Hacker News Reader

> "Driving is so much more dangerous than flying even if there’s a MAX crash every 6 months (doubtful)."

I'm not sure that's correct. Let say there's 350 MAX aircraft in service, flying 5 flights per day on average. In 6 months, thats about 320,000 flights. So a 1 in 320,000 chance of a fatal crash. Let's say the average route length is 1000 miles. That's 1 crash per 320 million miles flown, or 3.125 crashes per billion miles flown.

If we assume an average of 150 passengers per flight, then it's 468 deaths per billion miles for the MAX, compared to only 0.02 deaths per billion miles across all aircraft. [1]

In the US, there are 12.5 deaths per billion miles driven, including pedestrian and cyclist deaths from motor vehicle collisions.[1]

If it crashed once every 6 months, the MAX would be around 37X more dangerous than driving, and 2340X more dangerous than an average commercial flight in the United States.

[1] https://en.wikipedia.org/wiki/Transportation_safety_in_the_U...

150,000 flights, 172 seats per is 25.8 million passenger seat-journeys. Given the type, I think it's fair to say they're at least 1200mi long (why fly a 7M8 on a sub-2 hour flight, when you can just throw a regional jet at it). The whole point of the MAX family is that it can fly further than a normal variant. This is 30.96 billion passenger seat-miles. There were ~350 deaths. This is 0.000000011 deaths per passenger seat-mile, or 11 per billion. This is in line with but still less than driving (12.5 per your data source).

I imagine 1200 miles is short for an average 7M8 flight though, which could make it substantially safer. With 2500 mile averages you get back to that 3X number I cited. It's still no more a death trap than your average Honda Civic (10% safer by your metrics, more by mine).

To your point, it is of course riskier than a different aircraft.

My whole thesis was to ask what effect this will have across the entire population and how will humans react, irrationally, to this news.

On the flip side, if the plane is intended for long flights, do you think people are just going to drive 1200+ miles and incur the gas, the additional time, a hotel stay or a few?

Some trips may be made using older aircraft. Some trips may not be made at all. But I think assuming a 1:1 correlation with driving is a bit silly.

A very good point, although, my fear wasn't that people would replace 1:1 7M8 flights with a roadtrip across the country, but that grounding an entire plane type would cause people to replace short flights with long drives. I suppose planes falling out of the sky may do that too :)

There are arguments against measuring crash rates by “number per vehicle miles traveled” and instead by “number per vehicle hours traveled.” VMT is a product of speed and time, and thus a fatalities-per-VMT metric will nearly always show the faster routes to be safer.

But compare spending 20 minutes traveling 10 miles on a local street, vs 10 minutes traveling 10 miles on a freeway. Both are considered equally safe using a fatalities-per-VMT model. But in the former, you’ve stayed safe for a longer portion of your life. (See http://pedshed.net/?p=1050)

Using that metric would make that Boeing far less safe compared with driving.

But you travel to get somewhere not to be in a vehicle for X amount of time. So why would any metrics besides per-distance be more relevant.

As a traveler, I care about the trip, not the time nor the distance. If I take this trip, what's my probability of dying if I fly it versus drive it.

Airplane travel have fixed segments (takeoff and landing, ascent, descent) and variable segments (the coasting phase that varies by distance traveled). Car travel is only the variable segment. Hard to compare apples to oranges.

> measuring crash rates by “number per vehicle miles traveled” and instead by “number per vehicle hours traveled.”

there's another issue, which is plane risk is uneven during their journey, so a type that does half the flights but twice the distance on average will end up almost 1.5 times as safe all other things being equal.

this source claim 45% crashes happening on takeoff and landing which are a minimal percentage of travel if you measure risk by mile or hour: https://eu.usatoday.com/story/travel/columnist/cox/2013/08/2...

a way to normalize that would have to use "number per flights" but then you'd get trains, with their multi-stop journeys, skewing the numbers the other way.

That argument seems flawed to the point of absurdity. Two extreme examples showcase this:

1) a teleporter with the exact same risk as driving 20min on local streets or 10min on the highway would be the worst thing ever invented. In reality, 100% of people would choose it unless they were intentionally wanting to spend time driving

2) A 100% safe ridable snail would be the ideal form of transit, given that you'd spend your whole life safely trying to get somewhere.

