xkcd #2940: Modes of Transportation
Title Text:
My bold criticism might anger the hot air balloon people, which would be a real concern if any of them lived along a very narrow line directly upwind of me.
alt-text:
A chart that categorizes various modes of transportation based on their practicality and danger level:
Zone of Practicality:
- Trains
- Airliners
- Boats
- Walking
- Cars
- Scooters
- Bicycles
Zone of Specialty and Recreational Vehicles:
- Motorcycles
- Helicopters
- Light aircraft
- Go karts
- Skateboards
- Rollerblades
- Skis
- Unicycles
- Sleds
- Bumper cars
?????:
- Hot air balloons
“Hot air balloons are the optimal mode of transportation, if your optimization algorithm has a sign error.”
The fact that airplane travel is safer than cars is a myth invented to promote airplane travel. Well, it is not fully a myth, but to get to that result they measure per mile, and that greatly favor airplane travel. If you instead measure how likely you are to die on your next trip, then the dangers of airplane travel will significantly exceed car travel and other means of transportation.
What’s the benefit of measuring per trip versus per passenger mile?
Per trip is more in line with how people think about danger. Like, am I going to die on this trip?
I would think real statistics would be more interesting then peoples emotions when talking about what is actually dangerous.
And the question is am I going to die on this trip? And there the real statistics are pretty clear, cars are safer.
I sort of answered this somewhere else but i will reiterate.
Using this metric you are sort of assuming all trips are equal. No matter how short, or long you are assuming the base danger is the same. This means that driving 100 meters is just as dangerous as driving for a whole day. (See what the problem is?)
And if we look at this premise in isolation: "Am i going to die on this trip"? If the trip is 100m, then a plane is probably out of the question either way. And if the trip is to a different country.. then hey, look at that, the sources you cited come into relevance (where pr distance a plane is safer) and you would have to calculate the danger of completing that specific trip in a car VS flying that distance with a plane.
You are generalizing on terms that make no sense, since "total number of trips" in cars include all manner of different scenarios of some times extremely varying degree of danger. So in order to have data that is statistically relevant and in any form comparable you have to choose a different metric.
So to answer the question again "Am i going to die on this trip?" or to extrapolate "should i drive or fly on this trip", if you cant use generic statistics, the answer will be "it depends. You have to calculate danger for the trip specifically".
This is complete horseshit.
Are you aware how many flights take place every day?
Vs
How many fatal accidents pr flight?
The fact is that almost every time a fatal accident happens in a (commercial) plane anywhere in the world, you hear about it. Because if a plane crashes a lot of people die in one dramatic (and rare) event.
Fatal car accidents litteraly happen every minute of every day. Almost none of them go on the news. (Cause reporting them all would be impossible).
Let me also post some sources, since you did not:
https://injuryfacts.nsc.org/all-injuries/preventable-death-overview/odds-of-dying/
https://www.icao.int/safety/iStars/Pages/Accident-Statistics.aspx/ Air traffic: (3187 fatalities over 10 years)
https://www.who.int/news-room/fact-sheets/detail/road-traffic-injuries (1.19 million people every year die on the road)
I think you underestimate the number of trips per car per day. Most people will take more trips by car per month than they will fly for their lifetime. In Sweden , a country of 10 million, we have about 150 people killed per year from car accidents, yet most adults travel by car daily. That is millions of trips per day, and only half a death.
Yes, and how many die every year from plane crashes in sweden?
If we take a relatively big plane (450 passengers) as an example. One has to fall out of the sky every 3. Years to match the car accident number...
3186 deaths over 10 years VS 1.19 million every year.
(This is globally. Sweden and Norway(where i live) will naturally have pretty radically lower numbers then globally when it comes to road safety.)
But look at that air travel number again: 3186. Over 10 years. Globally. Commercial Air travel is fucking safe. Its horrible for the climate. But its safe.
Whatever way you slice those numbers it comes up air travel i safer. Feel free to find actual statistics that contradict me. :)
I think I get what the guy is trying to say. Per journey, air travel might indeed end up being statistically less safe (how many times a year an average person flies vs. how many times they drive their car) but of course the question is whether that particular metric is any useful. Surely if you replaced all airplane trips with car trips, more people would die.
This Wikipedia article contains a table, which if true, confirms it:
https://en.wikipedia.org/wiki/Aviation_safety#Transport_comparisons
If you sort it by Journeys, you'll find that 117 people die in an airplane per billion journeys, while only 40 die per billion car journeys. But the article points out exactly what I said before.
Funny example that illustrates how important the choice of metric is, is the Space Shuttle which is statistically incredibly unsafe per journey (17,000,000 deaths per billion journeys) and even per hours (only skydiving coming first by a small margin) but is safer than bicycles and only twice less safe than cars per distance traveled because of those insane distances it covers in orbit.
Edit: Not that I do not know whether the table counts only commercial flights or all airplane/helicopter journeys. And also the statistics is pretty old (1990-2000) and only covers the UK, so you may still be right and commercial air travel in the last decade might be safer per journey than cars globally. Can't find a better statistics.
Just check the stats https://en.m.wikipedia.org/wiki/Aviation_safety#Transport_comparisons
Very interesting 🤔
And your point about metrics is pretty spot on.
In the end it becomes an exercise in trying to find the metric that best supports your argument.
We have also been jumping around a bit on geographical limitations. And in for instance Scandinavia, the original premise might be closer to real due to better road safety.
I think implying some sort of myth or ruse is missing the mark hard on this subject.
From your own source:
So I guess this is the point you are trying to make?
Well, what I want to know is "Am I going to die today?". The distance traveled is irrelevant to answer that question. The only reason to add that to the equation is to make air travel look safer.
I honestly think you are showing a fundamental lack of understanding of statistics.
"Per trip" is a horribly poor metric. Because there is a fundamental difference between a trip down to the store, or a cross country trip, even with a car. Also it would be extremely dependent on where you are going, where you live etc. etc.
For the discussion to have any meaning you have to abstract it to a metric that makes sense for all people, or else you would have to also figure in where you usually travel, how good a driver you are etc etc etc.
At that point its a completely meaningless semantics exercise because for instance taking a plane to work is not realy valid for me since i live in the same city as i work... Or lets do it the other way around: If i need to go to Spain tomorrow, its safer for me to fly then to drive there. (This is based on your own sources)
But per mile measurement for flying implies that every mile of a flight is equally dangerous, but the truth I'd that it is most dangerous to start or land, which is a per trip occurrence. The take off and landing is equally dangerous whether you travel a long or short distance in between.
It's still a terrible metric to compare the safety of modes of transport and the Wiki article just below the table explains it well:
If people made similar trips with cars as they do with airplanes, cars would lose in the per journey metric big time.
Of course cars would loose if you tried to use it to travel across the Atlantic...
If you are traveling across the Atlantic to get from Los Angeles to New York i would argue that you are traveling the wrong way...
Yes, and?
The point of distance is to take it into aggregate, for both modes of transport.
This is in fact the exact point i am making.
Per trip measurement implies that every trip (regardles of time or distance traveled) has equal danger.
You can argue that "per person miles" is a better metric, but that is completely orthogonal to their initial claim.