19 ms·

Traffic Hacks: How One Driver Can Vastly Improve Traffic

237 points by Mintz 17y ago

masomenos 17y ago
While I have not conducted any format experiments in the efficacy of the advocated driving method, I can report that it's far more relaxing than tailgating. Give it a try!
Mintz 17y ago
I've actually done this on my way to/from work, and it makes a legitimate improvement in my commute.
zck 17y ago
Some Japanese scientists did a study to show how traffic jams can appear of of nowhere. They told 22 drivers to drive in a circular path at 30 km/h. A traffic jam appeared, and moved backward. The video is here: http://www.youtube.com/watch?v=Suugn-p5C1M http://www.youtube.com/watch?v=Suugn-p5C1M
- kyro 17y ago
  It's not 'out of nowhere.' It's because we're humans. We're paranoid and reactionary. You wouldn't see that with robots.
  - RyanMcGreal 17y ago
    Emergent network phenomenon is emergent.
  - randomwalker 17y ago
    Cynical much? How about it's because we have bounded cognitive abilities, limited perceptive accuracy and slow reaction times?
    - apotheon 17y ago
      That too.
    - dkarl 17y ago
      You're forgetting the natural caution and fearfulness that protect us from depending too much on our limited abilities when safety is at stake.
  - kragen 17y ago
    You can certainly program robots to do that. The question is whether you can program them not to without losing something important (robustness against unexpected conditions or malicious robots, for example).
  - req2 17y ago
    If I say I made a cake 'from scratch' (idiom: 'from nothing'), do you quibble that flour and eggs were indeed used, and that they are clearly not 'nothing'?
    - kyro 17y ago
      Mentioning how you made a cake from scratch and linking to a video of seemingly scientific research claiming an event to happen out of nowhere and for no apparent reason are two completely different things, due to the two completely different contexts.
      - tlrobinson 17y ago
        I think his point was that we all understand that "out of nowhere" doesn't mean "magically", at some level it has to be the drivers' fault. The video doesn't even say "out of nowhere", it says "for no apparent reason".
- cema 17y ago
  Jams move backward, a well known phenomenon. That they appear "out of nowhere" is more interesting.
  - steamer25 17y ago
    It is interesting but not inexplicable. For instance take a scenario where one driver sloooowly passes another on a two lane road. This effectively clotheslines the road and causes anyone who would drive faster yet to pile up behind the two cars--maybe even apply their brakes. Again, consider a sudden lane change that cuts off someone who's already in the lane being entered. Yet again, consider that most on-ramps are essentially a reduction of available lanes. Any time one car slows down for any reason, inertia kicks in--they do not immediately return to speed. In this case, small variations in speed mean that some drivers eventually catch up to others. Then they slow down a bit which reduces the space behind them. The problem is that the road is nearing peak capacity and has little tolerance for deviations.
tezza 17y ago
How One Driver Can Vastly Improve Traffic Yes, but that one driver is never the guy in front of you.
zngtk4 17y ago
Interesting, my instinct has always been to try to "punish" the cheaters by preventing them from merging in, thinking that this would discourage this behavior. But I can't really stop this, and the article provides a good explanation for why you might as well just let people merge in front of you.
- swolchok 17y ago
  The problem is that allowing merges incentivizes zooming ahead in the merge lane, which should result in there still being drivers who can't merge, creating the jam. I don't have a good solution for this.
  - zngtk4 17y ago
    To an extent I think you're right and some people will do it regardless. On the other hand, smoothly flowing traffic creates an incentive to not zoom ahead in the merge lane. There's less of a feeling of a need to cut ahead and people don't seem to do it as much when there isn't already a jam. (That is, the jam provides an incentive to cut in front of everyone) And if people were to implement this strategy, the provided space gives an incentive for those zooming ahead to merge a little sooner rather than risk waiting half a minute for someone to let them in right at the exit. It's been my experience that most of the "cheaters" will merge a little sooner if there's space rather than zoom up right to the end. They'll still be passing you, which is IMO aggravating, but at least things will move a little sooner.
- dminor 17y ago
  I used to as well, until one of the cheaters got out of his car and yelled at me at the next traffic light for "cutting him off". I decided that this potential outcome wasn't really worth it, especially since they always find a way in anyway. Another upshot is I also don't have to focus as hard on not rear-ending the person in front of me.
mrshoe 17y ago
This man is a saint. For years I've wished more people understood these principles. Instead, I get drivers who are angry at me for leaving a big space in front of my car. If only I pulled up right behind the car in front of me, they'd be 60 feet closer to their destination!! I wish more people would check this site out. Traffic is a real problem that wastes billions of taxpayer dollars and hours on the road every year. Another idea to make the roads better: lane-specific minimum speeds. Think about it. It would actually solve a ton of common traffic problems.
- antdaddy 17y ago
  Another problem is drivers who insist on driving in the left lane at all times and refuse to move over. IIRC, the left lane used to be referred to as the 'passing' lane...
  - axod 17y ago
    In the UK it still is, and passing on the left is illegal. I like that system more that all lanes equal.
    - rdtsc 17y ago
      So if there is a slow car there, what do you do? Does everyone wait behind them until they turn/exit or decide to switch lanes?
      - axod 17y ago
        If there's a slow car in the outside lane, and they're not making any attempt to overtake anything, you generally flash them until they move over. They are, after all, breaking the law.
        dgallagher 17y ago
        What if a car is in the passing lane doing the speed limit? Asking them to move out of the lane so another driver could pass by speeding, breaking the law, doesn't make much sense either.
imgabe 17y ago
I read this a while ago and I try to do it whenever I find myself in traffic. If nothing else, it is immensely less frustrating to drive a constant 15mph than to drive 30mph half the time and be at a complete stop half the time.
tjic 17y ago
When I accidentally end up on Rt 128 at rush hour, I do this. (I note that I was on a mailing list with a bunch of relatively smart people once, and I mentioned what I did - and even among the smart folks there was a lot of outrage because I was driving "wrong" or "stupidly". That reaction contributed to me leaving the list shortly thereafter. )
- cema 17y ago
  How happy I am to live (and work) at the portion of Rt 128 which is separate from I-95! But every time I go down to Boston, or far away to other urban areas, I can never keep the distance that I want because someone would certainly cut in. So I increase the distance again, to avoid tailgating, then another guy gets in between us, etc, etc. And so we have a peloton of cars moving at 65+ mph a few meters away from each other. Not too pretty.
