10 ms·
Phantom traffic jams, and why lane expansion isn’t really successful
- Nerdfest 8y agoIf a lot more people drove manual transmissions, or at least learned that you can slow down by just raising the accelerator a little rather than hitting the brakes, things would be a lot better. It really doesn't seem like a hard thing to learn.
- sliken 8y agoI believe it's the reaction time lag, not using the brakes instead of the throttle to slow down. Basically people suck at staying in the ideal position half way between the car in front in the car in back. Because of that they brake hard as they come close to the car in front. The next car has to brake harder... till you end up with cars stopping. So it's the human reaction time combined with cars accelerate (generally) slower than they brake.
- rplst8 8y agoThis. There are lots of studies that have pointed to slow re-starters as the prime culprit for traffic. And it's only gotten worse with cellphones because now there is something to longer distract people when the come to a stop.
- lloeki 8y agoSlow-restarters and sudden late-brakers. This is really all about an oscillating wave being carried vs mean flow speed. The best part when you get the hang of it is that in many situations when you can anticipate the "slow-restart/sudden late-brake" of a dozen or so cars before you (and a couple behind you), you can act as a dampener as you strive to keep near the mean flow speed, and delightfully see the phantom jam downright evaporate in your rear view mirror.
- ams6110 8y agoI think it's that most drivers just don't pay attention. I'm able to maintain consistent following distances as speeds change if I pay attention to doing that. The other problem is that if you maintain a gap as traffic speeds are slowing, someone will cut in front of you to gain a car length. Then you have to slow down using the brakes.
- gascan 8y agoHuman reaction time is always there, but driving a stickshift generally teaches you to be more observant & plan ahead more- to anticipate and try to avoid extreme braking or acceleration situations. I drove stick for years, and your parent is right on. And, back in an automatic, I've naturally regressed, I freely admit.
- sliken 8y agoMy favorite related quote is that adding lanes to decrease traffic is like letting your belt out to lose weight. For more info see: https://en.wikipedia.org/wiki/Induced_demand https://en.wikipedia.org/wiki/Induced_demand
- ars 8y agoBut you are ignoring the human and economic benefits of increased travel. A better example would be someone pulling their belt so tight they can barely move, and then letting it out.
- SerLava 8y agoBut we let the belt out all that way, and the belt still touches their waist!
- ars 8y agoThen there is more demand. Keep going. There is nothing that has unlimited demand. You can't reduce demand by constraining supply, all you are doing is making people unhappy. If your policies are very free, you'll find people will figure out how make things more efficient. For example inexpensive delivery will become possible if there was enough space for traffic. I know this because I've lived in dense cities (no delivery service - it's simply too hard to find parking, and thus too expensive) and wide urban areas (every store offers delivery because it's cheap and easy). And that's exactly the opposite of what everyone thinks will happen - they want cars off the road, and replaced with delivery, without actually making delivery practical (i.e. inexpensive).
- gok 8y agoRight. So when the subway gets crowded, the ideal thing to do is reduce the number of trains per hour so that people stop trying to ride it.
- maxsilver 8y agoThat's like saying, "we shouldn't build more hospitals, because we don't want to make more sick patients". It's only "true" if you grossly botch your before and after counts.
- 1123581321 8y agoI can’t speak to the aggregate, but going from two to three lanes in each direction on I-90 in northern IL has cut travel times for us by about 8% despite increased total traffic. This is because trucks passing now leave the left lane open and because the speed limit was increased 5MPH.
- rplst8 8y agoYeah, I think that saying more lanes means more traffic is just ignoring the fact that the current road system never met the original demand. Also, we obviously have to plan for growth. We do have more people on the roads than we did in decades past. I haven't checked the numbers but I bet highway-lane-miles has not kept pace with population growth.
- sliken 8y agoWell say you have a particularly busy 10 mile section of highway. You spend $millions adding lanes. Commuters are no happier because their commute time is unchanged. Riding on public transit drops which is worse for the environment. Was adding the lanes a win for the city? The commuters? The environment? It is counter intuitive, but if you think about it. Everyone has a commuting pain threshold that will cause a change in behavior. If a stretch of highway takes 10 minutes less because of additional lanes, then people will start commuting from 10 minutes further away. Fortunately it's really easy to exactly quantify trip times, cars per lane, and similar. Quite a few studies have shown short term benefit, but within a year the traffic is within 0.1% of what it started with before adding the lanes.
