Schneier on Security
A blog covering security and security technology.
« Buying an ATM Skimmer |
| How Much Counterterrorism Can We Afford? »
June 22, 2010
The Real Risk: Traffic Deaths
The New York Times Room for Debate blog did the topic: "Do We Tolerate Too Many Traffic Deaths?"
Posted on June 22, 2010 at 11:50 AM
• 65 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
The common theme in the posts on there is "we accept this risk". But what does "not accepting this risk" look like? Does it mean we forfeit driving? Does it mean we impose maximum speeds on all cars, so that they don't go over the limit? I honestly don't think we do accept the risk. We do everything in our technological power (see, crash test ratings on cars) to ensure driving is safer. We create laws stating certain visibility of road signage and markings, as well as things like daytime running lights to improve visibility of cars. I personally think we are doing a damn fine job of "not accepting" the risk.
I think we "accept" any residual risk (whatever that means) because its a controllable risk. We control the speed, and way we drive. If we personally cannot control it, then we want others (or ourselves) to take whatever steps we can to mitigate the risks. But, in driving, we already do take those steps ourselves. Think about when you are a passenger with a reckless driver. Often you will ask them to slow down, or you may not ride with them again.
Besides, we need cars. One cannot possibly expect that everything and everyone you need or want to see on a daily basis is within walking distance, or that transit is acceptable when mobility is required.
It has also become a familiar risk. Most of us know someone, or have been in, a car accident. It is not like being, for example, shot with gun. We have changed our society so guns are much more regulated, and therefore, in principle, pose less of a risk.
I see it like the heavy metal scare on the novelty glasses at McDonalds. Arguably the food served there poses a higher risk than the glasses, but the food is familiar and necessary so we do not see it as a risk. On the other hand, the glasses are a novelty product, something that has not been marketed as the core McDonalds family experience, so is something that can be single out as the token effort to enforce safety.
I think we need to stop using the word 'accident' unless it was really something not reasonably avoidable. Someone sending a text and plowing into someone at a stoplight isn't an accident, and to call it that gives the connotation that "oh well, it was just an accident, nothing we could do about it". Tom Vanderbilt blogs about this under the heading "Accidental Journalism"
Do we tolerate too many deaths due to cancer?
Do we tolerate too many deaths due to vascular disease?
Do we tolerate too many homicides?
Can we do something that will make things significantly better? I don't know. Making motorcycles illegal will help, but it will take away the perceived freedom from some. How do you keep people under the influence of medication, alcohol, or insufficient sleep from driving?
"We do everything in our technological power (see, crash test ratings on cars) to ensure driving is safer."
Welcome to America. When I was 16, I went behind the MVA with a pickup truck. I put my seatbelt on, made the examiner do the same. I put the handbrake down. I drove to a stop sign in a parking lot and-- get this-- stopped! Then signaled, turned... FAILED to parallel park... eventually they gave up and just let me have that, even though you can't have a license if you fail that.
Anyway, now that I've failed my driving test, they made me drive 5 feet to ANOTHER stop sign, stop, signal, turn. 30 feet to another stop sign. 50 feet. Pull to the end of the lot, stop, put the handbrake on, turn the car off.
And then I got a license.
You know what we don't teach or test on in America? Driving.
Preparing to take a DMV test in Atlanta (Marrieta actually) on a dirt 'play lot' I talked to someone else who had failed their previous test because they drove down the wrong side of the imaginary road.
@Rich Wilson: "Marrieta actually"
Ohhh, my condolences. We stopped there on our way to Florida because they have an Archiver's, and it was some of the worst traffic (and worst drivers) I've encountered. Still can't get over the people who get in the far left turn lane because it is the shortest line then about cause a pile up because they want to get to the far right lane to make a turn a block later.
But lunch at Twisted Taco was good. lol
I have family in Europe and the driving tests they describe seem more rigorous than the ones here in the US. Not sure if that is so, and not sure what their accident rates are either.
An interesting article, though a tad one-sided, and seems to be offering up a lot of statistics that seem suspect to me.
I dislike statements like "speeding is a factor in 1/3 of traffic fatalities." What exactly is a "factor"? Is a "factor" the same as a "cause"? Could it be that speeding is just a "factor" in 1/3 of all driving?
I get the impression that as soon as speeding or alcohol is identified a "factor", then people don't bother looking into it any deeper.
And here's another form of statement that really irritates me: "One death from a traffic crash is unacceptable."
Well, no, that would be very acceptable.
If lowering the speed limit from 100km/h to 90km/h would save one life... well, I'd say it sucks to be that guy. It would have to save a lot more than one life. I don't know how much more, but definitely more.
According to Crap Cycling and Walking in Waltham Forest, in 2009, 140 of the 184 fatalies in London from traffic RTCs were people walking or cycling [ http://crapwalthamforest.blogspot.com/2010/06/... ]
That means 76% of the people who ended up dead in London last year didn't say "I'm taking a risk but driving is necessary" or "we need cars". They said "I will walk or cycle in London."
Someone else made the decision to drive, perhaps the decision to speed, to drink, to pick up the phone. Someone else died, someone else's family got the phone call. This is not controlling risk, it is that if you walk or cycle our city the danger comes from other people who are no good at risk assessment.
The people doing the driving are not the people dying. We have spent money on car safety -Volvo and Mercedes make so much fuss of it- airbags, ABS, etc., but that can lead to complacency, and it avoids the issue that the people dying are people outside the vehicle. All you can do there is reduce kinetic energy, either by reducing vehicle mass or or vehicle velocity, ideally both.