  - apotheon 17y ago
    Nice use of "peloton".
- Mintz 17y ago
  Unfortunately (and unintuitively), a person's intelligence doesn't directly translate to the type of driver they are. That's a perfect example. Shame.
barrkel 17y ago
Ever since I can remember, I've always thought this was obvious, a product of random slowdown being amplified by reaction time delay. Motorway pileups are well known to be exacerbated by close driving, because each subsequent driver has to brake harder and harder to stop. Jams are just a less dramatic form of the same thing.
krschultz 17y ago
It saves a lot of gas too. I've always tried to get through stop and go traffic smoothly without using the brake, but I always did it for my own gas consumption. I never had this macro look at it.
- seanc 17y ago
  Especially if you drive a standard transmission. Staying off the brake makes a huge difference, and makes for less clutching as well. For bonus points, you can put the car in neutral while coasting, although some don't think that's safe.
  - eru 17y ago
    On modern cars with manual transmission, the motor is powered with some gas when it is not engaged. However when you leave it engaged an just don't push the pedal, no gas is needed. (Although the motor brakes somewhat. I don't know which effect is stronger.)
mildweed 17y ago
This works, but the trouble comes when some guy in a big truck starts heavily tailgating you and flipping you off.
- ynniv 17y ago
  That would be unfortunate, since a heavy truck has the most to gain by not changing speed.
kyro 17y ago
I'm a fast driver, and one thing I will never understand are those who will do anything and everything to box in a fast driver, or to prevent them from cutting into a lane and taking off. I see fast drivers as excited molecules, looking to escape. What often happens when a fast driver is being blocked is they'll switch lanes back and forth, looking for their escape, creating this wave of hesitation, causing this big low speed anchor to form. Someone driving fast won't suddenly stop the speeding after being blocked. Just let them go!
- ynniv 17y ago
  I am a fast driver, but in traffic I switch to wave abatement. It seems only logical to help the system when there is no way to personally benefit. (Except that moving to the right-most non-merging lane will help you avoid waves and move faster.) My prioritized rules of the road are: 1) go whatever speed you like 2) empty lane to your right? move right 3) no car in front of you and a car behind you? move right as soon as possible 4) slower car in front? move left if you can These simple rules cover most situations in a way that benefits everyone involved. They do not cover some minority cases: aggressive tailgaters and high relative lane speeds (I leave a buffer lane if > 10 mph and no cars behind me).
noodle 17y ago
i think its interesting that i used to drive like this site describes for the exact reasons it talks about. and now i don't/can't. the difference? i moved to atlanta. rush hour drivers here operate under a greedy algorithm. they see an opening that will allow them to speed up or gain extra distance immediately, and they'll take it, for no other reason than to be going faster right now, to be farther down the road right now. and that includes eating up my buffer zone, forcing me to break/stop, killing my flow optimization.
- zngtk4 17y ago
  That (in addition to the fact that everyone lives in the suburbs) perhaps provides some of the explanation for why Atlanta rush hour traffic seems to be some of the worst in the country, despite having more lanes and a smaller population than cities that have better traffic.
  - apotheon 17y ago
    I have had to commute on two of the top ten worst rush hour stretches of highway in the country (and they're within a few miles of each other in southern California, on the 91 and I-10). Noodle is exactly correct that such driving "cultures" do exist, and they make this consciously regularizing driving style worse than useless.
VBprogrammer 17y ago
While I find letting the 'cheaters' in grates too hard on my soul I alway try to practice driving at the average speed and avoiding using the brake. If for no other reason than to avoid wear on the car (in the UK most cars are still manual and so each start is a tiny bit more wear on the clutch as well as the brakes).
RyanMcGreal 17y ago
I changed my driving patterns after first reading this remarkable essay some years ago. Now I follow what I affectionately call the "zen of driving", to wit: leave a huge gap in front of me, drive at a constant speed where possible [1], avoid using the brakes except in case of emergency, and so on. On long driving trips through mixed traffic, I have made several observations, some obvious and some surprising: * Driving in congestion is no longer stressful. It just isn't. (Hence the "zen".) * Dramatically improved fuel economy. * It's possible to drive through even wildly erratic traffic (swinging from 100 km/h to 20 km/h) without ever touching the brakes. It becomes a kind of game. Manual transmission helps. * Some other drivers seem to figure out what I'm doing, and they fall into place behind me at a reasonable distance. I've had people follow me for hundreds of kilometres this way. * Even though you'd assume leaving a huge gap in front is an invitation for other motorists to cut ahead, they rarely do - especially in congestion. If they do, I just let them race up to the front of the gap and then fall back by a single car length to restore the same gap. I absolutely swear by this driving method. It's easier on the car, easier on the heart, and even serves to create a bubble of calm around you in an otherwise turbulent flow of traffic. ------ [1] Edit to qualify "drive at a constant speed". As I noted later, my speed ranges dramatically based on the congestion level - what remains more or less constant is the safe opening in front of me. Sorry for any confusion with my original choice of words.
- apotheon 17y ago
  That's not "zen" -- it's artificial regularity. Zen would be driving at what feels like a natural speed, only adjusting that when circumstances dictate, and "letting go" of your attachment to forward progress when something stands in your way. Instead, you're driving at an artificially limited speed to try to regularize your driving experience. You're tricking yourself into feeling less frustrated by consciously eliminating frustration triggers from the driving experience; a zen approach would involve simply choosing to feel less frustrated by "realizing" that the frustration triggers really don't matter. > Even though you'd assume leaving a huge gap in front is an invitation for other motorists to cut ahead, they rarely do - especially in congestion. That's heavily dependent on time and place. I've lived in many different regions, and done a lot of driving in those areas. I've noticed that the tendency to fill in gaps is heavily dependent on local driving culture and the specific demographics on the road at a particular time of day. > I absolutely swear by this driving method. It's easier on the car, easier on the heart, and even serves to create a bubble of calm around you in an otherwise turbulent flow of traffic. That may, to some extent, be an illusion. Some people are infuriated by other drivers they perceive as holding them up -- which often includes people who are "wasting" lane space by letting a gap of thirty car lengths lie unused in front of them. Personally, I only get annoyed with such people when they try to keep other people from filling that gap, because as far as I'm concerned a steady gap (given a steady general traffic speed) creates irregularities in grouping through which ambient traffic can accelerate itself, but people who maintain a huge gap then rush to block out people who want to move into it are increasing the danger of driving and doing nothing to help improve the flow of traffic.