Fatalities per passenger mile is a good metric for passengers, but for the FAA, I think crashes per vehicle mile is a better target.

That is, the plane isn't inherently 100x safer if you just fly it empty.

Yes, the plane is much safer if you fly it empty.

It's still a risk to those on the ground!

Also, more broadly, you have some degree of control over dying in a car. Like the odds of dying on the road while driving a modern safe car, legally and carefully, would be lower still again. Airplanes we cede all out control over safety to the airline / pilot and manufacturer.

This would be true if you were on the car alone. Unfortunately we all share the road, so it only takes the most dangerous person on the road to ruin your safety.

He said "some degree of control", not "complete control".

Even so, just by following road rules, driving sober and having a well-maintained car, you likely go to the bottom decile in car-related risks. That much is definitely under your own control.

And even worse if you do the calculation based on time spent in vehicle rather than distance.

The alternative isn't driving, it's using a different (safer) plane.

Maybe! Looks like there've been 350 deliveries so far [1]. Are that many planes available on short notice to replace them? Airlines generally operate with 80% load factors, it'd be hard to just re-shuffle everyone.

The grounding and reaction may also shake people's faith in air travel in general, causing them to drive more even if planes were available. Bruce Schneier has a good write-up on this effect post 9/11 too. Humans are terrible at dealing with this kind of thing.

The reality is you're much more at risk in your Uber ride to the airport than you are aboard the 7M8.

[1] https://en.wikipedia.org/wiki/Boeing_737_MAX#Orders_and_deli...

I think in practice more of the flights will be replaced by other flights or not travelling than driving 2500 miles.

Commercial flying is safer on average than driving, but that's because of, not despite the extreme approach to safety concerns. (General aviation is actually substantially more dangerous than driving)

> Are that many planes available on short notice to replace them?

If there are, they'll have to wipe the dust off of them, and send type-rated crews into territories they're not familiar with day-to-day.

That, or defer maintenance of the existing fleet.

I'm unconvinced that this is the safe resolution, but that's what will happen.

At least the US and Canada allowed planes in-flight to land, instead of making them turnaround and come back anyway via non-revenue ferry flights.

Maintenance deferral is not an option (unless they're electively carrying out maintenance checks well ahead of OEM-specified intervals) and type-rated crews landing in unfamiliar territory happens on a daily basis and certainly isn't a risk factor comparable with a [possibly] flawed aircraft. If the delay's long enough to actually see parked aircraft brought into service, it'll largely be the same crews flying older 737s anyway.

Mostly people will just cancel their holidays or agree a teleconference because getting flights on that route at short notice is too difficult.

> type-rated crews landing in unfamiliar territory happens on a daily basis and certainly isn't a risk factor comparable with a [possibly] flawed aircraft.

Flying in/out of unfamiliar territory is a known risk. Whether the 737 MAX series is higher risk or not is an unknown risk because of insufficient sample sizes.

Hopefully some statistician will start tabulating P values so we can see if the confidence intervals overlap or not. Treat this situation like a new "better" drug for X where 2 users died in quick succession. Is there enough evidence to take it off the market or not? AFAIK, nobody has (yet).

>If the delay's long enough to actually see parked aircraft brought into service, it'll largely be the same crews flying older 737s anyway.

Sorta. On a Canadian airline, Westjet, the MAX is flying routes that are too far for the regular 737s without a fuel-stop (by my estimates). I'm booked for such a flight in 2 months. I'll get to learn what happens. I would have been happy to fly the MAX8 based on the current lack of evidence for grounding.

Have you examined the actual outcome of grounding this model of plane?

I would think at least some people are driving instead. US airlines run at a pretty high utilization, grounding dozens of planes isn't going to be easily absorbed into other flights.

Grounding a newly release airliner is not new. It's only news because of the two crashes and 300 victims. I'm pretty sure these airlines are careful and still have some of the older 737/A320 that these MAX were supposed to replace.