So yes, we do undertolerate vehicle related deaths. We use the phrase "accident". If your computer got 0wned because you weren't keeping flash up to date or running a decent AV toolk, is that an Accident, or an event which can be predicted based on your inadequate processes? Same for Road Traffic Collisions. If someone drives to a bar, drinks and drives home, that's no more an accident than browsing to porn sites with a non-updated copy of WinXP service pack 1 . If someone drives excessively fast and then "looses control" at a corner, then that's no more an accident than running IE6 with the "accept ActiveX downloads" feature enabled.
What is interesting about RTCs is not how so many people say "Driving is a necessity", but how we don't say "Computing is a necessity" when we worry about computer security events. That we still believe that we can fix, that we can manage.
I would go on to argue that speed is a 'factor' in 100% of accidents, as a collision is not possible if no parties are moving. However, just speed is rarely ever the full cause of an accident. If speeding is the only traffic law you are breaking, you'll never hit someone, it's not as though following the speed limit prevents you from hitting others when not paying attention.
I've always supported a three strikes and you're out policy. A system of laser sats monitors all traffic and when you've done three stupid things in one day you get zapped out of existence. Some cities would have sparse roads within a week, while in others traffic would flow much faster and safer.
@Piper: I believe "factor" is defined to be when one of the charges in the accident is when the person admits to being over the posted limit before the accident. Which is why that 40% makes sense. According to one of the articles I came across, speed limits get set such that the 50th percentile tends to be around the speed limit (meaning 50% of people are speeding at any point in time).
It makes sense that speeding is a factor in less than 50% of the events, because sometimes a smart driver can lie about their speed, and find no evidence to the contrary.
It probably doesn't help that "Failure to control speed to prevent an accident" is the actual charge in AZ when you rear end someone, regardless of speed. My one accident on record was as a teenager, rear ending someone. I received a 10mph speeding ticket because the officer estimated I was probably going 10mph at the time of the collision, and clearly that was 10mph over what was safe. I thought it was the most simple and logical method of defining "safe speed" I'd ever considered!
If people wore their seat belts, didn't speed, and didn't drive impaired most crashes could be avoided (at least the sort that kill the occupants of four wheeled motor vehicles). These three factors can be dealt with technologically or administratively as soon as we're serious about reducing the number of deadly crashes. It all depends on how much you are willing to spend, how soon, and where. I suspect the pain vs cost threshold is about balanced in the States.
Great article and discussion.
I wish drivers would get the hands free phone or just pull over and talk, and stop texting while driving. There's nothing scarier than seeing a 17 y/o driver coming at you in a minivan with a cellphone stuck to their ear or texting while looking over the top of the steering wheel at intersections.
I'm not buying the "we accept the risk" argument, not when we are coerced into either accepting the government we get or leaving the country.
I have no say in traffic laws, speed limits, enforcement policies, or the penalties for violations.
Traffic enforcement has been corrupted into revenue enhancement, where law enforcement values earners as much as organized crime does, and, like the mob, imposes quotas on earnings. Speed cops game the system to meet their quotas with the least possible effort. Why not change the incentives: the fewer traffic deaths, the more the cops get paid, and the more deaths, the less pay?
Of the 140 pedestrian/cyclist fatalities in London in 2009, that you quote, how many of those were responsible for their own deaths by foolishly ignoring the same laws and safety guidelines that you unconditionally berate motorists for not following?
I have never, in twenty years of motoring, struck either a pedestrian or a cyclist but, despite a conservative and defensive driving style, I have to take emergency action to avoid a gratuitously wayward pedestrian or cyclist on at least a weekly basis and frequently more often.
With regard to your logic and rhetoric on the subject of not using the phrase 'accident' to refer to most road traffic collisions, perhaps you could therefore tell us how many of the 140 out of 184 deaths you quote should, in fact, be classified as 'moronic inadvertent suicides without whom the gene-pool is palpably improved'?
@Me!: I don't know the details of how people died in London. I think most of the cyclist deaths were involving people going under lorries, "trucks" in US language. Other than that, I don't know, therefore cannot reach any ill-informed conclusions.
You may say its an improvement for the gene pool, that these people have chosen to die, but another perspective is that most cities were designed for people, not motor vehicles, and trying to dedicate most of the area of city streets to fast moving vehicles has a price, one that the people who died got to pay.
Blaming the victim does not help.
How about requiring rigorous driving tests every 2 years? Every year if you are over 65 or under 21. If you fail, sorry. Try again in 3 months.
People should prove they can handle night driving, icy and wet surfaces, surprises, getting a flat while driving in the center lane, changing a tire.
How about raising the driving age to 18?
How about making it a felony to do anything behind the wheel of a moving vehicle other than driving? ANYTHING!
Drunk driving or reckless driving, first offense 1 year in jail plus heavy fine. 2nd offense, 5 years, third offense, life.
We need to stop coddling the lowest common denominator.
In the UK the statement "speeding is a factor in 1/3 of traffic fatalities." is often bandied about. What they seldom state is that "speeding" in this statement actually means a speed that is unsafe for the conditions. In fact, from UK government figures, less than 5% involve a vehicle exceeding the speed limit.
I hate it when pressure groups deliberately misquote statitics.
Am I in the right forum? Is this the one with the professional security people?
"If people wore their seat belts, didn't speed, and didn't drive impaired most crashes could be avoided..."
Seat belts reduce injury as the crash happens, I didn't think they were intended to prevent the crash.
Impaired driving is probably an issue and should be examined and addressed based on measured facts.
However, the reporting of speed "as a factor" gives little actual evidence. It would probably cost too much to investigate every accident and determine if the speed was a cause. Easier to just note it as a factor and let people jump to conclusions.
Also, slower speed means a longer travel time. That means more exposure to the risk and more cars on the road at the same time. We would have to measure the decreased risk from reduced speed against the greated exposure.
I don't know what the right answer is. But as security people we should know that common sense doesn't always give the best risk assesment.
'That means 76% of the people who ended up dead in London last year didn't say "I'm taking a risk but driving is necessary" or "we need cars". They said "I will walk or cycle in London."'