iigs 17y ago
I respect this guy's patience, but the primary reason that one car can make such a big difference in this case is because 520 (one of the most important commuter roads in the Seattle area) is just two lanes in each direction. One car on a three lane road will just be passed. The counter-point to this, in my opinion, is that at some point psychology comes in to play and you have people starting to get aggressive and angry because of a few people forming a "rolling-road-block". It's not rational but it definitely happens. I would support 3PM-7PM dynamic speed limits on 520W toward the bridge. If the speed was brought from 60 (just past 405) to 30 as you approach the bridge it could probably keep things moving smoothly through the traffic burst.
erlanger 17y ago
If you want to help traffic, stop hitting your damn brakes so much. Brakes are for strong deceleration, and most deceleration in traffic can be handled by letting off the gas. Once you get good at this, follow as closely as possible in a traffic jam without applying your brakes.* There's nothing better a skilled driver can do to alleviate a traffic jam. The jam is physically shorter and people behind you are less likely to apply their own brakes as a result of these measures. * This has the added benefit of making for a smoother ride, which also encourages those behind you to ease off the brakes.
- BRadmin 17y ago
  Driving in Los Angeles, this is exactly what I do... I'm convinced that most traffic jams are caused by people needlessly applying their brakes and slow cars clogging the fast (far left) lane.
huherto 17y ago
I read this site a few years back and I've done a bit experimenting. I'm convinced that leaving space in front and avoiding sudden changes on speed are really helpful. Some times I have even seen traffic jams disapear if I leave a lot of space in front.
TravisLS 17y ago
This is a representation of classic game theory. If all participants cooperate to drive as described, the traffic clears. If nobody else helps to open up space, however, you will personally arrive at your destination slower by driving passively, even though many others behind you will benefit. And since this type of driving requires some degree of altruism, you can pretty much rule it out as a realistic outcome without introducing another constraint (eg. as mentioned in the article, state troopers merging into traffic and forming a rolling barricade).
- jcl 17y ago
  However, if everyone carried 10-15 extra car-lengths ahead of them, the total capacity of the road would be greatly diminished; you would just be moving the jam to the on-ramp. The peak performance of the system is actually a mix of altruism and selfishness.
  - masomenos 17y ago
    That's a fallacy, addressed by the site -- in short, the benefit gained by minimizing jams more than offsets the increased space between vehicles.
    - jcl 17y ago
      I think you may have misunderstood my point, which is that peak performance is attained when a few drivers drive altruistically (as the author does) and the rest drive "normally". If all drivers choose the same behavior (all altruistic or all normal), then the system doesn't work as well. This is interesting from a game-theoretic standpoint. The author says much the same thing on his FAQ page: If EVERY driver was to constantly maintain a HUGE space regardless of speed, then it would probably cause problems. The merge-zones might stop jamming, but the capacity of major highways would be reduced. On-ramps would become choked as traffic backed up into them, and there would be slowdowns extending far out into the countryside. It is, of course, a purely theoretical problem, since there is no shortage of non-altruistic drivers.
- devin 17y ago
apotheon 17y ago
My biggest complaint about this is that it's going to give many people who don't actually help traffic, and just drive slowly, feel self-righteous about their driving habits. Yes, in some circumstances this driving style can help traffic; no, it doesn't mean that driving 15MPH slower than ambient traffic is justified "for the greater good", especially if you're the type who jealously guards the thirty car lengths in front of you, accelerating to keep people from changing lanes in front of you.
simplegeek 17y ago
BTW, sorry if this sounds off-topic, do you know any good books/articles on Traffic? I'm very much fascinated by this too. Thanks.
- tptacek 17y ago
  Not even close to off-topic. Traffic problems (and queuing theory) have a long history in CS, and the related problems have applications all over the place.
  - simplegeek 17y ago
    Can you recommend any good books related to Traffic issues for a beginner and intermediate? Thanks.
- latortuga 17y ago
  a coworker of mine recommends Traffic, on this topic, by Tom Vanderbilt.
tptacek 17y ago
Doesn't this imply that cities could alleviate traffic pathologies by hiring people to drive carefully down the highway?
- jcl 17y ago
  He suggests on one of the later pages that cities should have state troopers drive in a constant-speed band across the lanes. (He also points out that group of citizens doing this independently in Toronto was arrested for it.)
- designtofly 17y ago
  This was the focus of a short video a group of college students made a few years ago, "A Meditation On the Speed Limit." http://video.google.com/videoplay?docid=-5366552067462745475 http://video.google.com/videoplay?docid=-5366552067462745475
  - kingnothing 17y ago
    Is that the video in Atlanta where they drive side by side around I-285 at the speed limit of 55 mph? It's an interesting social experiment, but I'm surprised none of them were driven off the road. Traffic on the left on 285 generally moves at 70 - 80 mph here.
jimbokun 17y ago
"WHY must a bottleneck develop at a merge zone? Well, obviously because there's too many cars on one road. And because everyone must take turns slowly merging together. WRONG! Wrong wrong wrong. Even during extremely low-traffic conditions, everyone still takes turns, yet everyone merges as a high speed flow,like a zipper. A bottleneck never appears." New York city drivers are really excellent at the zipper merge thing. While living there, I got really used to not having to slow down much at merge points. Everyone just drove at a decent speed using all available lanes then smoothly merged at the point where the lane was closed. Moving back to Pennsylvania, people start queuing up in one lane a mile before the merge point, and traffic grinds to a stand still. This is insane. Usually, though, if you do the sane thing and use the other lane up to the merge point, you can get in without trouble. Maybe I was one of the early mergers before I lived in New York, and have just blocked it out of my memory. Now, though, seeing it happen really grates on my nerves. One of many reasons I'm glad I hardly ever drive anymore. :)
steamer25 17y ago
