Hacker News Reader

Comment by sethherr | Hacker News Reader

This is my huge problem with the article - it argues we have no control over speeding. And then it suggests places that ticket aggressively are behaving irresponsibly, to the point of villafying a town in Missouri for enforcing its speeding laws. All while skirting the fact that there is a clear correlation between higher speeds and higher fatality rates.

Everyone knows a place cops like to hang out, and slows down accordingly.

Why not enforce speed limits more frequently so people obey them? Why suggest a worse solution (encouraging all drivers to go faster) instead of punishing more people for driving irresponsibly?

View on HN

The real way to make people drive a certain speed - is to design a road that makes people drive that speed.

It's been a really slow process, but bit by bit we are beginning to recover from the mistake of treating the design of transportation networks as a technical engineering problem rather than a social design problem.

If you want people to drive 25mph, you have to design a 25mph road. You can't treat it like an engineering problem and include a "50% safety buffer", because then you designed a 37.5mph road and that's how fast people will drive.

In the UK we have "average speed cameras". They are dotted along a road, and monitor cars individually by their license plate.

I would argue they make drivers far less safe, because whilst on such a road, drivers are glued to their speedometers to make sure they're not going over the limit.

Drivers should be concentrating on driving safely. Not on driving at a specific arbitrary speed.

Surely this isn't something we need to guess at, there are roads with such systems and similar roads without and those can be compared.

Also I would think the majority of cars have cruise control (the last three cars I owned were all less than £2000 second hand and all had it) - driving at a specific speed is a button press away.

> Surely this isn't something we need to guess at, there are roads with such systems and similar roads without and those can be compared.

It's mainly done for political reasons or money raising reasons rather than safety. These cameras are usually on motorways which are the safest roads to drive on. It's all just for show, it doesn't actually make us any safer.

At least in Scotland, they're predominantly used during road works (and they do vanish afterwards). I can imagine they're worried about incidents especially when there's at times only a line of cones separating cars from workers.

Or they could just, you know, use cruise control.

I almost never touch the gas pedal once I've reached the speed limit: I just turn on cruise control and focus on the road, rather than my speedometer. Makes me a far safer driver, with the added bonus of also being a law-abiding one.

I would also say Cruise control results in more dangerous driving. If you're traveling at a constant speed, things get boring FAST. It's easy to just get bored, stop paying attention and zone out.

Cruise control isn't as big a thing in the UK. We have bendy roads and drive manual cars not automatics.

My suggestion for people who find cruise control makes driving boring, disengaging, and thus more dangerous is to instead focus on other aspects of driving - perhaps unobvious ones. It's surprisingly effective.

Personally, I use cruise control very often, and the attention freed up by it I typically put toward maintaining lane position and checking my mirrors more often. It keeps the trips engaging and makes me a safer, more consistent driver.

I'm also convinced that traffic flow would be much better if people used cruise control more often. Speed variability seems to cause excessive lane changes as a driver inadvertently slows down and the cars following attempt to pass so that they don't have to slow down too.

I rarely ever use cruise control. I actually find it useful to be able to modulate the speed of the vehicle via the accelerator. I find that not using it also helps me to adjust to potential situations that may be developing ahead e.g., letting up on the accelerator to reduce speed before even having to apply the brakes). I also will speed up while passing in order to reduce the time I have to spend in the passing lane.

I don't know whether or not this is related to the use of cruise control, but I have observed many drivers using their brakes on the highway even when it's clearly not necessary. For example, they'll slow down behind a car before switching lanes instead of anticipating the traffic situation and either letting up on the accelerator before catching up to the vehicle in front of them or adjusting their speed and moving into the adjacent lane to pass.

Some of this is simply the fact that it's still the case that many cars in Europe don't have cruise control: it isn't driver perception, it's simply the car not having it. On many cars it's something still restricted to higher trim levels or an optional extra.

I'll also say that by and large my experience in the UK is on most motorways traffic is sufficiently busy that cruise control is of relatively little use: there's enough variability in the speed of the traffic in front that you can't actually stay at a constant speed and there's enough traffic that you can't just pull out to overtake.

"Lane changes" aren't really much of a thing in the UK either.

Firstly, in the UK they aren't 'equal' lanes. They are for overtaking only. Undertaking is illegal. So we don't just change lanes willy-nilly. We change lane to overtake someone, and then change back.

Secondly, only motorways and dual-carriage ways are multi-lane. Most roads are single lane.

What makes a road a 25mph road? One dangerous enough that people don't feel comfortable driving faster -- how is that a solution?

Yes, basically. You make traffic lanes narrower, with much wider shoulders that are textured to discourage use. If there are bike lanes, you separate them with rounded dividers. Trees in the center divider also help. As do curves, and bumpouts around on-street parking.

People will just drive over textured bumps, I've seen it. You need to design the road such that people will damage their vehicle if they drive dangerously (things like on street parking, concrete bollards, tall curbs etc).

Bollards 3, bozos 0. [1]

The Sidney Harbor Tunnel had a serious problem with oversized trucks getting wedged in the tunnel. Watch a truck drive past two sets of giant flashing overhead warning signs. They stop only when they reach the final sign at the tunnel mouth, a water curtain with a giant STOP sign projected on it.[2]

[1] https://www.youtube.com/watch?v=i_Cw0QJU8ro [2] https://www.youtube.com/watch?v=pRKA7m-tbqM

That water screen is really brilliant.

An earlier bridge for them to crash into will also do the job quite nicely.

To be nice make it out of crumpling noisy material, not concrete.

> An earlier bridge for them to crash into will also do the job quite nicely.

Except for the part where it requires significantly more maintenance to fix up the bridge and it's incredibly dangerous for drivers following, whether the truck hard-stops or gets its top sheared off.

Hanging bars seem to solve this problem. Makes a loud terrifying noise when you hit them, but doesn't leave any permanent damage.

Truckers may not notice if the problem is at the rear trailer, as it was in the video. The water curtain, although unusual, is not expensive. It's just a sprinkler pipe and a very bright slide projector.

Also, it's at the start of a blind curve -- the road design naturally leads you to slow down as you approach, so stopping short be won't be as much of a problem.

Those are good points. Bollards, tall curbs, etc are appropriate for urban streets, because they protect pedestrians on sidewalks. But curbs are hazardous to bicyclists. More generally, you want streets that seem more dangerous than they actually are.

"seems more dangerous" probably only works for so long... Surely people would get used to the "dangerous" streets and realize they aren't as dangerous as they seem.

> But curbs are hazardous to bicyclists.

You can easily cut a bypass through the curb or chicane.

True. I've gone down hard, though. At night, they can be hard to see.

In neighborhoods near me we have gently sloping curbs. They are about a foot wider than the normal 90 degree curbs I see in more dense urban areas. It is easy to notice from the car if you pull partially up onto them. I suspect it also makes it slightly easier for cars to go of the road, but that hasn't been a real problem. Cars, bike and people seem to be at less risk from the curb directly.

You could think this way: When you are driving at day on a parking lot full of cars you drive slowly, right? And when you drive on a empty highway you drive fast. Regardless of law or street signals, you would drive at this same speed, right? Why? Because the environment are different. This is the point of the article. We have been giving tickets since forever and yet nothing has really changed, and nothing will change, unless we started to design the roads with the speed they need to have.

Yes but I don't see how making every road as safe to drive in as a parking lot (i.e. not very safe) vs. a big empty highway is a solution. The problem we're trying to solve is safety, not speed. Slowing people down by making roads less safe misses the point.

Traffic calming doesn't make roads less safe.