In the same way, without data you may be missing the point. How many might say "I have the absolute right of way, I can step into the street without looking and the drivers have to watch out for me."
The system only works when both sides, drivers and pedestrians, watch out for each other and respect each others rights.
And to determine how to change the system, we need accurate data.
It's always easiest to blame the driver. It means that none of the Powers that Be has to do anything -- car design, road design, traffic laws and enforcement, etc. It's all the driver's fault.
Many years ago, I saw a study where someone had studied about 120 fatal traffic accidents the way the FAA studies airplane crashes. They found that roughly one third of the accidents were caused by the driver, one third by mechanical failure, and one third by poor road design and engineering.
And two cases of murder by car.
I'd expect the numbers to shift some -- cars have gotten a lot more reliable and there is a lot less tolerance of drunk driving, but road engineering (at least in the US) is still pretty bad.
I'd be surprised if there have been any recent studies -- they'd be too likely to question deeply held assumptions.
"Also, slower speed means a longer travel time That means more exposure to the risk and more cars on the road at the same time. We would have to measure the decreased risk from reduced speed against the greated exposure I don't know what the right answer is."
As a simple rule of thumb your risk of death in a car goes up as the square of your velocity. Your risk in terms of increased exposure time goes up proportianatly to time.
So drive from A to B at 20mph or 60mph?
Normalised to 20mph 60mph has 3^2 increased speed risk but one third the time exposure risk, so on a simple premise 60mph is three times more dangerous to the "nut behind the wheel".
However the risk to others does not get reduced by the drivers reduced exposure time, so the risk to society in general goes up with the square of velocity so for the same 20 to 60mph societies risk is nine times what it was...
Then there is the issue of traffic capacity on a given stretch of road, as speed increases so does the stopping distance. There are two parts to this the reaction time (200-350mS) and the deceleration to zero speed. The reaction time is (supposedly) a constant and is proportional to speed but again decelaration is related to the square of speed.
Thus you can do the same 20 to 60mph calculation to show just how many extra cares you can pack in safely at 20mph.
Which ever way you cut it the personal freedom of the driver to put their foot down costs society many times more than that the driver gains.
In actual practice the risk of death to a pedestrian changes to a much higher power law than the simple square. If I remeber correctly the figures quoted are risk of death to a child at 30mph impact 10%, chance of a child surviving a 40mph impact 10%...
So the argument about an individual drivers exposure time risk means next to nothing when you look at the total risk to society of drivers moving at speed...
There have been several calls in the UK recently to cull urban foxes after a couple of accidents involving urban foxes and young children:
The foxes and the children were both just doing what is natural in their environment, but their interactions led to human injury, and hence the calls for culls.
Let us use the same logic for accidents arising from the interactions between motorists and pedestrians.
And so let's legislate to make legal the urban motorist cull.
"One cannot possibly expect that everything and everyone you need or want to see on a daily basis is within walking distance, or that transit is acceptable when mobility is required."
Why not? In much of urban Europe, that's what the situation is. I do have a car, but it is a luxury I don't really require. I can reach everything I need on foot, biking, by public mass transit, or with the occasional taxi ride. (There would be a lot less accidents if there were only taxi and bus drivers on the roads.) It's not even cost-effective to have your own car.
> One cannot possibly expect that everything and everyone you need or want to see on a daily basis is within walking distance, or that transit is acceptable when mobility is required.
I live in a city (Glasgow, Scotland), and either telecommute, walk or travel by train to work, and do most my daily shopping in local stores by foot, and much of the rest on the net. I have a car which gets used a couple of times a month, so that doesn't count as daily. You really need to raise your expectations.
According to the statistics published in the UK, speed is only a minimal factor, and even then 'inappropriate speed' is almost always just an exacerbating factor. The main causes are inattention and bad driving.
If we made the RoSPA or IAM tests mandatory, that would go a hell of a long way to sorting the problem. Reducing speeds through the use of speed cameras has been shown to have little effect, and sometimes increases accidents.
It would also make me feel much safer those days I can use my bicycle to commute in to Edinburgh (the A71 is scary when unaware drivers zoom past)
@Pete - I live in central Scotland and need to use my car for almost everything as our public transport system is shockingly bad - I could not work if I had to rely on it (caveat - I did used to live and work in Munich, so I have very high expectations on timeliness and efficiency!)
Speed kills. Its physics, not statistics. People who argue that speeding should be tolerated (why should i get a ticket, cops are just tring to make money) should work for a ambulance callout service for a few months.
The physics is that impact energy goes up with v^2. However reaction time is constant, so you travel a distance proportional to v before you start breaking. Then breaking dynamics means that hard braking is more effective at low speeds, as is vehicle control (a far bigger issue in accidents than is given credit). This add up to more of a v^3 effect. Double your speed, and all things begin equal, impact energy is ~8 times higher.
Now add that dissipating energy via crumple zones does not scale with speed. Crumple zones have some speed where they become ineffective. So there is a speed cutoff where risk to serious injury grows very very fast with speed.
The idea that you don't need to manage the speed of your vehicles is insane. I did fly planes for a short while. And you can speed in a circuit ( you fly into the guy infront of you). If you do, you can lose your pilots license forever if you are found too incompetent on a first offense (Well in my country, i hear the US is very lax).
If people are so sure that speeding tickets is about revenue(and that that makes it unfair or speeding justifiable), then lets stop with the demerit points and fines. Take the license the first time for a month. The second time for 3 months. The third time you never get it back.
Same with red light cameras (assuming non tampered with light timings ). There is never a good reason to run a red. Its insane that people claim its nothing but a money grab.
He is an idea. Don't speed and don't run reds. You won't support their money grab.
If you can't drive. Don't drive.