Nice, patient guy but there's a better way yet if you can afford to gamble with your time a bit (don't try this if you can't afford to lose a little time). When confronted with a single backed up lane, use the uncontended lanes to zip to the front of the wave and merge in _at speed_. Maintaining speed is important. If you don't find an opening, you'll have to abort and find another route to your destination. The up-side is that this almost always pays off and you waste little time slowed by traffic or slowing others. The reason is because many people who queue up resign to the monotony and pull out their phone or their air drums, etc. They pay enough attention to brake consistently but not enough to accelerate as soon as is possible. The vast majority of the time, you can slip in front of one of these 'sleepers' before they notice the opening or have fully accelerated. Because of inertia, they will go through an acceleration period where you pull away from them leaving them a very sufficient gap. Another way to look at this is using the available real-estate to ascertain a proper zipper-merge. If done right, you wait less and no one gets stuck behind you. This is especially useful if you commute five days a week--the payoffs over time are well worth the occasional crap out.
- devin 17y ago
  I take issue with your theories. The "payoffs" you allude to do not exist. Your "zipper-merge" costs you gas money, and if you took the time to watch the video, you'd realize that you are considered a "cheater". Cheaters feel better about having gotten in front, but what they don't realize is that the "sleeper" is providing a way for cars to merge, thus alleviating the jam in the first place. All you're doing is typical busy-body traffic maneuvers that soothe your rattled frame, but do nothing to actually get you to your final destination any faster. The difference between maintaining adequate distance and doing your patented "zipper-merge" have been proven to shave at the most, a minute and a half off of your commute. Even if you do this five days a week, you're saving roughly 8 minutes. If you think that's worth it, then you should also be aware that the vast majority of accidents occur while changing lanes.
  - Dove 17y ago
    Actually, this study (http://ops.fhwa.dot.gov/wz/workshops/accessible/McCoy.htm http://ops.fhwa.dot.gov/wz/workshops/accessible/McCoy.htm) lends credence to the claim. Late merges have better throughput than early merges when encouraged by traffic signs and uniformly adopted. The study does not directly address the effect of "cheating" late merges which can create waves, but did note that traffic signs which strongly encouraged early merging by everyone increased overall transit time through construction. That suggests that the default case is better. That makes a little sense to me: a single lane can bear less traffic, so extending the amount of distance everyone has to spend in a single lane might result in higher latency overall, if not necessarily lower throughput.
boredguy8 17y ago
"Constant velocity" applies to lane changes, as well. I hate when people are going 70mph and slow down to the 55mph of the car in front of them, THEN change lanes. No, change lanes as you're coming up on them so that you maintain a constant velocity. Related to this is the need to "stay right". Especially in CA, some of my fastest travel happens in the far right lanes because people think of lanes as a "speed" rather than a "function", and so the right lanes are empty while the left lanes are crowded and slow. The number one lane is not the "fast" lane, it's a "passing" lane. Guess what? The number two lane is a passing lane, too. If you're not passing anyone, you should be on the far right side of the road.
chr15 17y ago
For anyone else fascinated by traffic patterns and driver behavior, I recommend reading Traffic: Why We Drive The Way we Do and What It Says about Us The author goes into fascinating detail about driver psychology and discusses similar ideas like those mentioned in the essay. Link to book: http://www.amazon.com/Traffic-Drive-What-Says-About/dp/0307277194/ref=sr_1_1?ie=UTF8&s=books&qid=1251250678&sr=8-1 http://www.amazon.com/Traffic-Drive-What-Says-About/dp/03072...
mattmcknight 17y ago
Makes me think of Drum-Buffer- Rope from Theory of Constraints. Maybe we need a pull system for roads. There are a few onramps on criminally undersized (due to nimby nutsos) highways in my area where a traffic light limits flow onto the road at busy times- smart. Then the same people put in stoplights at the ends of the offramps, backing traffic up onto the highway.
nalbyuites 17y ago
This essay should appear as a Public Service Broadcast through all media outlets.
c0nsilience 17y ago
I've been doing this type of driving for years in L.A. I noticed when I was a teenager that, no matter what the cause for stoppage, people always hit the brakes when they see brake lights ahead of them. So whenever drivers ahead of me hit the brakes, I decide instantly whether I also need to brake or whether I can just let off the gas. This way I don't cause drivers behind me to brake and cause that long ripple effect.
devin 17y ago
I've been driving like this ever since I heard a show on Wisconsin Public Radio where they interviewed a traffic researcher. One thing that is not true, however, is that "cheaters" are always wrong. For instance, there is significant research that suggests in a 2 lane situation where both lanes must merge into one single lane (road work ahead, for example), it makes more sense for both lanes to proceed uniformly toward the point where they must merge. Most people would assume this would cause a deadlock at the point where the merge occurs, but it doesn't. When both lanes are allowed to proceed without switching lanes towards the point where they both must merge, they will achieve far better results than if they were to organize into a single lane prior to that point. This makes people /really/ angry. Once, when there was approximately 2 miles notice for road work, I stuck with the right lane and did not merge into the left lane like everyone else. I actually had a large pickup truck attempt to cut me off from proceeding in an EMPTY lane. Of course, this only holds when there is an inevitable traffic jam. When it comes to /avoiding/ traffic jams, you are better off maintaining distance to allow other drivers to merge.
ThomPete 17y ago
One has to wonder whether same principle can be applied to economics. I.e. that continous and moderate spending instead of savings somehow keeps the wheels moving in the economy.
aoeu 17y ago
Kinda related and interesting/amusing - "This Guy Can Get 59 MPG in a Plain Old Accord. Beat That, Punk.": http://www.motherjones.com/politics/2007/01/guy-can-get-59-mpg-plain-old-accord-beat-punk http://www.motherjones.com/politics/2007/01/guy-can-get-59-m...
bvanevery 17y ago
I completely agree with the moving wave! I have pondered this very same pattern myself. I'm excited to see it documented and shared on the web. Taking this "zen" driving a step further and imagining an ideal world of cars that never break down on the highway and where the process of merging onto and exiting from a highway could be scientifically mastered such that the speed of the nearby cars is at most nominally affected, it is not inconceivable that cars could travel at speeds far beyond what is considered safe today.
poltergeist 17y ago
My friend and I have designed, in our 6th semesters, a 8085 microprocessor based, congestion-aware traffic control for crossroads. The system uses switches on the roads ( replaced using RFID or other sensors placed on the ground or at the sides of the roads ), that detect the load on each path and route traffic accordingly. Not much, but it worked like a charm !!

