Hacker News Reader

Comment by ds9 | Hacker News Reader

"since airline travel is far safer than car travel" -- source quoted by Schneier

This old line has been exaggerated because people make a fallacious comparison. Ask whether people feel safer driving or flying, and many will say driving - this may still be statistically wrong, but it is intuitively based on the correct comparison.

The comparison people intuitively make is "events I can't control in each alternative", while the fallacious-debunkers compare total casualties. But almost all of the plane casualties are beyond the traveler's control, while on the other hand he/she can make sure the car is in good condition, drive while alert, obey speed limits and so on. The plane casualty number may still be lower than the corrected car casualty number, but not nearly by as much as is sometimes claimed.

View on HN

The issue with this comparison is... How many times has "feeling safe" resulted in saving a life?

Statistics isn't a feeling, and the comparison is correct in that, on average, air travel is significantly safer than car travel. One should note, however, that the "on average" is the key bit, as one might argue that drivers in accidents have some special characteristic at work making them more susceptible than most, but claiming that Schneier or his source is making a false comparison somehow is simply unfair and untrue. Moreover, plenty of fatal auto accidents are not the explicit fault of those that die (drunk driver, someone else texting while driving, semi-trucker not paying attn). So "feeling safe" is worth nothing when you're staring down the grim reaper in the grill of the semi that switched lanes without signalling to pass without seeing you.

You can't control other drunk drivers, either.

It does always take two people to get in a [two vehicle] collision. You can't control the other driver, but you can control how you react - and usually whether you've seen them coming in time to do so.

Sometimes it's unavoidable - you're stopped at a light and someone crosses the median to plow into you. In that situation, you are screwed.

But often, there is something you as the not "at fault" party could have done to avoid the collision (even with the drunk driver), had you been paying attention enough to see the situation. Change lanes. Slow down. Speed up. Leave more room. Leave less room. Always, no matter what, be looking.

Once I prevented myself from getting rear-ended at a traffic light: I was stopped and saw someone coming up behind me way too fast. I leaned on my horn, they clued in, and braked in time to stop just touching my fender. (I was in a motorcyle at a time, so REALLY happy with the outcome.)

Again, sometimes it really is unavoidable - but in a surprising number of situations there are actions a driver can take to avoid or reduce the severity of the incident.

Insurance companies understand this - that's why your rates raise when you get in an accident, regardless of whether you're legally "at fault".

Driving well and alertly doesn't end at the bounds of legal fault. Sure if you were rear-ended someone else will pay damages - but if you're in a wheelchair or dead, that money isn't going to make things better.

So to GP's point: there are a LOT of things that are far more under your control when you're driving. Falling out of the sky... not so much.

Indeed- I teach car control clinics, and it's alarming how many people simply freeze when put in a panic situation behind the wheel.

At some point the high rate is just called "normal". I would expect that majority of people freeze. And unless we're all trained in specific situations and see the situation perfectly and have a good reflex and ... there will be no reaction. And how would we prove that any kind of reaction is better than none?

You don't have to be trained to see the situation perfecty or even trained in a specific situation. It is possible to train you brain to still function in a panic situation and try something. You can think about whether that was the right action at a later time.

I've had people tell me "I don't even remember the last 60 seconds" after inducing a spin in a vehicle. It's a scary thought.

I don't doubt that you can influence your own safety on the road to a degree. But the question is, how much? How safe is an extremely safe driver compared to flying on an airliner?

I doubt that even the best driver can get safer than an airliner. It's going to be hard to measure. For one, it's almost impossible to measure the risk of a "safe" driver compared to an unsafe one. For another, it's almost impossible to measure the risk of American airline travel today, because fatalities are so rare that it's no longer possible to turn them into a meaningful number. That last part suggests to me that even the safest driver will be far more at risk on the road than in a plane.

Correction: You can't control other drivers.