You say "In the same way, without data you may be missing the point. How many might say "I have the absolute right of way, I can step into the street without looking and the drivers have to watch out for me."
The system only works when both sides, drivers and pedestrians, watch out for each other and respect each others rights."
Good point, and again we lack data. We do know that yesterday's conviction in a UK court for someone killing a pedestrian involved the driver doing 20 mph above the speed limit, them attempting to hide the registration numbers afterwards and making threatening gestures at the family of the deceased.
"The system only works when both sides, drivers and pedestrians, watch out for each other and respect each others rights."
Good point. Pedestrians have right of way. You have no right to kill anyone walking across a road. That doesn't mean that they shouldn't look before they step out though there are some game-theory arguments why people do that in NY at junctions -cars are actually more likely to give way if they think you haven't seen them. It does mean that if they are in your way, you have to let them live.
Incidentally, speed doesn't kill, it's kinetic energy that kills. Vehicle mass is a risk factor too.
@greg: "Speed kills. Its physics"
If you go into the field of physics better be sure what you say. Speed does not kill. It is the deceleration that kills.
As to the speed limits: depending on the part of the world you live in, the fraction of them posted for good reason is different - some are evidently absurd and an excuse for putting some speed camera or police patrol to make up the quota.
In my area it is common to post low speed limits were the road has gone bad due to poor maintenance - and it is good reason. After some time, the road undergoes an improvement - but the signs get "forgotten". In an effect you have several kilometers of a straight, level road in the middle of nowhere with a 40 km/h limit where the default would be 90 normally. And guess what happens? Nobody is respecting the limit - to the point that actually respecting it is dangerous to you as you can be easily rear-ended, or induce others to pass you dangerously. The police takes an advantage of appearing there occasionally (not too often as this would sensitize drivers to that stretch of road) and thus fulfilling their quota easily.
This practice of absurd limits, even if occasional, is actually very dangerous as it desensitizes the general populace of drivers to the limits that are really relevant.
As for the red-running: some traffic lights in my city are so short-timed, that you have to brake really hard and fast on yellow, even if you go the legal speed. Truck drivers may have no chance of stopping before the intersection. For that reason I have a habit of glancing into the rear view mirror before stepping on brake in such case. Several times I have almost run the red and the truck behind me had run the red for sure. At least I have my trunk in one piece still.
First, I don't think that cities are for pedrestians, in a big city you can't normally go on walk to your work, you have to use other transport.
In addition, I think that the "priority" of regulation is incorrect: pedrestians, bikes motorbikes cars and trucks. These priority rules arise from the weakness instead of the inercia. Regulations should make it easier to drive a truck/bus, car, motorbike, bike and pedrestian.
In my country, mainteinance signals are often ignored because they stay much longer than the mainteinance and are not adecuated. I think that this is a greater risk. When signals are ignored because are not adecuated it is dangerous because people gets used to ignore them and also because if you follow them you can become an obstacle for other drivers.
Finally, a small budget for roads produces very dangerous spots.
@Brian: "I think we 'accept' any residual risk (whatever that means) because its a controllable risk."
I would have qualified that as "perceived as controllable risk" because, as I long ago asked my passengers to wear seat-belts (before it was enacted as law): "You don't wear them because you mistrust *me*, you wear them because you can't trust anyone else".
And, really, it is often a matter that people's skill and situational awareness vary EVEN BEFORE cell phones became common (some people, I believe, should not drive with passengers capable of holding a conversation)... and this is before we factor in little mechanical issues or damage to the road surface or just out-and-out debris or *deer* in the roadway (or, as tangles traffic here in Florida, some *birds* that slowly walk across roads).
So you may have control of your own vehicle (up to a point) but you have NO CONTROL AT ALL of any of the other vehicles on the road.
Thirty years ago I commented to friends that, in a snow storm, only two kinds of people were on the roads: fools and idiots. Idiots were driving even though they weren't exactly *competent* to be driving but had their own reasons for doing so... and fools for being on the roads WITH the idiots. I will grant that I might've spent time on both sides of that divide at different times in my life...
So a "controlled" risk is limited to one's span of control... and of trust.
"If you go into the field of physics better be sure what you say. Speed does not kill. It is the deceleration that kills."
I am in the field of physics (day job) and you are just being pedantic. Cause and effect and all that.
Otherwise thoughtful post.
One thing we need, to reduce risk, is safer alternatives. We don't need to use them all the time, but if they're available when we need them we can increase safety.
Suppose I need to go somewhere. Assume I'd normally drive. Now, assume that I'm impaired somehow (alcohol, fatigue, stress, whatever): do I have realistic choices to go somewhere without driving? If I develop a chronic condition that impairs my driving (possibly old age), can I continue to function in society?
The US is really badly set up for this. Public transportation is usually bad. Urban sprawl means that walking is often not practical. Zoning regulations frequently mean that I can't live within walking distance of places I want to go to.
I don't have a solution for this, but it's worth considering in any comparison of US and European traffic deaths.
If I tell you that reducing the speed limit on a given road will save a child's life; roughly 50% of you will agree to it; but then if I tell you that the child in question is yours... 100% will agree.
Yes, we do tolerate too many deaths on our roads - we as a society tend to take the same view that Ford took with the Pinto. Which isn't smart.
On thing I'll be watching with interest is how cars developed with the intent of being fuel efficient will correllate with increased fatalities in accidents, particularly as these cars become more common.
Not saying good gas mileage is a bad thing. Just considering that sometimes the bigger cars that require more gas to power are safer to be in when there is an accident.
Of course, I can see where they will adapt counter measures accordingly, but it is a concern none the less.
@uk visa: "...if I tell you that the child in question is yours..."
If you could do that, why not just tell people when and where, so they could avoid that given road. Why with your latent psychic abilities you could be a superhero!