237 points by Mintz 17y ago

While I have not conducted any format experiments in the efficacy of the advocated driving method, I can report that it's far more relaxing than tailgating. Give it a try!

Mintz 17y ago

I've actually done this on my way to/from work, and it makes a legitimate improvement in my commute.

zck 17y ago

Some Japanese scientists did a study to show how traffic jams can appear of of nowhere. They told 22 drivers to drive in a circular path at 30 km/h. A traffic jam appeared, and moved backward. The video is here: http://www.youtube.com/watch?v=Suugn-p5C1M http://www.youtube.com/watch?v=Suugn-p5C1M

kyro 17y ago
It's not 'out of nowhere.' It's because we're humans. We're paranoid and reactionary. You wouldn't see that with robots.
- RyanMcGreal 17y ago
  Emergent network phenomenon is emergent.
- randomwalker 17y ago
  Cynical much? How about it's because we have bounded cognitive abilities, limited perceptive accuracy and slow reaction times?
  - apotheon 17y ago
    That too.
  - dkarl 17y ago
    You're forgetting the natural caution and fearfulness that protect us from depending too much on our limited abilities when safety is at stake.
- kragen 17y ago
  You can certainly program robots to do that. The question is whether you can program them not to without losing something important (robustness against unexpected conditions or malicious robots, for example).
- req2 17y ago
  If I say I made a cake 'from scratch' (idiom: 'from nothing'), do you quibble that flour and eggs were indeed used, and that they are clearly not 'nothing'?
  - kyro 17y ago
    Mentioning how you made a cake from scratch and linking to a video of seemingly scientific research claiming an event to happen out of nowhere and for no apparent reason are two completely different things, due to the two completely different contexts.
    - tlrobinson 17y ago
      I think his point was that we all understand that "out of nowhere" doesn't mean "magically", at some level it has to be the drivers' fault. The video doesn't even say "out of nowhere", it says "for no apparent reason".
cema 17y ago
Jams move backward, a well known phenomenon. That they appear "out of nowhere" is more interesting.
- steamer25 17y ago
  It is interesting but not inexplicable. For instance take a scenario where one driver sloooowly passes another on a two lane road. This effectively clotheslines the road and causes anyone who would drive faster yet to pile up behind the two cars--maybe even apply their brakes. Again, consider a sudden lane change that cuts off someone who's already in the lane being entered. Yet again, consider that most on-ramps are essentially a reduction of available lanes. Any time one car slows down for any reason, inertia kicks in--they do not immediately return to speed. In this case, small variations in speed mean that some drivers eventually catch up to others. Then they slow down a bit which reduces the space behind them. The problem is that the road is nearing peak capacity and has little tolerance for deviations.

tezza 17y ago

How One Driver Can Vastly Improve Traffic Yes, but that one driver is never the guy in front of you.

zngtk4 17y ago

Interesting, my instinct has always been to try to "punish" the cheaters by preventing them from merging in, thinking that this would discourage this behavior. But I can't really stop this, and the article provides a good explanation for why you might as well just let people merge in front of you.

swolchok 17y ago
The problem is that allowing merges incentivizes zooming ahead in the merge lane, which should result in there still being drivers who can't merge, creating the jam. I don't have a good solution for this.
- zngtk4 17y ago
  To an extent I think you're right and some people will do it regardless. On the other hand, smoothly flowing traffic creates an incentive to not zoom ahead in the merge lane. There's less of a feeling of a need to cut ahead and people don't seem to do it as much when there isn't already a jam. (That is, the jam provides an incentive to cut in front of everyone) And if people were to implement this strategy, the provided space gives an incentive for those zooming ahead to merge a little sooner rather than risk waiting half a minute for someone to let them in right at the exit. It's been my experience that most of the "cheaters" will merge a little sooner if there's space rather than zoom up right to the end. They'll still be passing you, which is IMO aggravating, but at least things will move a little sooner.
dminor 17y ago
I used to as well, until one of the cheaters got out of his car and yelled at me at the next traffic light for "cutting him off". I decided that this potential outcome wasn't really worth it, especially since they always find a way in anyway. Another upshot is I also don't have to focus as hard on not rear-ending the person in front of me.

