've heard from the fine folks on the pilot forums is that for most airlines (granted, passenger airlines, not necessarily cargo), the reverse is usually true. An unstable approach generates a number of these warnings and the plane tattles on the pilot in real time, there will be repercussions for not going around.
This amuses me, as it clearly addresses a tiny minority of HN commenters, and I can only imagine the arcane, irrelevant documents each one of us has saved to our disk and open in tabs right this very moment
Editing Wikipedia for 17 years has left me with research superpowers, in any arena I should choose, and it unfortunately supercharged my ability to ruin dinner parties by blurting out series of factoids about any item that should be mentioned.
In fact, my research efforts and other actions may have been directly responsible for a security lockdown at an international airport once upon a time. I have no way of knowing for sure, and I was innocent of all malice, but imagine my surprise when I arrived to find the chaos.
Well yeah the report is not a bunch of nothing also. NTSB usually takes 12+ months to complete their investigation so they release a preliminary report within a month of the incident usually to share relevant factual information. But the investigation remains ongoing so they have the time to check all angles and come up with final conclusions.
> According to Automatic Dependent Surveillance-Broadcast (ADS-B) Out and FDR data, flight 7598 was about 105 ft above ground level (agl) at a groundspeed of 177 kts as it crossed the threshold of runway 30. The accident airplane touched down about 3,750 ft past the threshold at a groundspeed of 160 kts.
Runway is 9360 ft long, so touchdown 40% down the runway.
> winds were from 190° at 19 kts with gusts to 22 kts
So, that's strong crosswind from slightly behind - not great.
> 1:42 [1 min 42 sec before end of recording]: The captain called for flaps to be set to 20 degrees, and the FO cautioned they were too fast. The FO made additional comments regarding the airplane’s excessive speed throughout the remainder of the recording. There was not a consistent verbal response to the comments about speed.
> 00:31 Sounds similar to the airplane touching down.
> 00:15 The captain called for a go-around.
> 00:10 Sounds similar to the airplane departing the paved surface.
Looks like they landed way too long and fast, without deploying speed brakes and thrust reverser. Should have gone around. Then there was a go-around attempt that came way too late and was aborted.
> The landing gear weight-on-wheels (WOW) parameter intermittently switched between “air” to “ground” for about 7 seconds. The WOW parameter switched to ground for the remainder of the recorded data when the airplane had about 2,730 ft of runway remaining. The FDR data showed the speed brakes remained stowed and the thrust reversers did not deploy during the landing and subsequent landing rollout.
> Wheel brakes were manually applied at a groundspeed of 146 kts. The brake pedals for wheel braking were subsequently released about 8 seconds later at a groundspeed of 120 kts and the throttle resolver angle for both engines increased to go-around thrust values; the airplane had about 930 ft of runway
remaining.
> About 4 seconds later, the throttle resolver angles for both engines reduced to idle values at a groundspeed of 117 kts, when wheel brakes were reapplied; the airplane had about 415 ft of runway remaining.
> without deploying speed brakes and thrust reverser.
I don't think they were able to, since their aircraft never really touched down firmly enough to satisfy the weight-on-wheels interlock system. The aircraft won't allow the crew to reverse thrust until it settles into a "ground state" which we have today because some past guys managed to crash airplanes that way. Also, these guys commanded go-around power when they should have been hitting the brakes, so there's no way the flight systems were going to open the reverser.
If they had stood on the brakes and had the thrust reversers armed, 2730 feet of runway probably wouldn't have been quite enough to get it all the way stopped, but it may have stopped while still on asphalt, and at the very least would have been going substantially slower. Likely no fatalities, no destroyed plane, just landing gear repairs and pilot ego repair.
Speed brakes can be deployed whenever, even in flight, because they are flight controls, too. Deploying them would have settled the aircraft on the ground, where the brakes would have worked better and the thrust reversers could have been used.
The captain held an airline transport pilot certificate with a type rating for the Boeing 767 that he received on May 18, 2026. The captain was hired by 21 Air in April 2026. He had accrued about 7,267 total hours of flight experience, as reported in his interview with NTSB investigators, of which 150 hours were in the accident airplane make and model.
The FO held an airline transport pilot certificate with a type rating for the Boeing 767 that he received on April 2, 2025. The FO was hired by 21 Air in February 2025. He had accrued about 2,800 total hours of flight experience, as reported in his interview with NTSB investigators, of which 800 hours were in the accident airplane make and model.