". Jus considering that sometimes the bigger cars that considering that sometimes the bigger cars that require more gas to power are safer to be in when there is an accident. there is an accident."
You could call it the "tank mentality".
That is the driver perceives their own driving is adequate BUT their perception of other drivers is that they are homicidal maniacs out to kill them. Thus they require armoured plating to protect themselves and their loved ones.
Now I'm not going to say that it is ther wrong view point especialy where the US road mortality rate is something like twice that of northern Europe.
But if the driver of a tank does have an accident then they have to accept that they are going to do both themselves and others a lot more harm than if they where driving a vehicle of a lot less mass.
Basicaly you cannot have it both ways and it is about time we started to get a grip on "the nut behind the wheel".
Now here is a mad technology idea for you,
Have several levels of driver the lower the level the lower the maximum speed you are alowed to do. However the size fine you pay is multiplied by the maximum speed you are alowed to do except for breaking your speed limit in which case you get fined at the next level above the speed you where actually doing.
It can be done with tachographes and smart cards quite easily (the technolog is already in European lorries and buses).
The point is you chose the level you want to drive at if you want to do more than 30mph then you have to pass an additional test and pay a higher fee for the licence.
Oh and the privilege for paying a higher rate is that you can drive on roads where there is a minimum safe speed like freeways etc, which lower licenced drivers cannot.
That way drivers can make "market" choices about how they drive...
Good food for thought, but I must say I disagree with your "levels of driver" idea, though some of the rationale makes sense.
The reason is that it would significantly increase the speed disparity between the fastest drivers and the slowest drivers which leads to more lane changes, more frustration, and ultimately, in my opinion, more accidents.
"Seat belts reduce injury as the crash happens, I didn't think they were intended to prevent the crash."
Fair enough. Clumsy editing on my part. But wearing seat belts significantly reduce injuries and deaths when crashes do occur...reducing the number and severity of _deadly_ crashes.
I'll stand by the generalizational that speed contributes to crashes - both their cause and their severity. As others have noted if you unintentionally collide with something or someone you were driving too fast for conditions - road surface, nature or quality of the vehicle you're operating, visibility, reaction time, your skill, or your ability to pay attention to the task at hand. The faster you are traveling when you have your unintended collision the greater the energy involved and greater the risk of injury or death (especially if all involved are not wearing their seat belts).
As for lower speed limits increasing exposure to accident I agree there are many factors to consider. One of them: Traveling to a destination at 40 mph may take twice as long as at 80 mph, but the kinetic energy involved in a collision at the slower pace will be only a quarter that of the higher speed crash. Someone else will have to do the rest of the risk assessment maths.
Don't drive impaired, always wear your seatbelt, slow down and live... We can choose to do or the nanny state can make the necessary arrangements.
@ Clive Robinson
When it comes to mad technology ideas...
How about something simple like a governor that limits a vehicle to the speed limit? Or, if you desire market oriented strategies let a smart card rack up the speeding tickets automatically whenever we exceed the posted speed. In that way the well-heeled can still choose to speed but the infraction will be automatically registered and fine debited from our account. Hmmn, the smart card involved could even be our operator's license. Parents, employers, or the courts could still implement the speed limit governor for persons or vehicles for whom they are responsible.
When it comes to impaired driving vehicle interlocks can easily be made that will not start for a person with a BAC over the legal limit. An even smarter car might call the police to tell them an inebriate has tried to start his car.
Finally there are many ways to make use of seatbelts a mandatory step in a vehicle's operation.
Of course we can accomplish these same effects through our personal actions. For now, it's our call.
A lot of the commentary on the NYT site and here has focused on speeding, and the drawbacks of speeding. Speeding has benefits too.
The risk/reward calculations performed by many in the comments above forget an important thing - most occasions of speeding, even serious speeding (in excess of 20 mph over the limit) do not produce an accident. In fact, they go completely unnoticed.
Speeding has serious benefits - it shortens the transaction time/costs of transportation, and as a society of speeders, we here in the US do not demand - and often actively oppose - strict speeding enforcement. This is because of conscious decision making. This concept expands to driving at large, and why people accept traffic deaths.
Vehicle transportation is under the control of the driver, and much (but not all) of the risk inherent in driving is controllable by the driver. Driving slower, avoiding late-night driving, etc, can drastically lower the risk of accident/death. Of course, other drivers can cause an accident without the involvement of the responsible driver, and there's some risk in that as well.
All in all - when measuring the risk of an action, you must not only measure the potential risk of the act, but the likelihood of such a risk. You must weigh that against the benefits of that act, and only then you can figure out if the act is worth restricting.
@mcb: Note that your "don't drive impaired" can conflict with driving slower. Once I leave work, I can probably drive four hours or so reasonably safely. If I'm at my destination then, that's fine. If I drive at half the speed, I'm going to be on the road for another four hours and driving while impaired.
There's also the question of whether an extra hour's driving to avoid a risk that will reduce my lifetime by an average of ten minutes is worth it.
The risk assessment math is complicated. If I start simplifying it, I can get a large variety of different conclusions depending on how I simplify it.
"If you go into the field of physics better be sure what you say. Speed does not kill. It is the deceleration that kills."
If I am a pedestrian, or in a slower car, and your car hits me, it might be the *acceleration* that kills. If you are going to be pedantic, get it right. :)
MRSG says "Vehicle transportation is under the control of the driver, and much (but not all) of the risk inherent in driving is controllable by the driver. Driving slower, avoiding late-night driving, etc, can drastically lower the risk of accident/death. Of course, other drivers can cause an accident without the involvement of the responsible driver, and there's some risk in that as well."
Control is an interesting issue -we demand and pay for far more safety in planes and trains because we don't like being passengers in them, we know we aren't in control. But, to return to London's 2009 stats, if 140 of 184 deaths weren't people driving, the people dying aren't the ones making the balanced decisions about what speed to drive at. They aren't the ones with the most kinetic energy in the "Event", to use a physics term.