Advocates of traffic calming measures aren't trying to literally make roads as unsafe as parking lots. What makes parking lots unsafe is the constant threat of someone pulling out in front of you, no one is advocating simulating that.

Think about roundabouts, they force you to slow down because they prevent you from driving directly though an intersection. They are statistically much safer than 4 way stops.

> Traffic calming doesn't make roads less safe.

It absolutely does when you measure it against the alternatives instead of the status quo.

The only way you can get an apparent benefit from it is through people overcompensating for danger. People are driving 30MPH, you add some danger, people overcompensate for the danger by driving 25MPH, so the result is as though people were driving 28MPH on the original road. You call this a benefit.

The problem is that it's crazy. You're designing an intentional safety reduction and paying the full cost of it just so that people will overcompensate for it.

What you should be doing is designing a safety improvement that people will undercompensate for. Make the road as safe to drive at 35MPH as it was at 30MPH, but don't make that fact obvious, so that people only drive 2MPH faster even though the road is 5MPH safer. Then people drive 32MPH and the safety level is as though they were driving 27MPH on the original road. Compared to traffic calming you have a safer road and people get where they're going 7MPH faster.

Traffic calming isn't decreasing the safety of a road, it's decreasing the perceived safety of driving at higher speeds.

>Make the road as safe to drive at 35MPH as it was at 30MPH, but don't make that fact obvious

What would that even look like? What would you do to increase the safe speed of road? You'd widen the lanes, limit access etc... All of these things are antithetical to making a road safer for non-motorists.

> Traffic calming isn't decreasing the safety of a road, it's decreasing the perceived safety of driving at higher speeds.

It is decreasing the actual safety of driving at higher speeds. And at lower speeds for that matter. The hope is that people will overcompensate for the danger, so that the safety improvement from driving at lower speed is more than the safety loss from the inherently dangerous design. That is obviously not a desirable trade-off, even when that actually happens.

> What would that even look like? What would you do to increase the safe speed of road? You'd widen the lanes, limit access etc... All of these things are antithetical to making a road safer for non-motorists.

You separate the motorists from the non-motorists. Put sidewalks and bike lanes on the opposite side of each building from the road, so that a building has motorist access on the north and east and non-motorist access on the south and west. Then the building to the north has motorist access on the south and pedestrian access on the north etc.

Build pedestrian tunnels or bridges at busy intersections so that pedestrians and cyclists are never in the road at all.

Houston has a system of underground pedestrian tunnels: https://en.wikipedia.org/wiki/Houston_tunnel_system

None of those pedestrians has ever been hit by a car at any speed.

Safer for drivers - maybe. Safer for pedestrians? No.

>Think about roundabouts, they force you to slow down because they prevent you from driving directly though an intersection. They are statistically much safer than 4 way stops.

That's not entirely true, with half decent vision I can drive straight onto and off the roundabout again. Certainly in a 30/40mph limit I don't have to slow down particularly. The point of roundabouts isn't to slow people down, it's to keep traffic moving efficiently and safely.

Four way stops are just obviously an awful solution that I've never understood

The second benefit of well-made roundabouts is that they impede head-on and T-bone collisions. All accidents become side-by-side accidents.

Depends on the size of the roundabout, some are designed for 40 mph others for a more sane 20 mph.

I mean, a roundabout is a system for efficient "giving way". 20mph is only "more sane" when a roundabout is in a low speed limit area. We've got plenty of them on 60mph zones - you've got no interest in having cars slow down and accelerate again when not necessary (if nothing else because it's a waste of energy as well as time).

How is it efficient for "giving way"? If one entry is highly traficated, entries between the traficated one and the popular exits cannot get in.

By god, I hate roundabouts. Not because I have to slow down. If your exit is 45-90 degrees to the right (which a few of the roundabouts I have to go through are), you can commonly go in a straight line towards it without slowing down the slightest anyway, and going in as straight a line as possible is encouraged over here to make your intentions clear.

However, I feel much less safe going through a traficated, small-radius two-lane roundabout than I do going through a traficated intersection (making a turn or not). First of all, you have to find a gap, which is stressful and entirely unnecessary in a normal intersection. Traffic usually mean that just don't get a large enough gap to get in without accelerating hard, which isn't really safe. If you want an exit that is not the first (sometimes also the second), you have to use the inner lane, meaning that you need to find a gap in two lanes, rather than one. It seriously lacks some fair scheduling.

Depending on the roundabout, exiting can only be done by going "LEEROY JEENKIIIINS". Your mirrors point at nothing useful due to the turn of the roundabout, so the lane change back out involves going fast forward to the left, with your neck twisted as far as you can back to the right. Everyone is moving at their own random, highly variable pace, so a quick orientation is immediately invalidated. On a straight road, you can rely almost entirely on your mirrors so your peripheral vision still covers what's ahead of you.

Collisions between cars are likely milder, as the vehicle velocity is limited by the roundabout geometry, but side collisions are awful, and the chance of collision seems much higher, feeling like a game of Russian roulette. I'm normally (over)confident behind the wheel, with no problem driving like an idiot, around other driving like idiots, and have several hundred thousand kilometers of driving experience, but I get uncomfortable and stressed in those roundabouts.

Some two-lane roundabouts have fixed this by drawing the lines so that the lanes spiral out, with the outer lane going to the next exit, and the inner lane going to the one after it. In this way, going around the roundabout is a constant lane change to the left with no one beside you, and an exit is simply keeping your lane. This helps a lot, but is by no means common. Entry still sucks, though

>Collisions between cars are likely milder, as the vehicle velocity is limited by the roundabout geometry, but side collisions are awful, and the chance of collision seems much higher,

One study of 17 intersections found an 84% reduction in injury accidents. And a 100% reduction in fatal accidents. And every study I've seen is pretty similar in terms of injuries avoided.

Hmm. Those seem like very solid numbers. I wonder how the total accident count compares. Scratched cars are of course not important in comparison to a human life, but I'm curious whether my gut feeling of not being able to obtain an appropriate overview actually translate to accidents.

Fatal traffic injuries are quite rare over here (175 in 2015, which gives about 1/4th of the fatal injuries in traffic per capita of USA), with most seemingly being bicycles killed in intersections by trucks turning right in the city (We have far more bicycles than cars, and they seem to enjoy ignoring rules while simultaneously going "I HAVE THE RIGHT OF WAY AND PRIORITY OVER MOTORISED VEHICLES, SO I SHALL PROCEED BLINDLY").

I might be underestimating the intersection problem as a result of our "gentler" traffic (I don't recall ever actually seeing an intersection incident - I mostly see two-car incidents on the highway, where some idiot is too busy on their phone to see that the car in front is braking).

I agree that neighborhood roads should have roundabouts and so on. But that doesn't apply to major arteries, highways, etc.

I would agree, but my father used to drive a bottled water delivery truck. And he's seen lots of neighborhoods where roundabouts are simply not built large enough to accommodate anything with more than two axles. In California, roundabouts tend to be added into new construction, but squeezed into a space that is no larger than a normal intersection.

If they're designed properly, they should be low, almost the same height as speed bumps. That way something with 16 wheels can go straight over them while regular cars will probably go around them.

A neckdown or narrow lanes do not make a road less safe. Instead we see the opposite, traffic naturally slows down and thus the road becomes safer.

They also make the road less useful -- you have to balance safety with something. If not, it's just safer if everybody stays at home.

If the road has to close because of a traffic accident, that makes it less useful too - especially when it seems to happen almost every day. A road with less throughput that at least keeps moving can be better.

I'm a pedestrian. Many roads around me have been made more useful to you, to the point where they're useless to me. If they're all made useful, then I'll be the one who has to stay at home.