mrshoe 17y ago

This man is a saint. For years I've wished more people understood these principles. Instead, I get drivers who are angry at me for leaving a big space in front of my car. If only I pulled up right behind the car in front of me, they'd be 60 feet closer to their destination!! I wish more people would check this site out. Traffic is a real problem that wastes billions of taxpayer dollars and hours on the road every year. Another idea to make the roads better: lane-specific minimum speeds. Think about it. It would actually solve a ton of common traffic problems.

antdaddy 17y ago
Another problem is drivers who insist on driving in the left lane at all times and refuse to move over. IIRC, the left lane used to be referred to as the 'passing' lane...
- axod 17y ago
  In the UK it still is, and passing on the left is illegal. I like that system more that all lanes equal.
  - rdtsc 17y ago
    So if there is a slow car there, what do you do? Does everyone wait behind them until they turn/exit or decide to switch lanes?
    - axod 17y ago
      If there's a slow car in the outside lane, and they're not making any attempt to overtake anything, you generally flash them until they move over. They are, after all, breaking the law.
      - dgallagher 17y ago
        What if a car is in the passing lane doing the speed limit? Asking them to move out of the lane so another driver could pass by speeding, breaking the law, doesn't make much sense either.

imgabe 17y ago

I read this a while ago and I try to do it whenever I find myself in traffic. If nothing else, it is immensely less frustrating to drive a constant 15mph than to drive 30mph half the time and be at a complete stop half the time.

tjic 17y ago

When I accidentally end up on Rt 128 at rush hour, I do this. (I note that I was on a mailing list with a bunch of relatively smart people once, and I mentioned what I did - and even among the smart folks there was a lot of outrage because I was driving "wrong" or "stupidly". That reaction contributed to me leaving the list shortly thereafter. )

cema 17y ago
How happy I am to live (and work) at the portion of Rt 128 which is separate from I-95! But every time I go down to Boston, or far away to other urban areas, I can never keep the distance that I want because someone would certainly cut in. So I increase the distance again, to avoid tailgating, then another guy gets in between us, etc, etc. And so we have a peloton of cars moving at 65+ mph a few meters away from each other. Not too pretty.
- apotheon 17y ago
  Nice use of "peloton".
Mintz 17y ago
Unfortunately (and unintuitively), a person's intelligence doesn't directly translate to the type of driver they are. That's a perfect example. Shame.

barrkel 17y ago

Ever since I can remember, I've always thought this was obvious, a product of random slowdown being amplified by reaction time delay. Motorway pileups are well known to be exacerbated by close driving, because each subsequent driver has to brake harder and harder to stop. Jams are just a less dramatic form of the same thing.

krschultz 17y ago

It saves a lot of gas too. I've always tried to get through stop and go traffic smoothly without using the brake, but I always did it for my own gas consumption. I never had this macro look at it.

seanc 17y ago
Especially if you drive a standard transmission. Staying off the brake makes a huge difference, and makes for less clutching as well. For bonus points, you can put the car in neutral while coasting, although some don't think that's safe.
- eru 17y ago
  On modern cars with manual transmission, the motor is powered with some gas when it is not engaged. However when you leave it engaged an just don't push the pedal, no gas is needed. (Although the motor brakes somewhat. I don't know which effect is stronger.)

mildweed 17y ago

This works, but the trouble comes when some guy in a big truck starts heavily tailgating you and flipping you off.

ynniv 17y ago
That would be unfortunate, since a heavy truck has the most to gain by not changing speed.

kyro 17y ago

I'm a fast driver, and one thing I will never understand are those who will do anything and everything to box in a fast driver, or to prevent them from cutting into a lane and taking off. I see fast drivers as excited molecules, looking to escape. What often happens when a fast driver is being blocked is they'll switch lanes back and forth, looking for their escape, creating this wave of hesitation, causing this big low speed anchor to form. Someone driving fast won't suddenly stop the speeding after being blocked. Just let them go!

ynniv 17y ago
I am a fast driver, but in traffic I switch to wave abatement. It seems only logical to help the system when there is no way to personally benefit. (Except that moving to the right-most non-merging lane will help you avoid waves and move faster.) My prioritized rules of the road are: 1) go whatever speed you like 2) empty lane to your right? move right 3) no car in front of you and a car behind you? move right as soon as possible 4) slower car in front? move left if you can These simple rules cover most situations in a way that benefits everyone involved. They do not cover some minority cases: aggressive tailgaters and high relative lane speeds (I leave a buffer lane if > 10 mph and no cars behind me).

noodle 17y ago

i think its interesting that i used to drive like this site describes for the exact reasons it talks about. and now i don't/can't. the difference? i moved to atlanta. rush hour drivers here operate under a greedy algorithm. they see an opening that will allow them to speed up or gain extra distance immediately, and they'll take it, for no other reason than to be going faster right now, to be farther down the road right now. and that includes eating up my buffer zone, forcing me to break/stop, killing my flow optimization.

zngtk4 17y ago
That (in addition to the fact that everyone lives in the suburbs) perhaps provides some of the explanation for why Atlanta rush hour traffic seems to be some of the worst in the country, despite having more lanes and a smaller population than cities that have better traffic.
- apotheon 17y ago
  I have had to commute on two of the top ten worst rush hour stretches of highway in the country (and they're within a few miles of each other in southern California, on the 91 and I-10). Noodle is exactly correct that such driving "cultures" do exist, and they make this consciously regularizing driving style worse than useless.

VBprogrammer 17y ago

While I find letting the 'cheaters' in grates too hard on my soul I alway try to practice driving at the average speed and avoiding using the brake. If for no other reason than to avoid wear on the car (in the UK most cars are still manual and so each start is a tiny bit more wear on the clutch as well as the brakes).