@Geek Prophet: "If I am a pedestrian, or in a slower car, and your car hits me, it might be the *acceleration* that kills."
Perhaps it is the difference between the directional accelarations of the entities involved in impac that kills. :) If I'm going 64 mph north and you hit me going 65 mph north, death is unlikely. However, if you're going 64 mph south then the mathematically directional differential is approximately 129 mph and my body will likely be spread over a wide area...
okay, so i'm just being a goof off. :)
@Clive: "As a simple rule of thumb your risk of death in a car goes up as the square of your velocity. Your risk in terms of increased exposure time goes up proportianatly to time."
I'm wondering where you came up with this rule of thumb. The first part assumes that risk of an accident is constant or rises with speed and also that risk of death is proportional to the force of impact. It is entirely possible that neither of these conditions hold for reasonable speeds. (e.g. it might be possible to design the passenger compartment so that the risk of injury only goes up as n log n instead of n^2)
The second part begs the question of whether it is more appropriate to measure traffic incidents per mile, per minute, or on some other basis. If traffic accidents happen at a constant rate per minute then you are (perhaps) better off driving faster and getting off the road. If they happen at a constant rate per mile then your speed is irrelevant to your odds of being in an accident.
@ted: "How about raising the driving age to 18?"
I have reservations about this solution because the only way to learn to drive is to drive. This would just move the accidents due to inexperience from the 16-18 age group up to the 18-20 age group.
@JimFive: "I have reservations about this solution because the only way to learn to drive is to drive. This would just move the accidents due to inexperience from the 16-18 age group up to the 18-20 age group."
I would have to agree. Further, the 16-18 age group usually doesn't travel far from their home where they are becoming more experienced on roads they are familiar with. To up it to 18 would put more inexperienced drivers in areas they are unfamiliar with when they go to college.
@uk_visa: "If I tell you that reducing the speed limit on a given road will save a child's life; roughly 50% of you will agree to it; but then if I tell you that the child in question is yours... 100% will agree."
While that may be true it just points up the problem with making policy based on personalized emotional arguments. You're argument is useless because you really CAN'T say those things with any degree of certainty. In addition, that argument is the same for anything that you want to use it for. "Ban spray bottles? If it was your child being blinded you would be for it." Policy shouldn't be made that way.
Most of the comments are missing the point by debating specific ways to try and make transportation safer or assigning blame for deaths. The point is, lots of people are dying and we don't even *care* because it's part of the background noise of daily life. People aren't outraged and looking for solutions.
That's how it usually is with public health issues, which are more or less *defined* as "stuff that kills a lot of people that we rarely think about." It's not news, so nothing gets done.
Unless you want to argue there's *nothing* worth doing to reduce traffic fatalities and so the problem isn't worth society's collective attention, Bruce has a point. And I don't think many of us have the expertise to say we know nothing should be done; "nothing should be done" is a pretty strong statement to make, and I for one don't even know all the proposals that are out there to improve safety.
indeed, US driver tests do not actually expect the driver to know how to *drive*. Just how to stop at a stop sign and parallel park, and other skills that will not serve you an iota of good when someone is barreling head-on towards you in your lane because they misestimated the time they need to execute a pass on a two-lane road.
People need to *learn how to drive*.
Traffic reporting in my local metropolis calls them 'crashes' rather than 'accidents.' Sometimes, it's an intentional crash, sometimes it's an accidental crash. It's always a crash, but it's not always an accident.
>>We do everything in our technological power (see, crash test ratings on cars) to ensure driving is safer.
Hum, not really. No roads yet have electronic gizmos that tell cars the speed limit. It'd be no fun, but would help reduce speeding. Or simply limited cars based on max speed limit in that state (sure, then everyone would try to register in Montana, except, not really...)
Or simply let people pay higher insurance premiums & registration fees for higher speed governors (if we truly think absolute max speed is actually the problem).
But all that is junk. My personal favorite is making people more accountable. Seat belts, airbags, and insurance take away most of the risk of anything unfavorable happening to the driver. So drivers drive with impunity. Take out seatbelts, shave everyone's tires, and force insurance to only cover half the damage you cause. Then people might quit eating McDonalds while texting and applying eye-liner...
@JimFive: "I have reservations about this solution because the only way to learn to drive is to drive. This would just move the accidents due to inexperience from the 16-18 age group up to the 18-20 age group."
@HJohn: I would have to agree
Just to confirm your thinking: The legal driving age in Aus is 18. The high risk age group here (according to the insurance companies and the road safety advertising) is 18 – 25.
18 is also the legal drinking age – er… whoops…
@ HJohn, mcb,
They are good points and we need to look at all the ideas to see what our options are before we can make choices.
I'm old enough to remember "we have the technology to rebuild him" of Steve Austin, the point being you can look at a car etc as an extension of a persons capabilities as are all force multipliers or tools.
Ultimatly the question comes around to do we want simple tools with all their implicit dangers, or do we want tools smart enough to protect us not just against others but against ourselves as well.
Is it a Faustian Bargin we make with technology, or is it the good of the many verse the freedom of the individual?
The technology to reduce the carnage of driving down to an acceptable level is out there but at the moment it is expensive. However we are introducing it in Europpe and other places for commercial vehicles to make ordinary motorists safer.
We need to start talking now about what we are prepared to accept and what we are not. We know that being a member of society requires us to give up some freedoms (those that we generaly call crimes or torts) , but we also know that if we give up the wrong freedoms we are nolonger a society in which we would wish to live.
It has been said "there is no such thing as an act of God" that is they are either acts of man or things we can reasonably predict.
In the UK we are already talking of the desirab ility of zoning legislation to stop people building on flood plains etc. That is society is thinking about stoping a personal freedom to reduce the downside cost to the rest of society.