This stood out to me --- both pilots are extremely new.
Not much more than me if you'd sum up my Kerbal Space Program and Flight of Nova hours.
Given, much of these is in building rockets and orbital flight, which doesn't translate that well into planes.
Yeah that stood out to me too. It feels like they should be putting veteran pilots with newer ones so that you gain experience while having it in the cockpit already.
To my understanding a modern 767 usually has EGPWS (enhanced ground proximity warning system) which includes more than just flaps/gear/speed, it even knows about particular airports which otherwise trigger false alarms due to their configuration. An airline pilot should very rarely hear it.
For those of you who don't read NTSB reports as a hobby, know that Preliminary reports generally do not assign probable cause or list contributory findings. They're usually just full of basic, factual information like the date, time, location, weather, info about the aircraft and pilot, and so on. So don't go into the PDF thinking you're going to learn the cause of the crash.
While this is generally true and having read this report I do not believe it to contain any new material information related to the probable cause, it continues to paint the picture of the captain making a pretty big series of mistakes.
It really sucks that the FO could not intervene further as he clearly identified the problem and articulated the solution.
I certainly doubt there will be a satisfying ending to this one.
> It really sucks that the FO could not intervene further as he clearly identified the problem and articulated the solution.
First officer could have easily intervened, all they needed to do was call for a go-around. If the captain wouldn't initiate a go-around(AFAIU typical airline policy from my understanding is that if either pilot calls for a go-around then they must initiate a go-around) after being called for by the FO they should have initiated one themselves. The events from the CVR show only the captain ever called for a go-around after it was too late.
The CVR transcript is pretty damning I'm not sure what new details you expect them to uncover.
The alerts from the GPWS below 1000' signify an unstable approach, you go around, the end.
What they will investigate is the why they did not go around, but it's largely irrelevant.
It would be irrelevant if the pilots died, however, since they survived and potentially will continue flying, the why is relevant to not do it again
Even if they had died, it would still be relevant. It may point to deeper systemic issues at the company or could lead to identify additional ways to train aircrews to avoid this situation in the future.
It's always relevant but the NTSB's recommendations sometimes fall on deaf ears in a way that endangers future safety.
and potentially will continue flying
Hopefully not without some very thorough additional training and certification.
The fundamental question is why they pressed on and what we need to look for in other airlines to avoid this. Why was the FO so unassertive? Was this a common pattern at that airline? Was there pressure from the airline to minimize go-arounds?
Is this an issue with the re-training of rotor to fixed wing pilots?
> Was there pressure from the airline to minimize go-arounds?
The thing I've heard from the fine folks on the pilot forums is that for most airlines (granted, passenger airlines, not necessarily cargo), the reverse is usually true. An unstable approach generates a number of these warnings and the plane tattles on the pilot in real time, there will be repercussions for not going around.
> don't read NTSB reports as a hobby
This amuses me, as it clearly addresses a tiny minority of HN commenters, and I can only imagine the arcane, irrelevant documents each one of us has saved to our disk and open in tabs right this very moment
Editing Wikipedia for 17 years has left me with research superpowers, in any arena I should choose, and it unfortunately supercharged my ability to ruin dinner parties by blurting out series of factoids about any item that should be mentioned.
In fact, my research efforts and other actions may have been directly responsible for a security lockdown at an international airport once upon a time. I have no way of knowing for sure, and I was innocent of all malice, but imagine my surprise when I arrived to find the chaos.
I found the timeline of events from the flight data recorder to be fairly suggestive of a likely conclusion.
Though that's as a lay-person; I am not a pilot but would love to hear if the seemingly obvious issue is the likely root cause.
This one is blindingly obvious. Appalling airmanship, just like it reads.
Touching down 2/3rds of mile down the runway at around 25 knots over Vref. That was never going to go well.
Well yeah the report is not a bunch of nothing also. NTSB usually takes 12+ months to complete their investigation so they release a preliminary report within a month of the incident usually to share relevant factual information. But the investigation remains ongoing so they have the time to check all angles and come up with final conclusions.
Yikes.
> According to Automatic Dependent Surveillance-Broadcast (ADS-B) Out and FDR data, flight 7598 was about 105 ft above ground level (agl) at a groundspeed of 177 kts as it crossed the threshold of runway 30. The accident airplane touched down about 3,750 ft past the threshold at a groundspeed of 160 kts.