RyanMcGreal 17y ago

I changed my driving patterns after first reading this remarkable essay some years ago. Now I follow what I affectionately call the "zen of driving", to wit: leave a huge gap in front of me, drive at a constant speed where possible [1], avoid using the brakes except in case of emergency, and so on. On long driving trips through mixed traffic, I have made several observations, some obvious and some surprising: * Driving in congestion is no longer stressful. It just isn't. (Hence the "zen".) * Dramatically improved fuel economy. * It's possible to drive through even wildly erratic traffic (swinging from 100 km/h to 20 km/h) without ever touching the brakes. It becomes a kind of game. Manual transmission helps. * Some other drivers seem to figure out what I'm doing, and they fall into place behind me at a reasonable distance. I've had people follow me for hundreds of kilometres this way. * Even though you'd assume leaving a huge gap in front is an invitation for other motorists to cut ahead, they rarely do - especially in congestion. If they do, I just let them race up to the front of the gap and then fall back by a single car length to restore the same gap. I absolutely swear by this driving method. It's easier on the car, easier on the heart, and even serves to create a bubble of calm around you in an otherwise turbulent flow of traffic. ------ [1] Edit to qualify "drive at a constant speed". As I noted later, my speed ranges dramatically based on the congestion level - what remains more or less constant is the safe opening in front of me. Sorry for any confusion with my original choice of words.

apotheon 17y ago
That's not "zen" -- it's artificial regularity. Zen would be driving at what feels like a natural speed, only adjusting that when circumstances dictate, and "letting go" of your attachment to forward progress when something stands in your way. Instead, you're driving at an artificially limited speed to try to regularize your driving experience. You're tricking yourself into feeling less frustrated by consciously eliminating frustration triggers from the driving experience; a zen approach would involve simply choosing to feel less frustrated by "realizing" that the frustration triggers really don't matter. > Even though you'd assume leaving a huge gap in front is an invitation for other motorists to cut ahead, they rarely do - especially in congestion. That's heavily dependent on time and place. I've lived in many different regions, and done a lot of driving in those areas. I've noticed that the tendency to fill in gaps is heavily dependent on local driving culture and the specific demographics on the road at a particular time of day. > I absolutely swear by this driving method. It's easier on the car, easier on the heart, and even serves to create a bubble of calm around you in an otherwise turbulent flow of traffic. That may, to some extent, be an illusion. Some people are infuriated by other drivers they perceive as holding them up -- which often includes people who are "wasting" lane space by letting a gap of thirty car lengths lie unused in front of them. Personally, I only get annoyed with such people when they try to keep other people from filling that gap, because as far as I'm concerned a steady gap (given a steady general traffic speed) creates irregularities in grouping through which ambient traffic can accelerate itself, but people who maintain a huge gap then rush to block out people who want to move into it are increasing the danger of driving and doing nothing to help improve the flow of traffic.

iigs 17y ago

I respect this guy's patience, but the primary reason that one car can make such a big difference in this case is because 520 (one of the most important commuter roads in the Seattle area) is just two lanes in each direction. One car on a three lane road will just be passed. The counter-point to this, in my opinion, is that at some point psychology comes in to play and you have people starting to get aggressive and angry because of a few people forming a "rolling-road-block". It's not rational but it definitely happens. I would support 3PM-7PM dynamic speed limits on 520W toward the bridge. If the speed was brought from 60 (just past 405) to 30 as you approach the bridge it could probably keep things moving smoothly through the traffic burst.

erlanger 17y ago

If you want to help traffic, stop hitting your damn brakes so much. Brakes are for strong deceleration, and most deceleration in traffic can be handled by letting off the gas. Once you get good at this, follow as closely as possible in a traffic jam without applying your brakes.* There's nothing better a skilled driver can do to alleviate a traffic jam. The jam is physically shorter and people behind you are less likely to apply their own brakes as a result of these measures. * This has the added benefit of making for a smoother ride, which also encourages those behind you to ease off the brakes.

BRadmin 17y ago
Driving in Los Angeles, this is exactly what I do... I'm convinced that most traffic jams are caused by people needlessly applying their brakes and slow cars clogging the fast (far left) lane.

huherto 17y ago

I read this site a few years back and I've done a bit experimenting. I'm convinced that leaving space in front and avoiding sudden changes on speed are really helpful. Some times I have even seen traffic jams disapear if I leave a lot of space in front.

TravisLS 17y ago

This is a representation of classic game theory. If all participants cooperate to drive as described, the traffic clears. If nobody else helps to open up space, however, you will personally arrive at your destination slower by driving passively, even though many others behind you will benefit. And since this type of driving requires some degree of altruism, you can pretty much rule it out as a realistic outcome without introducing another constraint (eg. as mentioned in the article, state troopers merging into traffic and forming a rolling barricade).

jcl 17y ago
However, if everyone carried 10-15 extra car-lengths ahead of them, the total capacity of the road would be greatly diminished; you would just be moving the jam to the on-ramp. The peak performance of the system is actually a mix of altruism and selfishness.
- masomenos 17y ago
  That's a fallacy, addressed by the site -- in short, the benefit gained by minimizing jams more than offsets the increased space between vehicles.
  - jcl 17y ago
    I think you may have misunderstood my point, which is that peak performance is attained when a few drivers drive altruistically (as the author does) and the rest drive "normally". If all drivers choose the same behavior (all altruistic or all normal), then the system doesn't work as well. This is interesting from a game-theoretic standpoint. The author says much the same thing on his FAQ page: If EVERY driver was to constantly maintain a HUGE space regardless of speed, then it would probably cause problems. The merge-zones might stop jamming, but the capacity of major highways would be reduced. On-ramps would become choked as traffic backed up into them, and there would be slowdowns extending far out into the countryside. It is, of course, a purely theoretical problem, since there is no shortage of non-altruistic drivers.
devin 17y ago

apotheon 17y ago

My biggest complaint about this is that it's going to give many people who don't actually help traffic, and just drive slowly, feel self-righteous about their driving habits. Yes, in some circumstances this driving style can help traffic; no, it doesn't mean that driving 15MPH slower than ambient traffic is justified "for the greater good", especially if you're the type who jealously guards the thirty car lengths in front of you, accelerating to keep people from changing lanes in front of you.