"I'm wondering where you came up with this rule of thumb."
The simple reply is it's Newtonian mechanics for V^2 and Occams Razor on the probability of an accident as being uniform with time. Which would in all probability be the worst case argument against my argument of V^2, as it brings the normalisation of risk to the driver down from N^2 to N^2/KN.
However it still leaves the risk to society as being K(V^2).
Which I think you understand with,
"The first part assumes that risk of an accident is constant or rises with speed and also that risk of death is proportional to the force of impact."
However there are several issues with this that I tried to avoid as it would make the thumb somewhat "knarlly". The first two being,
1, Safe Stopping Distance.
2, Plastic limit on human physiology.
It is assumed that the SSD is coming to rest with no damage to the driver, vehicle or society in general.
However as we know there is a very firm limit on either acceleration or decceleration which is, at a point the human body fails irrespective of how you protect it externaly.
Now there are already production road vehicles that can under ordinary conditions deccelerate at a rate faster than the human body can tolerate unsupported which is why we do get whiplash injuries even though the vehicle does not impact with anything.
Which brings up your point,
"It is entirely possible that neither of these conditions hold for reasonable speeds. (e.g. i might be possible to design the passenger compartment so that the risk of injury only goes up as n log n instead of n^2"
In engineering terms this is a "non argument" that is although it can be done, the simple fact is beyond a very early (still almost linear) response area the costs become prohibitive for any single solution. There are however engineering "sweet spots" where multiple systems staying well within the cost envelope can be combind to give a much better level of protection for the same price as a single system (ie seat belts, head rests and airbags that deploy by deceleration not impact).
But and it's a big but the fragility of the human body can only be mitigated by restraint to a slightly higher level. That is even if you where fully encased at a certain G force your heart would still tear lose in your chest destroying your major blood vessals and you would die in just a matter of minutes irrespective of what medical assistance could be provided.
But importantly all of that technology you allude to only applies to those inside the vehical not those outside who still are subject to the "newtons cradle" effect of the energy transfer on impact.
So from a simplistic societal perspective all of that protection is detrimental to society because raising the speed at which the driver suffers minor injury (whiplash) takes those outside the vehicle into the high probability of death or serious injurie area.
Unless of course they are in another similarly protected vehice...
Thus we have an interesting perspective point.
"Protection is both offensive and defensive"
In a vehicle it protects the driver against their own stupidity and also protects other not so stupid drivers from the stupid drivers (I say stupid advisadly as surely nobody would argue that any driver can be unaware of the risks of speed and still be competant to drive?).
The contrary perspective is that of those not in a vehicles, in the proximity of vehicles. That is pedestrians, cyclists, motorcyclists and maintenance staff. They correctly view any improvment to driver safety by improvment to the vehicle as being offensive. That is it activly encourages faster driving by drivers and provides absolutly no additional protection for them. For them it is still V^2 and Newton's cradle effect that is important.
Which brings us back to the SSD, the simple solution is to segregate drivers from pedestrians etc by the SSD. However that involves real estate and as Mark Twain once observed "they are not making it any more". That is it is NOT an infinite resource so it's unproductive use should be minimised.
There are two basic solutions reduce the SSD by reduci.ng the speed drivers can go at or errect barriers to absorbe the impact.
The later is very very expensive so reducing SSD is in many ways the prefered option. It also has many many benificial side effects as well.
I could go on about this but hopefully I have sufficiently made the point that,
"From societies perspective V^2 is the important factor"
And it is the "rights of the many" (ie society) versa the "rights of the individual" (ie a driver) that we should be considering as the base case.
Which brings me around to the journy time element and the probability of an accident which you quite rightly bring up with,
"The second part begs the question of whether it i more appropriate to measure traffic incidents per mile, per minute, or on some other basis. "
Simple answer is that traffic incidents by and large depend on the "nut behind the wheel" thus you have to take a 20,000ft view aproach.
Very obviously incidents per mile varies wildly from road to road irespective of the level of traffic on it so per mile is going to be a bad choice. Time however is more reliable, even over a relativly short measure of a day things appears to average out. That is due to human nature we tend to do the same things at the same time every day. So although the risk late at night is higher than say mid afternoon, over the average journy of 15mins the accident risk rate is not going to change much on a minute by minute basis. So extending a short journy time from 15mins to 20mins will still be within a particular average of the day, so the risk would go up appropriatly for an individual driver if nothing else changed.
Which is the point you where making with,
"If traffic accidents happen at a constant rate per minute then you are (perhaps) better off driving faster and getting off the road."
But what if every body is forced to slow down and thus increase their journy time by 1/3?
What would happen to the accident rate, would it stay the same or would it go down or up?
Well from the limited analyses I've seen a reduction in speed from 75 to 50mph better than halves the number of accidents even though the density of vehicles per road mile is nearly doubled...
This 4:1 change (halved rate, double the number of vehicles) for a 1/3 reduction in speed tends to sugest that it is speed that is the major contributing factor to (reported) road incidents.
I would agree that the figures are open to argument as insufficient rigorous testing has been done. However it is sufficient for most UK Motorway control authorities to apply speed reductions not just to reduce incidents but actually to get a greater volume of vehicles through any given stretch of road in a given time.
It is without doubt a complex subject that it would probably be immoral to test rigorously as it would involve significantly raising the risk of injury and death.
For those that take issue with the term "Accident":
Keep in mind that the word simply implies that the incident was unintentional. Very few drivers/pedestrians/cyclists go out with the intention of plowing into x vehicle/person at y intersection on z day. It is usually the absence of thought that results in traffic injuries and fatalities.
Yea, you are right. Its perceived controlled risk that I meant.