>They also make the road less useful

And that's the real goal here. This is just another avenue of attack for anti-car zealots.

You say "anti-car" like it's a bad thing.

I guess it depends on if you value freedom of movement or not.

But thanks for dropping by as an example.

I value freedom of movement - but not only through car accessibility. I value being able to walk / bike / take a bus safely and within a reasonable distance from my destination. The modern American suburb is terribly designed from this perspective.

This is not an uncommon opinion: https://likewise.am/2016/05/08/why-suburbia-sucks/

There are tricks to making roads feel smaller / trickier / slower. Search for the phrase "traffic calming" to find material on the subject.

There is a layer of perception here. People drive faster when they see lots of open space. For example, adding some barriers that obstruct sight but aren't horrible to crash into makes people drive slower without making the road much more dangerous to drive on.

Adding some trees and shrubbery between the road and the sidewalk also does a lot. Because sidewalks next to a road visually widen the road. This has the added benefit of making pedestrians feel a lot safer. No-one likes walking 2 meters next to cars driving at near-highway speeds.

>For example, adding some barriers that obstruct sight

This sounds like a recipe to increase accidents, just saying.

Experience shows they're just the opposite.

The point is not to have much of the environment hidden from sight. The point is to prevent the 'All I see is open road' feeling. On a slightly too simple note, the point is to make you feel like you need to be more careful without actually adding a lot of risk.

One might say the current road design makes roads feel so safe that the actual safety benefits get canceled out by changed behaviour.

Yes. This is a reason why a lot of one way streets are being made into 2 way streets. And why closing streets temporarily increases traffic but decreases fatalities (and eventually decreases traffic)

There is evieence that removing safety devices reduces speeds. Drivers compensate to the conditions. I found this article interesting: https://www.theguardian.com/commentisfree/2016/feb/04/remova...

Case in point: the traffic lights at the bottom of Ludgate Hill were out this morning so cars and bicyles had to negotiate the intersection without any form of communication or signalling. Nobody died.

Traffic lights are not there for safety, they are there to make cars go faster, while being dangerous for pedestrians and cyclists. People are often killed by cars turning right at traffic lights.

Well, here in Sweden the 30km/h roads are often enforced by speed bumps if the cars wouldn't slow down enough by themselves.

Here in the USA, we have speed bumps too. This is why we need to buy huge SUVs with high clearance. With a big enough vehicle, the speed bumps are not an issue. You should try it.

I often wonder what the environmental cost of speed bumps is. It is normal to drive huge SUVs. It is normal to slam on the brakes, go over the speed bump, and then hit the gas. There is also an environmental cost for broken parts that need to be replaced and oil that leaks from broken oil pans.

I bet if you combined speed bumps with a decent gas tax (with tax discounts for business use) and eliminated the "light truck" tax loophole for SUVs [1], you'd probably wipe out the huge SUV syndrome.

Of course, SUVs are highly profitable for the auto and petroleum industries, heavily represented in lobbyist circles, so it's pretty much a no-go.

[1] https://turbotax.intuit.com/tax-tools/tax-tips/Small-Busines...

> gas tax

"How to not get re-elected in the USA 101"

Combining two things Americans hate very much, gas prices and taxes.

While I generally think speed bumps are a good thing your comment made me think.

Also I like the matter-of-factly / deadpan style.

It's a solution, obviously, because people will drive the speed you want them to, 25mph, on it!

Better idea: make the road actually safe at higher speed.

If you don't want people getting hit by cars, you need to make them visible. Get rid of parallel parking. Get rid of bushes and mailboxes near the road. If there is a sidewalk, put it 10 feet from the road. Get rid of roadside advertising.

With those changes, highway speeds are perfectly safe. People don't pop out from behind opaque objects if there are no opaque objects near the road, and people don't fall into the road if the sidewalk is set back. Drivers have plenty of warning when somebody approaches the road, and there they can pay full attention because the distractions have been removed.

Giving drivers sight of pedestrians is good, but giving drivers to much open sight lines is bad. Open sight lines are exactly what make people drive faster. Mailbox near the road seem like a great idea to me. It communicates that people might be here, creates a visual obstacle that will get people to drive slower, and if you hit one, it doesn't cause a shit ton of damage (unlike, say, a tree).

As a counterpoint, having mailboxes right next to the road is a great way to wind up with your mailbox absolutely obliterated by the snowplow as it goes by...

Many ways to skin a cat. Design roads properly, sure. But I'm pretty confident that more aggressive ticketing would decrease speeding. Red light cameras certainly have that effect on red light running, and the negation just seems absurd to me. As a bonus, aggressive ticketing requires no new infrastructure.

What you propose is taking something that has essentially 0 effect on speed right now. And then doing that thing harder. Furthermore the result of not having an effect on speed but posting an authoritative sign that has little to no reflection of the actual speed has a detrimental effect on safety for pedestrians and non-automotive traffic.

There's a famous saying that goes something like this: "The definition of Insanity is doing the same thing and expecting a different result."

The article is proposing not doing the insane thing and instead doing something that studies show have the actual effect desired.

I don't think either of the following have been established:

1. We should expect existing lax enforcement to have a significant effect on speed.

2. Existing lax enforcement has a nonexistent effect compared to the baseline (no enforcement at all).

Even if you established both of these, it still wouldn't be sufficient to make the logical inference that rigorous enforcement would be ineffective for decreasing speeds. It would not be at all surprising that behavior goes unchanged if enforcement rates are below 0.1%. It would be very surprising if enforcement rates closer to 10% had no power to change behavior.

The problem is that it's really an economic problem. The existing enforcement is a revenue generation method -- everybody exceeds the speed limit, so you can put a cop anywhere who can just sit there and hand out tickets. It's net income for the government, but there isn't enough enforcement to actually change behavior.

If you added more police then people would stop speeding, but then people stop speeding -- there is no more ticket revenue. Now how are you paying for all the traffic cops?

That's an interesting technical problem. PID controllers answer a similar problem in control theory. I don't have the answer here, and I'm not certain there is one. It really depends on how many more cops we'd need to get to a state where enforcement can be rigorous, and whether we as a society prefer the tradeoff of paying for those enforcers to get safer roads. Tough question. But I like tricky questions aimed in a true direction better than easy, wrong answers like, "ticketing people has no effect on speeding."

If people were to stop speeding, perhaps there'd emerge a culture in which speeding is considered bad, reducing the need for traffic cops.

As soon as you stopped enforcing, people would speed again.

In cultures where e.g. drunk driving is considered an awful thing to do, people mostly self-regulate - there's some policing necessary, but it's not like the police are trying to stop something everyone wants to do. There's no reason it would be impossible to create a culture in which speeding is treated similarly.

There's quite a few countries in which people avoid speeding, even when there's no cameras around.

Aggressive enforcement isn't how you do that. And it's a lot harder to do it with speeding than drunk driving because it actually is a lot more dangerous to drive drunk than to drive the 85th percentile traffic speed even though the speed limit is set 10MPH lower than that.

In most of the existing places with a culture of respect for the speed limit, they have a higher speed limit.

Fair enough. The article may not have sufficiently established such a thing. But enforcement when it works is likely to only work when enforcement is frequent enough to create an expectation of getting caught. Given that this would require a significant increase in manpower to truly be effective I'm not sure it's sustainable.

I suppose you could find a way to automate it similar to red light cameras but even those have been shown to be rather ineffective at actually lowering the incidence of violations so I'm not sure anything less than the presence of a police vehicle with flashing lights that pulls up behind you would have an effect.