I've been rear-ended multiple times. None of the drivers that hit me were drunk or intoxicated. I still had no control over their actions. Short of not driving at all there was no action I could have taken that would have prevented me from being hit. I was simply in the wrong place at the wrong time (e.g., I had been fully stopped for about 60 seconds at a stoplight, not in rush hour traffic, when the light turned green and I was almost instantly hit from the rear. I was about 16 cars back from green light, and there was no way I could have been rolling. But, the driver behind me saw the light turn green and decided that traffic would be moving before he reached my location.)

The fallacy is that people think they are in control in traffic, and that if they are careful then they'll be safe.

Dig up statistics about traffic fatalities, and see how many of them were caused by someone else. Here you can see that one third of casualties are not even drivers: http://www.census.gov/compendia/statab/2012/tables/12s1105.p...

Those stats reveal that the majority of fatal crashes only involved a single vehicle. Also, they don't specify which vehicle the occupants were in. You can totally control who you allow to drive you.

I don't believe it is exaggerated. Air travel is about 10X safer per mile. If you are not a high risk driver, that will narrow, but it's still true.

Comparing total casualties is very deceptive, since cars rack up many more passenger miles than airliners. Still, cars kill lots of people and car safety is a high-value goal.

Once we get self-driving cars that are significantly better at avoiding accidents and mitigating severity of accidents, this all will change in very significant ways. It will certainly result in many more people avoiding the TSA.

>It will certainly result in many more people avoiding the TSA.

There is another interesting way of thinking about this. They might be more inclined to use air travel than road because the 'element of control' is lost from them in both modes of travel. So they would then compare the time it takes to reach the destination and since air travel is faster, they might even end up sucking up the TSA pat-downs.

My car is much nicer than any seat in coach. I won't be touching or smelling the people around me. Nobody will screech "safety information" at me and then provide desultory food and drink service. The sandwich from Rein's will be tasty, and I can reek of garlicy pickles without offending anyone. I'll watch a movie from the Millennium trilogy in the extended cut with the profanity and nudity, and I'll talk on the phone as loudly as I please.

Once self driving cars come, airlines will rediscover service and make 80% of their customers pre-checked or they will lose half their business.

Going strictly by time to reach destination, people will consider "time from when I want to leave my house" or "time before I want to arrive that I need to leave my house".

In many cases over land, they'll find that an autonomous car will be faster.

It's an imperfect comparison yes. We're never given the whole picture, only comparing by miles traveled where of course airplanes go a lot further on each trip. I'd like to see a comparison by time spent on each type of vehicle, or by number of trips, where a lot of drivers make multiple trips every single day and their number of flights would be only a fraction of a percent of that. Not to mention the intuition that a plane crash would certainly result in death, whereas in a car crash you may have a much better chance of survival and recovery.

It would be just as "unfair" to compare driving against crop-dust pilots, which seem to more frequently end up in a crash.

"Not to mention the intuition that a plane crash would certainly result in death"

That intuition is somewhat wrong, depending on how you define 'crash'. http://www.ntsb.gov/doclib/safetystudies/SR0101.pdf table 2, shows 53,487 passengers involved in accidents, with 2,280 fatalities. Table 3 limits that to more serious accidents. There, of 2,739 passengers, 1,215 died, over half due to impact wounds.

Yes, a jet nosediving into hard rock will kill everybody on board, but one making a very rough 'landing' near first aid units? Not something to volunteer for, but not certain death, either.

You'd think this would be fairly obvious coming so soon after a high-profile crash in San Francisco which killed only three people out of nearly 300.

This is an interesting point though I don't think it would tip the scales in the favor of driving. However, another factor is that the total number of actual flights is a subset of all flights, the main filter being weather and mechanical problems. Whereas with driving, people drive regardless of the weather and sometimes, mechanical problems. I wonder how the risk changes further if you delayed your driving as much as airlines delay their flights?

Reply on news.ycombinator.com