It reminds me of an old joke: A passenger is riding in the car with a reckless driver. They fly through a red light. The passenger goes "why did you do that". The driver responds "my brother does it, why shouldn't I". They do this a few more times. Then they approach a green light and stop. The passenger asks "why?". The driver responds with "My brother may be coming the other way".
OK, so I suck, I used an absolute. We do everything that we can reasonably be expected to now. We can always do more, no matter what we are currently doing. I like the idea of 'electronic speed detectors'. I know certain Nissan cars will alert you that you are voiding ones warranty if you go onto a track with the car, and even flag the onboard computer of such a violation. It also reminds me of a story of OnStar, where the person was speeding in an egregious sense (180KPH), where they came on and said "if you continue, or do this again, your warranty is void". Probably a good thing. You don't even really need the detectors in the road, just a database in the car (or on the net if the car has an internet connection, assuming security on the car...). Perhaps it could even dynamically 'force' the speeder to slow down... Ensure the driver doesn't speed at all. But, of course, how could they make money that way? Its a good idea for safety, but makes no economic sense. After all, you have to pay for this technology somehow. Not that I'm saying its all a cash grab; on the contrary. You just have to pay for things in some way.
Yes yes, your cities are built with houses and businesses intertwined. I used to walk to work as well, then my office moved (but I could have gotten another job regardless), and now its a 40 minute drive to work daily. Around here, everything is sprawled out, and public transport is horrific. Driving is the only option that works. Spending 2 hours on a bus or bike when you can spend 30 minutes in a car, and have the flexibility, is invaluable.
"Keep in mind that the word simply implies that the incident was unintentional...It is usually the absence of thought that results in traffic injuries and fatalities."
While the latter statement is regrettably true, an accidental crash ought to be regarded as different from one caused by negligence. There's value in maintaining a category of mishap resulting from equipment malfunctions (blown tire) or a sudden change in road conditions (rock fall). Likewise, it's appropriate to hold drivers accountable for crashes caused by their being inattentive, careless, negligent, or wreckless. $0.02
"Likewise, it's appropriate to hold drivers accountable for crashes caused by their being inattentive, careless, negligent, or wreckless. $0.02"
Where does me sneezing and reaching down to get a napkin to wipe my eyes due to allergies fall in line?
There are cases where negligence does play a part. A pedistrian was killed by a car while the driver was reaching down to fiddle with the iPod. It went out of control. The driver was changed with negligent homicide.
Oops. "Reckless" came out "Wreckless." In the context of this discussion wreckless is what we're looking for...
> We do everything in our technological
> power (see, crash test ratings on cars)
> to ensure driving is safer.
The important thing we *don't* do is require people to drive in a safe and sane manner if they wish to retain their driving privileges.
In the first place, the amount of training and the rigor of the testing that we require for people who wish to operate a vehicle on the public roads is laughably inadequate. We hand licenses out pretty much like candy, to anyone who seems like they could with luck maybe get safely to the grocery store and back once, nevermind about every time. You don't have to take a test at night. You don't have to take a test in the rain, in high winds, on slush, or in the fog. Yet, having passed a fair-weather daytime test, you're allowed to drive under these adverse conditions. The vision test is an absurdly lenient farce based on stationary enlarged boldface type right in front of your face. At minimum, we should be testing your ability to read short words in normal-size type that flash briefly through your peripheral vision at an angle. The maneuverability test is totally mechanical, a memorized shtick. I could program a robot to pass it, but the robot wouldn't be able to parallel park in the real world, much less drive safely. We don't test the prospective driver's ability to ignore potential distractions and focus on the task of driving for a solid hour. We don't test the driver's ability to safely get the vehicle off the road in the face of common minor vehicle malfunctions, such as a flat tire. Basically, we don't test *anything*.
Additionally, we're far too reticent to revoke a license once it is issued. Things that currently *don't* get your driver's license permanently revoked and should include... driving while intoxicated; falling asleep behind the wheel; running a red light; driving with your head under the dashboard because you're fishing around on the floor for something you dropped; proving beyond any doubt your inability to drive safely by having a collision with a stationary object while traveling faster than parking-lot speed; rear-ending another vehicle at faster than parking-lot speeds; I could go on, but you get the idea. We don't take unsafe drivers' licenses away, because as a society we are willing to put up with continued unsafe driving.
> Where does me sneezing and reaching down
> to get a napkin to wipe my eyes due to allergies fall in line?
Same place it falls if you were flying a commercial airliner and crashed it while making the same move.
As a driver, you are swinging one or two tons of speeding metal around in close proximity to unarmored pedestrians.
If you cannot do that safely, then do not do it. Allergies are a reason not to drive, not at excuse to drive dangerously.
Driving safely is your responsibility.
"Allergies are a reason not to drive, not at excuse to drive dangerously."
Ideas like this are a reason not to post on the internet, not an excuse to post ridiculously.
I'm afraid you are the safety risk around here. My superior rectus muscles were damaged from rolling my eyes so hard after reading your comment. We need to curtail your freedom of speech immediately!
Posting intelligently is your responsibility.
So... most of the writers in the linked article use the absolute number of fatalities rather than the rate per million kilometers driven, to make it look like driving has become more dangerous, rather than the opposite. They call accidents "alcohol-related" if a passenger had too much but the drivers were completely sober. A speeding-related accident is one where the number on the speedometer is greater than the number on the sign, regardless of whether or not that number was set according to the MUTCD.
These writers hold up the national maximum speed limit and minimum drinking age as achievements in safety. They propose increased enforcement, which by a happy coincidence increases the revenue that pays their salaries.
Pretty much exactly the same theatre and counterproductive policies that we see in security, in other words.
Some good points: Mr. Graham mentions that small cars are less safe. Mr. Kissinger both recognizes that road safety is continually improving and calls for evidence-based solutions. Pity that's not a view that would earn anyone a nomination as Secretary of Transportation.
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.