> But I'm pretty confident that more aggressive ticketing would decrease speeding.

It's also a good way to push people onto alternate roads which may be more dangerous or go through a neighborhood. It's also a good way to get people to completely avoid your roads and make your city or town dry up and blow away along with anybody that your roads connect.

The question is: What's the goal?

The article shows you how to increase safety.

What is your goal in busting speeders?

Where in the article is there any experimental evidence, or evidence of any kind, that increasing speed limits increases safety? They found one guy who thinks it does, which is not very convincing.

My goal in busting speeders is to increase safety. I just think the article is poorly reasoned and unconvincing in its claim that raising speed limits is an effective way to do that.

>When the speed limit is correctly set at the 85th percentile speed, the minority of drivers that do conscientiously follow speed limits are no longer driving much slower than the speed of traffic. The choice of the 85th percentile speed is a data-driven conclusion -- as noted Lt. Megge and speed limit resources like the Michigan State Police’s guide -- that has been established by the consistent findings of years of traffic studies.

>This is important because, as noted in a U.S. Department of Transportation report, “the potential for being involved in an accident is highest when traveling at speed much lower or much higher than the majority of motorists.”

And again, because everyone thinks a speed it's safe doesn't mean a speed it's safe.

People may believe it is, but phisics laws are hard to fool. Meanwhile people are well known to not understand risk and probability.

Posting speed limits that reflect actual traffic speeds decreases speed variance and decreases the frequency of crashes.

That's actually safer.

speed and severity of crashes correlation is quite established in literature, so I'm gonna wait for a study to come up with conclusive statistical data stating speed variance correlates more with injuries than absolute speed.

until then, it's conjecture, feelings and wishful thinking.

edit: then again urban and highway situations are different, so there's that. but you have truck and slow traffic on the highway anyway thus increasing car limit does increase speed variation in that situation and would be detrimental whether variance or absolute speed are considered

> speed and severity of crashes correlation is quite established in literature, so I'm gonna wait for a study to come up with conclusive statistical data stating speed variance correlates more with injuries than absolute speed.

That is if a crash happens. If you compare crash rates and fatality rates on limited access highways where speeds are higher to other types of roads, you'll find that the former is lower on both counts.

> but you have truck and slow traffic on the highway anyway

Trucks do not drive that much slower than cars unless they're going uphill or have a speed limiter. With higher speed limits, slower traffic will most likely speed up to some extent and trucking companies may remove their limiters (or increase the speed where they kick in).

Which studies? What did they show? Again, the fact that some guy claims a study shows he is right does not count for evidence.

The way to go are automated radars, that are clearly visible.

They recently installed a new radar at a local highway, right at the beginning of a speed limit.

Everyone sticks to the speed limit now.

A different section of the highway has "section control". (Cameras at the beginning and end of a section that record license plates and determine average speed). Nobody speeds there either.

With todays technology it is trivial to enforce speed limits everywhere, and people really do stick to the speed limit if they know that there are consequences.

So you like the fact that the police know exactly where and when you drove? Because the licence plate tracking cameras require that, you know.

Also, the notion that there is some universal speed limit that makes driving safe on a particular road is silly.

Two scenarios with speed limit of x: 1) sunny day, brand new porsche, new tyres. I'm driving at a sped of x+5 2) windy, snowy day, nissan sunny from the 80s with almost worn out tyres. I'm driving at x-5

Pick the safer one. I doubt anyone would think option 2 is safer, yet 1 is illegal and 2 is perfectly fine.

I see. Your'e saying: since speed limits aren't perfect in all cases, so lets not have them at all.

I do believe in increasing personal responsibility of people and having less of a nanny state, so I do tend towards that answer, but it's really not that black and white. On highways and places where pedestrians and bicyclists are not allowed - yes, no speed limit makes a lot of sense (and is available in Germany for example)

In cities and suburbs where kids can run into the street in unexpected places - more thought is needed what is the best solution.

I've actually gotten a speeding ticket because I was going 35km/h in a "construction" zone with a temporary 30km/h limit on a freaking Sunday with nobody around. That is also clearly ridiculous.

> They recently installed a new radar at a local highway, right at the beginning of a speed limit. > Everyone sticks to the speed limit now.

Has this also resulted in traffic bunching up? That is, long lines of vehicles tailgating each other?

I live near a couple of rural small-town speed traps. Yes, I have learned to set my cruise control to 45mph at the sign that is a ridiculous distance from town on otherwise normal highway to avoid tickets. The problem is, not everyone is a local, so you have people driving 45 and people trying to drive 60+ on a two lane highway with all the tailgating and reckless driving that entails. The speed trap isn't making anyone safer, it's just making those police departments richer.

Red light cameras increase accidents because people speed up if they think they are even slightly close to the limit.

Think of it this way, which changes behavior getting a ticket in 4-6 weeks by mail or having a cop pull you over today. Or worse watching someone get a ticket, or possibly noticing a flash and having no idea what it's from.

Your ideas about what red light cameras do are contradicted by the actual studies done to see what red light cameras do. https://www.fhwa.dot.gov/publications/research/safety/05048/...

Red light cameras increase rear end collisions and decrease side impact collisions; but side impact collisions are more dangerous.

The problem, of course, is that there is a better way to reduce side impact collisions -- and red light running in general -- which is to increase the length of the yellow light.

But because doing that actually does reduce the number of people who run a red light, the intersections with red light cameras do the opposite of that, because they need people to run red lights in order to generate revenue and pay for the cameras. And if you compare the results at red light camera intersections with short yellow lights against normal intersections with appropriately timed yellow lights, the cameras are killing people.

In my experience, I don't see that longer yellow lights really help much - people just keep going through until the light turns red. At least in this state, as long as you have already entered an intersection before the light turns red, you haven't run the red light.

I would think that a better strategy would be to increase the delay before turning the next light green, after the last one turned red.

> In my experience, I don't see that longer yellow lights really help much

Studies done in Virginia and Texas contradict that assertion. Increasing the yellow phase time decreased the incidence of red light running and the effect was sustained over time.

I'm interested to read a citation that shows that increasing yellow light length is superior to RLCs. Please send it my way. Thank you.

Compare the percentages for right-angle crashes on page 63 of your link to the numbers here:

http://saferstreetsla.org/wp-content/uploads/reports/A%20Com...

Assuming the number of right-angle crashes is proportional to the number of people who run red lights, the reduction due to having a longer yellow light is larger than the reduction due to the RLC. And that's before considering the rear-end collisions, which a red light camera increases but a longer yellow light reduces.

This is interesting, but it seems to mostly be focused on violations rather than crashes. I can see how the two might be related. But I also somewhat suspect that early into red violations may not be the type of red light running that causes accidents the most.

That's not what your link says, there is a modest aggregate crash cost benefit of RLC systems. It found this by: The economic analysis examined the extent to which the increase in rear end crashes negates the benefits for decreased right-angle crashes.

Rear end crashes cost less than right angle crashes. Further this does not mean there was a lower severity of crashes simply fixing doors costs more.

It examined the extent to which the rear end crashes negated the economic advantages, and found that they did not fully negate the economic advantages. They instead had an economic advantage of ~40-50k USD per site per year. See pages 5, 68, 74. This is why it's important to go further into papers than just reading the abstract.

I say they increase accidents you say, that's not true they lower costs. I say that does not mean they lower accidents. You respond yes they do, they lower cost. Do you really want me to say that's not my argument using different words?

If your point is sure they cause more accidents but that's ok because X then use that argument don't pretend there are zero costs.

PS: Economic arguments are very easy to manipulate based on what you do and do not count. So, yes I also disagree with the study, but that's not the argument I am making.