- Rescis 8y ago> Was adding the lanes a win for the city? The commuters? The environment? It was a win for the local economy. By allowing more people to drive to and from locations they please, you are allowing more people to work jobs they otherwise wouldn't be able to, and purchase goods and services they wouldn't otherwise. I'm rarely a fan of government spending; but, roads, police, and firefighting are probably some of the least evil.
- gok 8y ago> Latent demand is the idea that if you widen a highway to reduce its traffic density, people who would otherwise have stayed on surface streets will now fill the newly available space on the highway, negating the benefit of having more lanes. ...if the only benefit of building more lanes was reducing mean transit time at rush hour. You could also just build a one lane road and have the toll so high that only a single driver could afford to use it; that would be optimally “beneficial” in this model.
- SerLava 8y agoThank you. I am mystified by the apparently widespread mental framework that roads don't actually do anything other than attract masses of people in cars, inexplicably like insects toward bright light.
- mratzloff 8y agoVersus what?
- taejo 8y agoPresumably if someone wants to go from A to B, or take goods from A to B, they derive some social or economic benefit from that (as do the people at B).
- CalRobert 8y agoI know you didn't mean it, but honestly your analogy is kind of brilliant. It does feel that way as somebody who enjoys walking.
- gascan 8y agoIt's true that roads convey people & permit economic activity. But that road down the way that is so congested right now? If we expanded it, and it could convey more cars at the same level of congestion- where did they come from? The latent demand represents potential traffic that either: 1) is doing just fine taking a different route right now 2) isn't particularly economically valuable, i.e. not worth waiting through traffic for 3) things like future sprawl enabled by wider roads today In order, #1 is not particularly valuable to reroute, it was doing ok. #2 is not very valuable either, though perhaps a little. #3 creates new demand, but the only thing we get in return is lower density. So I guess in sum the point is that all the most economically valuable travel is already happening, right now, and by widening the road we add only the travel that has such marginal economic value it's not worth doing today. So the more we widen the road, the less & less economic value we get in return.
- toast0 8y agoLane expansion isn't likely to eliminate rush hour, but it will increase the number of vehicles served, and can also reduce the duration of peak congestion. As vehicle populaton grows, congestion will grow as well, however. A recession is even better than construction at improving traffic, however.
- KayEss 8y agoHere in Bangkok the size of the (already legendary) traffic jams is certainly a great proxy for the economy as a whole.
- rplst8 8y agoYour point about recession is interesting. As we came out of the previous recession after the crash in 07/08, I noticed commute times increasing in the DC metro area. My argument is and has been that highway lane miles have never met demand. More lanes bring more cars simply because the other routes we're equally congested/slow before and now there's a slightly better option with the new mega highway.
- sliken 8y agoThe article has a lame explanation of latent demand. Sure some switch from surface streets to the highway. But people also decide to make more trips, commute further, or use public transportation less. So the result is after an adjustment period of about one year there's almost precisely as much traffic per lane as before.
- martyvis 8y agoI just got a new Subaru Outback with their Eyesight Adaptive Cruise Control. I've gotta say it is fantastic and just the sort of fairly simple and safe technology that could ease traffic jams. Certainly outside the city where I live, the way it nicely follows cars at a safe distance seems to make a lot of sense. As soon as the lead car moves away, it neatly accelerates to the preset speed and smoothly decelerates on the engine braking (or even real braking if needed) when it sees the gap closing. It's very obvious on my long commute that probably only 30% of cars seem to be using just regular cruise control, based on how their speed varies on the clear open road.
- stephengillie 8y agoAdaptive Cruise is life-changing. I couldn't go back to normal cruise. A major downside to Adaptive Cruise is when another driver uses your gap to change into your lane, and your car slams on its breaks because suddenly their car is just inches away from yours. Recklessly aggressive drivers are learning they can take abuse these safety features in the cars of other drivers, making roadways potentially more unsafe.