Issues with new airliners and engines -specially engines- are not rare at all. So this sort of grounding is not rare: see 787 first grounding (battery issues), current partial grounding (Rolls-Royce Trent 1000 issues), or Pratt & Whitney ongoing issues with the A320NEO engine (PW1000G issues).

You can imagine the sort of clusterfuck when PW needs to produce 60 engines a month for Airbus, only manages to build 20 in its factory... but need to send these 20 engines to replace already delivered engines because they are defective. The situation has improved a lot lately, but 2017/2018 were tough...

Only the A350 introduction has so far been relatively event free. Probably because the Trent-XWB is a derived from the Trent 1000, so the 787 took care of the teething...

> Only the A350 introduction has so far been relatively event free.

I am biased on this since I have spent a couple of years working on the development of A350XWB systems (and only marginally contributed to other programs) but I always had the feeling that, after the trauma of the A380 (rampant production issues, lengthy delays and management mishaps/stock fraud), the focus for A350XWB was put on accountability at all levels.

This might not be sustainable for a large corporation but retrospectively, that seemed to have had the desired effect at least for this particular aircraft program.

Of course, the other thing to consider is how many more people will drive due to a loss of trust in the FAA and airlines over the failure to ground the plane once it's percieved as unsafe.

"I could drive or fly...oh, no, but what if they try to put me on a 737 Max? I better drive."

No, most people don't pay that much attention, honestly. "Drive or fly? Hm... that plane did just fall out of the sky right, planes freak me out, let's just drive." This is an excerpt of a conversation I've had with friends after the MH crashes. I had to explain to them planes generally don't get shot out of the sky or just disappear and that in spite of this all, flying remains safer than driving.

People pay attention to things like this. My kid is flying to Denver in a few weeks and was looking to see what plane type it was. An acquaintance is heading west for work and the first question was ‘Max 8’?

Sure, a avoid reason to ground the affected model.

“The ones that fall out of the sky have been grounded, so that’s great”

I see what you’re saying, but there are intangibles that are hard to factor in. For example, a big reason that air travel is so big an industry is precisely because it is so safe. But that image could get tarnished very easily, at which point you would see a sudden and large drop in paying customers (think 9/11). Consequently, grounding the planes until everyone can be certain of the cause and the remedy is in the best interest of the entire industry (although not great for Boeing in the short term).

Why would grounding the planes cause people to drive more? There's very few of them in operation, and airlines have other planes to fill the void.

Why will grounding the plane cause more people to drive? It does not have large enough share of existing planes to cause significant rise in prices to make people choose driving. Especially considering that it is used for long distance flights, which are more likely be replaced by train instead of car.

It still pushes the balance. Not everyone lives right by an airport and has a true destination right by an airport.

OR the planes doing the short-haul flights will now have to make these longer-haul flights, increasing the short-haul flight costs.

>Driving is so much more dangerous than flying

Flying is only as safe as it is thanks to regulations written in blood. I don't think flying today versus driving today is the right comparison to make, rather flying today versus flying in a decade. If we allow leniency on the basis of alternatives to flying being more dangerous now, improvement of flight safety may be squandered so that the safety of future transport as a whole is reduced.

I also think these regulatory bodies have to make grand gestures, at times, to reassure the bags of blood and bones that get on planes in a way that statistics do not.

> Let's assume that each 7M8 journey averages 2500mi

Let's then also assume that people are not commontly drive transcontinental distances. Your calacluations are interesting but all over the place. E.g.:

> How many more people will take connecting flights — the > danger is in the take off and landing after all

So in this case distance does not matter. How does car safety compare to airplanes in terms of number of trips?

Anyway, the biggest problem is assuming that all grounded flights will be replaced by driving. Either the flight will use different aircraft or the trip is cancelled in majority of the cases.

The holistic view is rather that the only reason we have such safe aircrafts so you can get such nice numbers is because there's very well crafted procedures for handling crashes.

Reacting strongly when those procedures seems to have missed something is important not because of the crash itself but because the process is important. Could the same oversights lead to additional fatalities is the question, the robustness of the boeing is a minor point.

This is why I come to Hacker News. Thank you