Around here red light cameras where removed. I heard it was because they where found to significantly increase accidents.

It's a complex matter. The company demands to retain control over the camera and makes the municipality pay rent if not enough tickets are issued.

Rlc issue is that incentive structure is all wrong, which lead to shortened yellows and people doing emergency braking at the slightest hint of light change.

Around here is not US and while I haven't seen anything written about such incentives I think that might be because any setup like that would be found illegal.

Your "I heard" is not a relevant response to my linked meta-analysis.

You have a constitutional right to face your accuser. Traffic cameras have got to go.

This has been covered to death on here, but your accuser is the officer of the court who signs the ticket, not the camera.

For traffic violations where a law enforcement officer initiates a traffic stop, it's clear that the officer witnessed the violation and when issuing a citation, list details such as the make and model of the vehicle, the identity of the driver, the location of the offense and what what the offense was.

If the case is taken to court, the officer will testify that he observed the vehicle committing a traffic violation (the vehicle appeared to be exceeding the speed limit), the supporting evidence of his assertion (e.g., RADAR tracking history), the fact that he initiated a traffic stop, then identified the driver, etc.

He cannot testify, under oath, that he did these things based on video or pictures that someone else took. The person that did take the video can testify to that effect though.

Has there been something more recent than this case?

http://www.thenewspaper.com/rlc/docs/2011/ca-macias.pdf

"Here, Officer Berg did not qualify as the appropriate witness and did not have the necessary knowledge of the underlying workings, maintenance, or recordkeeping of the Redflex Traffic System. The foundation for the introduction of photographical evidence and the underlying workings of the Redflex Traffic System was outside the personal knowledge of Officer Berg.

In absence of a proper foundation, the photographic evidence was held inadmissable."

The officer cannot be the accuser and introduce the evidence if he has no expertise in how the evidence was obtained.

Yes, there have been a lot of unsuccessful defenses based around the camera being your accuser, here's the first one I found.

http://caselaw.findlaw.com/ca-supreme-court/1669146.html

> there is a clear correlation between higher speeds and higher fatality rates

But according to the article, there is little or no correlation between the posted speed limit and the actual speeds at which people drive. So lowering the speed limit does not actually reduce fatality rates. But it does, as others have commented, creaet a misperception of what the actual expected speeds are on a road, which increases risk.

> Why not enforce speed limits more frequently so people obey them?

Because that would be way too expensive, not just in the increased cost of cops but in the lost time of all the people getting stopped even though they are driving at about the average speed of other drivers on the road--which, as the article notes, is the biggest factor in traffic safety.

The enforcement algorithm that most cops appear to actually use is to go after people who are going significantly faster than average, i.e., who are increasing the variance in speeds and therefore the risk of an accident.

> instead of punishing more people for driving irresponsibly?

This assumes that driving over the posted speed limit must equate to "driving irresponsibly". I strongly disagree with this assumption, and so, I suspect, do many others.

> Because that would be way too expensive, not just in the increased cost of cops but in the lost time of all the people getting stopped

You can automate it for much less than what traditional cop labour costs, and speeding ticket revenue would pay for the automation. If there was more enforcement, people would keep to the speed limits more frequently and there would be less stops. It's a control system with a feedback loop.

France is a natural experiment for that, they have a lot of automated speed traps.

They really ramped up automated radar deployment in 2003-2007, I spent 10 minutes trying to find data that would let me see at a glance if accidents went down more than usual starting in 2003 but didn't find anything straightforward.

Anecdotally, most people who speed just slow down at the traps, then accelerate back. I'm not sure that's much safer than going steady at the higher speed.

I found this about France, reporting very encouraging results:

https://hal.archives-ouvertes.fr/hal-00430965/document

> This public policy has yielded remarkable road toll statistics. The number of fatalities has diminished by over 40% since 2002, and the authorities are speaking of 12,000 lives saved since the new policy was implemented. The number of injuries avoided is also considerable, even if the fatality figures are those that draw most attention from official decision makers and the media.

The "slowing down for camera" behaviour might be less of aproblem, if camera locations are biased towards safety critical places. But it also has a fairly simple technical solution: average speed cameras. These work by timing vehicle sightings between two cameras.

> Why not enforce speed limits more frequently so people obey them?

As long as we do it consistently, actually stop people, and do it across the country... which is pretty hard. The are quite a few issues with just extra ticketing because speeding is dangerous.

The worst ideas imo are the automated checks during holidays. So you get caught on a random road, in a place you're never coming back to. You pay the ticket you get a week after the fact, and what changed? You were still speeding during the busy holiday weekend, you'll likely forget about it next year.

Another one is small speeding (5 km/h) fines in residential area. You're constantly changing direction, slowing down and starting again. Would you rather people paid attention to their dashboard to keep the right speed, or actually looked around for people/cars?

As much as people made fun of it, I liked the idea that was used in Poland (maybe still is?) where instead of actual cameras or patrols, a simple plastic shape of a car painted like the police cars was placed near the road. [0] Works great on non local people and sometimes you would get an actual police car behind it to stop people who get used to / ignore that warning.

[0] https://creators-images.vice.com/content-images/contentimage...

Consistent enforcement is not hard: photo traffic enforcement devices.

Sure, if you don't mind penalizing people who have caused no harm. I strongly disagree with that approach.

This comment is utter nonsense. First off, how is that any different from enforcement from a cop? Secondly, why do we have to wait until someone actually causes an accident? Breaking the speed limit, driving drunk, running red lights isn't bad enough?

I think I've addressed these questions already in other posts in this subthread.

So you're okay with people driving while drunk or high as well?

They might not cause any harm too.

If they don't cause harm, then they shouldn't be penalized, at least not by the law. If they cause harm, they should.

You seem to think that anything that is legal must be "ok". That is certainly not the position I am taking. Lots of things are legal that I do not think are "ok", and some things that are illegal and I think should not be (such as mere possession of drugs), are still not "ok" to me. I would not advise anyone to drive while drunk or high, and if someone does it and manages to not cause harm, they were lucky and shouldn't expect to be so lucky if they try it again. And if their parents or spouses or significant others want to penalize them for doing something so stupid, that's fine with me. I just don't think the law should penalize them unless they actually cause harm.

>If they don't cause harm, then they shouldn't be penalized, at least not by the law. If they cause harm, they should.

I think that drunk driving checkpoints are often taken advantage of to work around the 4th amendment, but making it legal to drive drunk unless you harm someone is insane. Drunk driving directly causes around 10k deaths per year in the US.

What if I decide to walk down a crowded sidewalk swinging a sword over my head, under your system is that fine until I actually hit someone? Do you have no legal recourse to stop me?

Is it OK for me to drag race down a residential street at 120mph as long as I don't hit anything?

Lets say the guy next door decides that he's going to build a nuclear reactor. Do you have to wait until your hair starts falling out to stop him?

Clearly there is a limit where a behavior is inherently dangerous enough that we've decided to outlaw the behavior before it causes harm. Does drunk driving not cross that threshold for you? Where is the limit for you personally?

You added some more items while I was posting my first response, so I'll address them separately here.

Clearly there is a limit where a behavior is inherently dangerous enough that we've decided to outlaw the behavior before it causes harm.

Clearly in our current system we do this, yes. And clearly it doesn't always work (as we see in the drunk driving numbers). And even in cases where it does "work", I think that's more because of the cost of the dangerous behavior--very few private individuals have enough wealth to be able to build their own personal nuclear reactor, and the ones that do appear to have better things to do with their wealth--than because of legal prohibitions. ("My house has anti-tiger protection." "But there are no tigers within thousands of miles of your house." "See how well my protection works?")