- Already__Taken 8y agoWhat are you saying, you want it to crash into them for you? That's a major downside?
- zyberzero 8y agoI'm not parent, but just a note from me: I drove a brand new Volvo XC40 last week that had adaptive cruise control. If someone just cut in to the front of me the car slammed on its brakes even if it wasn't necessary. The car could just as easily slow down gradually until necessary space was free in front of the car (as I usually do when I don't have adaptive cruise control).
- vinceguidry 8y agoTry Mazda, they're much better in that regard.
- alphabettsy 8y agoI find that on my commute most of the time when one lane is moving slow it’s because the vehicle at the front of a long line of cars, usually the left two lanes, in traveling 5-10 mph slower than other traffic. Speed differential and the use of brakes seems to cause traffic jams, so having vehicles all traveling the same speed or at least moving out of the way if they’d prefer not to go with the flow would seem to solve these issues.
- bryanlarsen 8y ago"A long line of cars ... traveling 5-10 mph slower" In other words you're saying that they're closely packed. If the distance between cars is reduced by 50% and their speed has reduced 10-15%, the throughput of that section of road has increased substantially. As you said, it's speed differential and the use of brakes that causes traffic jams. By that reasoning, the left lane should go the same speed as the right. That would also discourage the use of lane changes which is generally what causes people to use their brakes on the freeway... And remember the square law of braking. Because the safe following distance increases so much when you increase speed, faster speeds on the freeway decrease throughput rather than increase it.
- alphabettsy 8y agoNo. I’m saying that when faster traffic comes across these drivers they either change lanes immediately, if possible, which causes problems or they stay behind that driver, but still with many cars hunting for a faster moving lane behind them. I don’t find the slow moving line to be any closer together at all mostly because of people constantly changing through that lane. If the far right or middle lanes are traveling faster than the left and also moving smoothly then someone’s doing it wrong.
- Sohcahtoa82 8y ago> And remember the square law of braking. Because the safe following distance increases so much when you increase speed, faster speeds on the freeway decrease throughput rather than increase it. Got an article to explain this? It seems counter-intuitive. Following distances should be based on time, not physical distance. It shouldn't matter if cars are going 30 mph or 300 mph. If everyone maintains a 3 second following distance, there should still be a throughput of 20 cars per minute.
- blasdel 8y agoBill Beatty's comprehensive explanation from 20 years ago is still the best: http://trafficwaves.org/ http://trafficwaves.org/
- peterwwillis 8y agoAll he's doing here is obviating braking. In theory it could make for slightly reduced trip time as some stop-and-go's have more inefficiency than just moving slowly. He's reducing the pressure of the flow. But he's doing nothing about the actual causes of the slowdowns, which is on-ramps, off-ramps and rubbernecking. So while you reduce stop-and-go, you're still traveling at maximum capacity, just like in stop-and-go's. A kink in a hose doesn't go away if you reduce the water pressure.
- badrabbit 8y agoI wonder if toll payment based on amount of time spent in a lane would improve things? Right/exit lane would be free and incrementally increase up to the high cost left lane. The lane stripes would have connected rfid sensors to measure amount of timr spent in each lane. Might work because less lane switching means less slow downs to adjust to new cars entering a lane. And people who don't absolutely need to will remain on right lanes while leaving the left lane for those committed to the lane or for those using it momentarily to pass other cars. Heck,you can even change lane pricing dynamically based on traffic and time of day.
- sjwright 8y agoThis would almost certainly cause extreme congestion on the exit lanes and as a result a huge speed differential between lanes. Moving to the faster lanes would be dangerous. Cars merging back into the slower lanes to exit would cause permanent jams in the second lane at every turn-off. Nobody would benefit in the end and the road would be horribly unsafe.