Runway is 9360 ft long, so touchdown 40% down the runway.
> winds were from 190° at 19 kts with gusts to 22 kts
So, that's strong crosswind from slightly behind - not great.
> 1:42 [1 min 42 sec before end of recording]: The captain called for flaps to be set to 20 degrees, and the FO cautioned they were too fast. The FO made additional comments regarding the airplane’s excessive speed throughout the remainder of the recording. There was not a consistent verbal response to the comments about speed.
> 00:31 Sounds similar to the airplane touching down.
> 00:15 The captain called for a go-around.
> 00:10 Sounds similar to the airplane departing the paved surface.
Looks like they landed way too long and fast, without deploying speed brakes and thrust reverser. Should have gone around. Then there was a go-around attempt that came way too late and was aborted.
> The landing gear weight-on-wheels (WOW) parameter intermittently switched between “air” to “ground” for about 7 seconds. The WOW parameter switched to ground for the remainder of the recorded data when the airplane had about 2,730 ft of runway remaining. The FDR data showed the speed brakes remained stowed and the thrust reversers did not deploy during the landing and subsequent landing rollout.
> Wheel brakes were manually applied at a groundspeed of 146 kts. The brake pedals for wheel braking were subsequently released about 8 seconds later at a groundspeed of 120 kts and the throttle resolver angle for both engines increased to go-around thrust values; the airplane had about 930 ft of runway remaining.
> About 4 seconds later, the throttle resolver angles for both engines reduced to idle values at a groundspeed of 117 kts, when wheel brakes were reapplied; the airplane had about 415 ft of runway remaining.
> without deploying speed brakes and thrust reverser.
I don't think they were able to, since their aircraft never really touched down firmly enough to satisfy the weight-on-wheels interlock system. The aircraft won't allow the crew to reverse thrust until it settles into a "ground state" which we have today because some past guys managed to crash airplanes that way. Also, these guys commanded go-around power when they should have been hitting the brakes, so there's no way the flight systems were going to open the reverser.
Indeed, by the time the WOW settled into "ground", they had 2730 ft of runway left. :-/
If they had stood on the brakes and had the thrust reversers armed, 2730 feet of runway probably wouldn't have been quite enough to get it all the way stopped, but it may have stopped while still on asphalt, and at the very least would have been going substantially slower. Likely no fatalities, no destroyed plane, just landing gear repairs and pilot ego repair.
Speed brakes can be deployed manually I think.
Speed brakes can be deployed whenever, even in flight, because they are flight controls, too. Deploying them would have settled the aircraft on the ground, where the brakes would have worked better and the thrust reversers could have been used.
Yeah its like the anti Sully in terms of decision making skills.
The captain held an airline transport pilot certificate with a type rating for the Boeing 767 that he received on May 18, 2026. The captain was hired by 21 Air in April 2026. He had accrued about 7,267 total hours of flight experience, as reported in his interview with NTSB investigators, of which 150 hours were in the accident airplane make and model.
The FO held an airline transport pilot certificate with a type rating for the Boeing 767 that he received on April 2, 2025. The FO was hired by 21 Air in February 2025. He had accrued about 2,800 total hours of flight experience, as reported in his interview with NTSB investigators, of which 800 hours were in the accident airplane make and model.
This stood out to me --- both pilots are extremely new.
Not much more than me if you'd sum up my Kerbal Space Program and Flight of Nova hours. Given, much of these is in building rockets and orbital flight, which doesn't translate that well into planes.
Yeah that stood out to me too. It feels like they should be putting veteran pilots with newer ones so that you gain experience while having it in the cockpit already.
Related: “NTSB issues investigative update on B-767 runway excursion accident in Miami” (ntsb.gov)
150 points | 22 days ago | 292 comments
https://news.ycombinator.com/item?id=49650418
I would have thought the terrain warning would have finally been the thing that would break his fixation, but I guess not. Just wild.
I'm gonna guess that the crew have been regularly landing while ignoring similar automated warnings in the past
To my understanding a modern 767 usually has EGPWS (enhanced ground proximity warning system) which includes more than just flaps/gear/speed, it even knows about particular airports which otherwise trigger false alarms due to their configuration. An airline pilot should very rarely hear it.