The fundamental problem is that some people have a propensity to do stupid and dangerous things. How do we stop them? Legal prohibitions don't seem to work very well. (If personal nuclear reactors were affordable, do you think making them illegal would really stop all the people who want to build one?) Plus, legal prohibitions preempt other possible methods, and reduce respect for the law, and undermine personal responsibility. So what are they really gaining us?

>The fundamental problem is that some people have a propensity to do stupid and dangerous things. How do we stop them?

And fundamentally the only way to stop some people from doing stupid and dangerous things is with force.

Let's say my neighbor sets up steel target in his backyard. Directly behind the target is my kid's swingset.

My neighbor insists that he is such a good shot that he will never miss. He hasn't actually caused any damage yet.

Let's assume my neighbor is an asshole and asking him (and other forms of social pressure won't work).

You've taken the libertarian anti-regulation stance much further than anyone I've ever met. Most libertarians would say that the mere threat of him missing the target causes me harm, just like the threat of swinging a sword around on a crowded sidewalk causes harm.

> Legal prohibitions don't seem to work very well.

Sometimes they do, sometimes they don't. You can't seriously look at the history of regulation over the past century or so and not realize that there are some successes and some failures.

I used to be a hardcore libertarian when I was younger, but I've learned that the world is rarely binary. Dogmatic approaches to governance rarely work, and no regulation is just as bad as a command economy.

The search for a simple set of fundamental governing principles is a search for elegance--the same as trying to design an algorithm with no edge cases and no exceptions. The problem is that I've yet to encounter a system that didn't require edge cases and exceptions.

> fundamentally the only way to stop some people from doing stupid and dangerous things is with force.

Yes, that's true. But the law is not the only way to apply such force.

> Most libertarians would say that the mere threat of him missing the target causes me harm

I don't think they would. But I think there are plenty of libertarians who would say that a homeowners association, neighborhood group, etc.--i.e., a private organization, not the government--has the right to tell your hypothetical neighbor that the rules prohibit him doing shooting practice in his back yard. And people who really, really can't live with such restrictions will, under a libertarian system, end up living far away from anyone else, because no one wants to live near them.

> I've yet to encounter a system that didn't require edge cases and exceptions.

I agree with this. But different systems can have very different edge cases and exceptions; and it seems reasonable to me to judge systems based on which ones have edge cases and exceptions that are more manageable.

In our current system, the edge cases and exceptions are instances in which regulations prevent people from doing things which would benefit them but which some rare people can't be trusted to do safely. Such cases are common: for example, to go back to the topic that originated this discussion, lots of people get stopped for speeding that would not have caused an accident had they not been stopped. And the fact that everyone knows such edge cases and exceptions are common means that people have less respect for the law--which spills over into the cases that are not edge cases and exceptions.

In the kind of libertarian system I envision, the edge cases and exceptions are instances in which private preemptive action is not sufficient to stop someone from causing harm. I would expect such edge cases and exceptions to be rare, since most people don't want to cause harm and will respond to reasonable requests to curtail potentially harmful activities. Furthermore, I would expect there to be much more respect for the law in such a system, because everyone understands that the law has a clearly defined boundary: cause harm, get punished.

making it legal to drive drunk unless you harm someone is insane. Drunk driving directly causes around 10k deaths per year in the US.

And every drunk driver who causes a death should be penalized, no question. But penalizing people who have caused no harm undermines people's respect for the law. It also undermines personal responsibility: people figure that if it's legal, it must be ok, instead of exercising their own judgment.

Actually, the very fact you quote should make you doubt the effectiveness of penalizing drunk drivers who have not caused harm. Everyone knows you can lose your license for drunk driving, yet drunk driving still causes 10k deaths per year? What sort of draconian system would it take to actually deter all those drunk drivers if the law is the only tool we have to do it? I don't want to find that out.

What if I decide to walk down a crowded sidewalk swinging a sword over my head, under your system is that fine until I actually hit someone? Do you have no legal recourse to stop me?

Once again, you're conflating "legal" with "ok". Obviously walking down a crowded sidewalk swinging a sword over your head is stupid and carries a huge risk of harm. And everyone around you should be ducking and screaming at you to stop, for goodness' sake, you're going to hurt someone. But no, the law should not punish you unless you actually do hurt someone.

Your underlying fallacy here is that the only way to prevent people from doing something is to make it illegal. That is both too weak and too strong. Too weak, because, as noted above, even severe legal penalties aren't sufficient to deter some people (including, apparently, 10k drunk drivers a year). Too strong, because there are lots of ways to deter people without any legal penalties whatsoever--but that requires us, as individuals, to take responsibility for regulating our own behavior and pointing out obviously stupid and dangerous things that others around us are doing and asking them to stop, and imposing non-legal penalties if they don't. Sorry, you can't come into my house with that sword, it's too dangerous. Sorry, you can't come into my store with that sword, it's too dangerous. Sorry... Pretty soon the idiot will get the idea that nobody is going to interact with him if he persists in swinging his sword around all the time. (Applying similar logic to people who are prone to try driving while drunk is left as an exercise for the reader.)

Look at the number of drunk driving deaths (and accidents) from 30 years ago or so and you'll notice that the rate has been dropping since enforcement became more regular. Of course some of that is likely from social pressure, but most studies have shown that enforcement does reduce the number of drunk driving deaths.

But let's not argue in vague generalities about using social pressure instead of laws.

A company builds a nuclear plant on Lake Michigan. There is a 10% chance per year that the plant will fail. The failure will catastrophically damage Lake Michigan and the surrounding area. The company decides that it's worth it.

Under your system, a boycott is the only way to stop them?

A lab decides to manufacture biological weapons to sell to dictators. There is no chance of a boycott doing any good because the owner doesn't sell to anyone in the US. There is a 5% chance per year that safety procedures will fail, releasing a virus that will kill an estimated 50 million people in the US.

Under your system, what do we do about it?

These hypothetical situations are extreme, but I'm trying to find out what your limits are. Because I really don't think you've thought this through.

>most studies have shown that enforcement does reduce the number of drunk driving deaths

But there are still 10k of them a year. Why, by your logic, don't we tighten up enforcement until there are none of them? What justifies stopping anywhere short of that?

> Under your system, a boycott is the only way to stop them?

How did they manage to build a plant with a 10% per year chance of catastrophic failure? It takes thousands of people to build a nuclear plant. Why would all of them participate in such a thing?

> Under your system, what do we do about it?

Under our current system, what do we do about it? You have described the current world situation: someone could have such a lab right now and we couldn't stop them. So why hasn't it happened?

> I really don't think you've thought this through.

I'm not sure you have either. You are assuming a world in which we have eliminated all penalties against people who have not done actual harm, but nothing else has changed from our current world. That is not at all what I have been assuming. I am trying to contrast what we have now--a system in which the law penalizes people who have done no actual harm, and which has evolved for a long time under that assumption--with what we would have if a system in which the law does not penalize people who have done no actual harm had evolved for a long time under that assumption. In such a world, many, many things would be different from today's world.

As an illustration, I'll redescribe your two scenarios:

A company wants to build a nuclear plant on Lake Michigan. However, when they try to acquire the land, they find the landowners asking what they are going to use the land for, and when they say a nuclear plant, they start getting asked all kinds of hard questions about the risks involved. They can't give satisfactory answers, so the landowners refuse to sell. Similarly, when they try to recruit people to help work on the plant, they find the prospective employees asking the same hard questions, and refusing to work for them when they don't get satisfactory answers.