- badrabbit 8y agoThat's why I suggested dynamic pricing and charging based on time spent. The right most lanes would congest but they would also have exiting cars. Second from right could cost a bit more than free (1cent per mile for example). It would be an incentive to use it for passing but also to avoid right lane congestion for a negligible fee. Third and fourth from right can cost $0.20 and $0.40 per mile which discourages long term use but also allows those in a hurry who can afford the fees to leave the highway faster and allow for cars in right lanes to use less congested left most lanes to pass slow drivers and lane exit/entrance in their lanes. Think of it from a network traffic CoS and QoS point of view (also,sorry for the late response,hope you see this)
- sjwright 8y agoAgain, a nice idea in theory but it will just concentrate congestion towards the right, which would exacerbate the speed differential between lanes and make the road much less safe. It would exacerbate economic disparities, forcing poorer people—who are already chronically time-poor as well as money-poor—to spend more time in their car for the convenience of rich people. Rich people already have innumerable advantages in our society, we don't need to go around looking for more to give them. (As a comparatively rich person, I appreciate living in a society surrounded by people that are—in the main—well educated, healthy and not overly stressed, and I happily pay a lot in taxes to help maintain that. Which is why the USA holds zero appeal to me; you couldn't pay me to live there.)
- dwd 8y agoOne solution is to restrict lane changes. All it takes to create a phantom traffic jam is someone to change lanes where the cars behind need to slow or brake to accommodate them.
- a3n 8y agoIf another freeway lane attracts drivers up from the surface streets, then that must be good for the surface streets. I assume the new drivers must have determined that that's a more efficient way to get from A to B.
- JTbane 8y agoBetter idea- expand the lanes but give them to a light rail system instead.
- bayfullofrays 8y agoLimiting the use and possession of private cars would go a long way into reducing environmental impact and income inequality.
- aeorgnoieang 8y agoI'm curious what the specific mechanisms are that you imagine would cause what you claim. First, how exactly are you going to limit the use and possession of private cars?
- magduf 8y agoThat'd be great, but no one wants to actually build any workable alternatives, at least in the US. A lot of people would probably be happy to give up their cars (or at least keep them parked until they need them for weekend trips) if they had good public transit available to use instead. But US cities are not doing a very good job with that.
- andrewla 8y agoThese problems only occur when the load get close to the effective maximum load. It's easy to look for blame in driver behavior, but I think misleading -- it's like taking that first crystal to form in 0 degree water, and blaming it for ice formation. You may be able to reduce crystallization (and thus sneak in a couple more Joules) by distilling the water (adaptive cruise control) or cleaning your glassware (on-ramp metering or "do not change lanes" signs), but really you're just fighting the statistics of phase changes. Note that I don't have a ton of supporting evidence for this sort of thing. One imaginary project that I would love to see if I had the time or the patience for deep research is to just gather a data set and make it public of anonymized traffic behavior -- over a large stretch of highway, have periodic cameras that track every car by license plate (if possible) or car shape, aggregate and remove the identifying data, and output a data set that allows you to slice it by time or by moving segment to directly observe these things.
- hedora 8y agoI was hoping for more detail. The argument against adding more lanes is that making the freeway faster than surface streets causes people to prefer the freeway to surface streets. Tying lane count back to fuel economy, average door to door time, and number of vehicles served would be much more productive. Similarly, the article recommends hov meters on freeway entrance ramps, but doesn’t mention that they create jams sbout a half mile after the entrance. Again, some sort of metric involving “how many people”, “how long”, and “how much gas” would go a long way. Fun fact about the Bay Area: my car gets about 75% of epa freeway mileage driving from SJ to Santa Cruz (over a mountain), but only 50% from San Jose to Mountain View (in traffic). This is if I try my best to conserve fuel in the traffic case. If I drive aggressively, the gap widens. So, reducing conjestion could substantially increase the miles per gallon of the existing silicon valley fleet of cars. Instead, the “environmentally friendly” nimbys keep voting in people that pay for traffic quiescence “upgrades” that further snarl traffic and increase the local carbon footprint.
- StanislavPetrov 8y agoHere in New York I would attribute traffic jams in equal parts to the endless construction, massive congestion, and a huge number of terrible and/or elderly drivers. All it takes is a couple of slow/bad drivers to clog up the expressway for miles, even if the congestion is low and there is no construction.