A lab wants to manufacture biological weapons. However, they are unable to get funding to even start, because there are no dictators or similar insane rulers to sell to. Plus, once the lab's employees find out what the owners are thinking, they all quit.

In short, in our current world, where we allow the law to penalize people preemptively, everyone expects the few people who make the laws to be responsible for stopping people from doing harmful things. So nobody feels individually responsible for doing it. In the world I'm describing, where everyone understands that the law won't step in until actual harm is done, everyone feels individually responsible for stopping people from doing harmful things, which means stepping in preemptively based on their own individual judgment instead of expecting the law to do it.

Would such a world be free of harmful things? Of course not. But neither is our current world. The question is, would the world I'm describing have fewer harmful things happen? Would it have fewer than 10k drunk driving deaths a year? We don't know because we can't run a controlled experiment, but I think it's quite possible that the answer could be yes.

The point of the law is to discourage behaviour that we have agreed as a society should be discouraged.

In some cases, the person who might have an adverse impact of a decision is someone other than the person taking it. And sometimes the impact of that decision may be irreversible.

In those kinds of circumstances, we, as a society have to ask whether the externalised risks (i.e. the risk that person A imposes on person B, against their will or knowledge) outweigh the curtailing of rights (i.e. the right to do whatever I like) that a restriction would impose.

While I don't think that laws are the only way to prevent someone from doing something bad, I will leave you with a potential downside to what you're suggesting:

Lets say someone is rich/powerful, would the rules apply the same to them in your scenario? Should they?

> The point of the law is to discourage behaviour that we have agreed as a society should be discouraged.

Are there any studies that demonstrate that people will not engage in undesired behaviors solely because they're illegal as opposed to what they feel is right due to their upbringing and/or education?

For example, I'll wear a seat belt while riding or driving a vehicle regardless of whether there's a law that makes it a requirement.

> The point of the law is to discourage behaviour that we have agreed as a society should be discouraged.

That's the principle that underlies our current system, yes. But it's not the only possible principle on which to base a legal system. The alternate principle I am proposing is basically that the point of the law is to make clear exactly what actions cause harm and will cause those perpetrating them to suffer penalties.

> Lets say someone is rich/powerful, would the rules apply the same to them in your scenario?

Sure, why not?

The obvious scenario is firearms.

I can take concealed firearms into a public space start firing into the air with no intention of hurting people, and so long as I don't accidentally hurt anyone, the law is powerless to stop me.

No one wants to let me into their store? Ok. I'll just walk down the street and shoot into the air!

How much danger should I be allowed to put you in without repurcusions? Should I be allowed to point a loaded firearm at you? As long as I don't pull the trigger I'm not causing any harm. I'm pretty sure most people would appreciate the police getting involved in this type of situation.

Thanks for explaining your position. I mainly asked because in my opinion there is very little difference between driving high / drunk or violating speed limits. While former increase chances for incident later increase lethality and in the end only risk for human injuries / life are matter.

While I don't think that anything legal must be "ok" and law is perfect I still don't agree with you and there is serious reasons for that. First of all one can simply be "lucky" with careless driving or just straight defective brakes.

Second I seen how law enforcement changed driver behaviour in my country of origin. Even with extreme amount of corruption it's still forced almost everyone to use seat belts, pass vehicle inspection regularly and avoid breaking the law if there is chance that will end up with loss of their driving license or huge bribe. I totally sure that without proper enforcement that would be never achieved otherwise.

Third is that I seen the difference between driving behaviour in the UK and Asia and while UK system cause a lot of false positives lethality numbers make a big deal. Even if enforcement is not the only measure to calm traffic it's important one. I really can't see how one can justify that 3x-10x more people going die.

Last reason is that I simply not believe that society on it's own can change driver behaviour as effective (if at all) as enforcement. Fines and tickets work because they are market mechanism like fuel and horsepower taxes, paid parking and London's congestion charge, etc.

Also there are plenty of other laws that limit overtaking on sidewalk, illegal parking and many more that only work because there are enforced. Remove them and streets of any country will slowly turn into hell.

UPD: Grammar fixes.

> I mainly asked because in my opinion there is very little difference between driving high / drunk or violating speed limits.

I see a big difference. Driving drunk/high impairs your judgment; it means you are less capable of knowing what a safe speed is (as well as being less able to judge many other factors). So it obviously increases risk.

Violating the speed limit, assuming you are sober, might not increase risk at all, because the driver has more information about the conditions in the particular situation than whoever set the posted speed limit, and therefore is better able to judge what a safe speed is. His judgment might be (and often is, if typical behavior on highways in the US is any indication) that the posted speed limit is too low--that speeds higher than that limit are still safe. I don't see any basis for saying that he must always be wrong and the posted speed limit must always be right.

> I seen how law enforcement changed driver behaviour in my country of origin

I didn't say that law enforcement doesn't change behavior. Of course it does. The question is, at what cost? And is the cost less than the benefit? Could the same (or greater) benefit be achieved at a lower cost by some other method.

Your own description suggests very large costs that you do not appear to be taking into account: you speak of "extreme amount of corruption" and "huge bribe". What cost does your society as a whole bear as a result of these things, compared to a society in which people are held strictly responsible when they cause harm and know they can't bribe their way out of being punished?

The article does not claim that we have no control over speeding. We do have control over or speed in numerous ways. The article merely says that signs have little impact on how fast a community goes on any stretch of road... other factors end up being more of a factor. At the same time, road safety is not as strongly tied to the speed of traffic.

This article has an attention-grabbing headline (cop says to increase speeds!), but really it's drawing attention to how there are many factors that go into road usage and laws and there are places where people are starting to take these other factors into account. Sounds like a micro-level revolution in how the executive branch of US government operates.

> All while skirting the fact that there is a clear correlation between higher speeds and higher fatality rates.

The NMSL was repealed in 1995 (22 years ago). Most states have raised their interstate highway speed limits from 55 or 65 mph to 70 mph and beyond. Yet, the fatality rate per 100 million vehicle miles traveled has dropped substantially since 1995.

Therefore, higher speeds do not have a correlation with higher fatality rates.

Vehicle safety is a significant confounder here. It is possible that increased deaths from higher limits was more than offset by mandatory airbags, anti lock disc brakes, crumple zones, etc.

Actually it's pretty neat how much safer cars in America have gotten since then.

While that is true between 1995 and 2005, if you look at the data [1] between 2005 (roughly 10 years after both drivers and passenger side airbags were required equipment) and 2015, you can see the fatality rate dropped from 1.46 down to 1.08

Given the average age of a vehicle on the road is about 10 years give or take, then by 2005, the majority of vehicles had both drivers and passenger side air bags and crumple zones. Traffic speeds have also increased to some extent in the last 10 years as well.

[1] https://www-fars.nhtsa.dot.gov/Main/index.aspx

There are other new safety features introduced on many cars since then. Vehicle stability control is a big one.

Vehicle stability control is a big one.

In some situations, but in other situations it seems to be a negative (at least as implemented on my car, a '13 Hyundai Genesis Coupe).

I can see that it's helpful for low-traction situations where it's just protecting the driver from losing control.

But on my car it has a really annoying, and I think dangerous, side effect. You'll note that difference in wheel speed from left to right is what triggers the stability control behavior, and that this delta is proportionately greater for tighter-radius turns, the ultimate example of which is a right turn from a stop sign or driveway. And the effect of it is to severely limit the force applied to the wheels (whether by pulling back the throttle or by applying a brake to the affected wheel).

But that right turn, when I'm trying to merge into traffic on a road with a 60mph speed limit, is exactly where I need the power to accelerate. So I see a break in traffic, turn onto the road, a tire slips on a bit of sand, and then I'm stuck wallowing there with no power, watching my mirrors as a car barrels down on me from behind.

I've learned that I get the best results if I make the first thing I do when I get into the car, to turn off the traction control unless the pavement is wet.

The NHTSA and the Insurance Institute for Highway safety estimate that stability control could prevent 1/3 of fatal accidents.

Studies show it's effective. Of course it depends on the car, but I've never found it to be a problem unless I'm driving like I'm at the track.

Also where are you turning right onto a road with a 60mph speed limit with no merging lane?

Studies show it's effective.

I agree that it's doing a good job in keeping me from breaking my tires free, which could lead to losing control. But I'm skeptical that enforcing these tactics blindly in all cases (as my car does) is the best overall strategy.

where are you turning right onto a road with a 60mph speed limit with no merging lane?

Specifically, US 290 west of Austin TX. But this seems to be the norm in much of Texas.

>But I'm skeptical that enforcing these tactics blindly in all cases (as my car does) is the best overall strategy.

It's not optimal strategy in all cases. But driving 100% of the time with electronic stability control (for which traction control is required) on results in an overall safety benefit as opposed to driving with traction control off 100% of the time. Which is why it is a required feature in all new cars.

>Specifically, US 290 west of Austin TX. But this seems to be the norm in much of Texas.

That's crazy. I've seen something like that in very rural areas, but even then the speed limit was only 55.

This is the most dangerous thing about stability control - if you've learned how to drive without it, your instincts are going to be all wrong when it tries to compensate for what it thinks is happening. It's the same problem with ABS brakes, which can be terrifying if it's a little icy and you're expecting them to lock up properly.

I also wonder what's going to happen when those sensors get a little screwed up and start to fail. Or when you put different tires on, or make any modifications to the suspension.

It's the same problem with ABS brakes

I get your meaning, but any difference with ABS isn't going to put you in harm's way (except some really rare cases of deep snow, where the effect of your tires pushing the snow provides extra resistance). All you have to do with ABS is trust it, but even if you don't it'll work out fine.

But in this case, you've got your mind calibrated to how you know your car can accelerate, and the safety of the maneuver is contingent on the car living up to it. But the stability control makes the car say, "no, I'm not going to do it", leaving you hanging out there in danger.

Vehicle stability control isn't mandated by NHTSA, and I don't think that it's a feature on a substantial portion of the vehicle fleet.

You would have to look at what vehicle safety features were mandated between 2000 and 2010 to really see if any of them played a part in the decline in the fatality rate in the 2005 to 2015 time period.

Electronic stability control has been mandated by the NHTSA since 2012 (they required at least 55% of 2009 models sold to be equipped with it, 75% of 2010 models, and 95% of 2011 models)

Also in 2007 about half of all new cars sold were equipped with it.

The NHTSA and the Insurance Institute for Highway safety estimate that stability control could prevent 1/3 of fatal accidents.

If a substantial portion of cars started to be equipped with it around 2005 (they did), it's a likely explanation for a decline in traffic fatalities since then.

> Electronic stability control has been mandated by the NHTSA since 2012 (they required at least 55% of 2009 models sold to be equipped with it, 75% of 2010 models, and 95% of 2011 models)

I should have looked that up before stating the opposite. My mistake.

> Also in 2007 about half of all new cars sold were equipped with it.

I've read that the average vehicle age is about 10 years. That means that only now are we getting to the point where the vast majority of vehicles would be equipped with electronic stability control. Looking at the FARS page I cited earlier, the decline in fatality rate is pretty clear, but how prevalent ESP was and how much it changed in that time period is not as clear (at least I don't have the data for it).

Some very back of the envelope math shows that each year about 7% of vehicles are replaced each year, so by 2012 we can estimate that about 35% of cars would have been 2007 models or newer.

So by 2012 we have a minimum of 17% of cars with electronic stability control (probably much higher given that by 2010 75% of cars sold were mandated to include it). If cars with electronic stability control are really involved in 1/3 few accidents, then we'd expect a very sizeable decrease in vehicle fatalities due to electronic stability control alone by 2012 and probably much earlier.

After the city put in a speed limit sign with radar that shows speed, kids put up a high score list under the signs. I thought it was hilarious and scary.

> Why not enforce speed limits more frequently so people obey them?

Personally, I prefer to see police enforce more meaningful laws and investigate crimes where there are actual victims involved.

Like hit and runs?

Yes. Seems like drunk driving is a key contributor to that: https://www.usatoday.com/story/news/nation/2013/11/10/hit-an...

> Why suggest a worse solution (encouraging all drivers to go faster) instead of punishing more people for driving irresponsibly?

The article addresses this. the posted speed limit doesn't affect how fast people drive (slow or fast). They drive as fast as is comfortable regardless.

Posting a higher speed limit doesn't make people drive faster. But it does give a clearer warning to pedestrians and parents as to what the probably speed of drivers is, so they can take the appropriate cautions.

> it argues we have no control over speeding. And then it suggests places that ticket aggressively are behaving irresponsibly, to the point of villafying a town in Missouri for enforcing its speeding laws.

He doesn't argue that we have no control. Only that posted speed limits aren't effective as a control measure. The point of discussing the towns that use traffic violations as a revenue source is that, if posted speed limits aren't effective at creating safer drivers, then the only reason to use super low speed limits is to generate revenue which is a conflict of interest and counter productive for the reasons stated above.

That small town in Missouri was in fact violating the law, and as the article pointed out.

> Why suggest a worse solution (encouraging all drivers to go faster) instead of punishing more people for driving irresponsibly?

This is the same logic that failed in the "War on Drugs". First off, there's a false dichotomy between proscribing a behavior and encouraging it. And strict enforcement of a "crime" that it turns out lots of people commit on a regular basis doesn't really work so well.

I think the problematic logic of the War on Drugs is a bit subtler than "If some people want to do something, it's a bad idea to make it illegal."

Off the top of my head: Many of the drugs in question are responsible for far fewer deaths than cars are, and so it's not obvious why the penalty fits the crime and there's a strong argument for outright legalization. Many of the drugs are physically addictive and not merely something people want to do with full free will; many people arrested would be happy to stop doing drugs if they got treatment instead. The war involved foreign policy and foreign military activity, whereas nobody is suggesting that we bomb Toyota if they refuse to make slower cars. The war has been waged particularly aggressively among historically-oppressed races and classes (cf. The New Jim Crow), whereas no directly analogous danger exists from enforcing low speed limits.

The "war on drugs" was largely won in Asian countries that have proper enforcement (Japan, Singapore, Taiwan, etc...)

Just compare the streets of Tokyo to that of San Francisco and tell me it is not working!

Also, compare driver (and pedestrian) behaviour in Japan. In urban areas, both largely follows the rules and it is safer.

(Strict liability for drivers may also be a part of it...)

Alameda CA has all of that too. 25MPH speed limit over most of the city, heavy enforcement of traffic laws, narrow lanes, wide bike lanes, walk buttons with flashers, raised crosswalks for visibility. People still speed but much less so that in neighboring cities

No, that isn't really what it's saying.

Vilifying

You're right (though I like the idea of villafying a town, which I suppose would mean filling it with large Mediterranean-style dwellings) but this sort of correction to something no one was in much danger of misunderstanding doesn't really contribute much and I suggest is best avoided unless (1) the perpetrator specifically asks for corrections, (2) there's serious danger of misunderstanding as a result of the mistake, or (3) correcting it gives the opportunity to say something genuinely interesting and relevant.

Reply on news.ycombinator.com