simplegeek 17y ago

BTW, sorry if this sounds off-topic, do you know any good books/articles on Traffic? I'm very much fascinated by this too. Thanks.

tptacek 17y ago
Not even close to off-topic. Traffic problems (and queuing theory) have a long history in CS, and the related problems have applications all over the place.
- simplegeek 17y ago
  Can you recommend any good books related to Traffic issues for a beginner and intermediate? Thanks.
latortuga 17y ago
a coworker of mine recommends Traffic, on this topic, by Tom Vanderbilt.

tptacek 17y ago

Doesn't this imply that cities could alleviate traffic pathologies by hiring people to drive carefully down the highway?

jcl 17y ago
He suggests on one of the later pages that cities should have state troopers drive in a constant-speed band across the lanes. (He also points out that group of citizens doing this independently in Toronto was arrested for it.)
designtofly 17y ago
This was the focus of a short video a group of college students made a few years ago, "A Meditation On the Speed Limit." http://video.google.com/videoplay?docid=-5366552067462745475 http://video.google.com/videoplay?docid=-5366552067462745475
- kingnothing 17y ago
  Is that the video in Atlanta where they drive side by side around I-285 at the speed limit of 55 mph? It's an interesting social experiment, but I'm surprised none of them were driven off the road. Traffic on the left on 285 generally moves at 70 - 80 mph here.

jimbokun 17y ago

"WHY must a bottleneck develop at a merge zone? Well, obviously because there's too many cars on one road. And because everyone must take turns slowly merging together. WRONG! Wrong wrong wrong. Even during extremely low-traffic conditions, everyone still takes turns, yet everyone merges as a high speed flow,like a zipper. A bottleneck never appears." New York city drivers are really excellent at the zipper merge thing. While living there, I got really used to not having to slow down much at merge points. Everyone just drove at a decent speed using all available lanes then smoothly merged at the point where the lane was closed. Moving back to Pennsylvania, people start queuing up in one lane a mile before the merge point, and traffic grinds to a stand still. This is insane. Usually, though, if you do the sane thing and use the other lane up to the merge point, you can get in without trouble. Maybe I was one of the early mergers before I lived in New York, and have just blocked it out of my memory. Now, though, seeing it happen really grates on my nerves. One of many reasons I'm glad I hardly ever drive anymore. :)

steamer25 17y ago

Nice, patient guy but there's a better way yet if you can afford to gamble with your time a bit (don't try this if you can't afford to lose a little time). When confronted with a single backed up lane, use the uncontended lanes to zip to the front of the wave and merge in _at speed_. Maintaining speed is important. If you don't find an opening, you'll have to abort and find another route to your destination. The up-side is that this almost always pays off and you waste little time slowed by traffic or slowing others. The reason is because many people who queue up resign to the monotony and pull out their phone or their air drums, etc. They pay enough attention to brake consistently but not enough to accelerate as soon as is possible. The vast majority of the time, you can slip in front of one of these 'sleepers' before they notice the opening or have fully accelerated. Because of inertia, they will go through an acceleration period where you pull away from them leaving them a very sufficient gap. Another way to look at this is using the available real-estate to ascertain a proper zipper-merge. If done right, you wait less and no one gets stuck behind you. This is especially useful if you commute five days a week--the payoffs over time are well worth the occasional crap out.

devin 17y ago
I take issue with your theories. The "payoffs" you allude to do not exist. Your "zipper-merge" costs you gas money, and if you took the time to watch the video, you'd realize that you are considered a "cheater". Cheaters feel better about having gotten in front, but what they don't realize is that the "sleeper" is providing a way for cars to merge, thus alleviating the jam in the first place. All you're doing is typical busy-body traffic maneuvers that soothe your rattled frame, but do nothing to actually get you to your final destination any faster. The difference between maintaining adequate distance and doing your patented "zipper-merge" have been proven to shave at the most, a minute and a half off of your commute. Even if you do this five days a week, you're saving roughly 8 minutes. If you think that's worth it, then you should also be aware that the vast majority of accidents occur while changing lanes.
- Dove 17y ago
  Actually, this study (http://ops.fhwa.dot.gov/wz/workshops/accessible/McCoy.htm http://ops.fhwa.dot.gov/wz/workshops/accessible/McCoy.htm) lends credence to the claim. Late merges have better throughput than early merges when encouraged by traffic signs and uniformly adopted. The study does not directly address the effect of "cheating" late merges which can create waves, but did note that traffic signs which strongly encouraged early merging by everyone increased overall transit time through construction. That suggests that the default case is better. That makes a little sense to me: a single lane can bear less traffic, so extending the amount of distance everyone has to spend in a single lane might result in higher latency overall, if not necessarily lower throughput.

boredguy8 17y ago

"Constant velocity" applies to lane changes, as well. I hate when people are going 70mph and slow down to the 55mph of the car in front of them, THEN change lanes. No, change lanes as you're coming up on them so that you maintain a constant velocity. Related to this is the need to "stay right". Especially in CA, some of my fastest travel happens in the far right lanes because people think of lanes as a "speed" rather than a "function", and so the right lanes are empty while the left lanes are crowded and slow. The number one lane is not the "fast" lane, it's a "passing" lane. Guess what? The number two lane is a passing lane, too. If you're not passing anyone, you should be on the far right side of the road.

chr15 17y ago

For anyone else fascinated by traffic patterns and driver behavior, I recommend reading Traffic: Why We Drive The Way we Do and What It Says about Us The author goes into fascinating detail about driver psychology and discusses similar ideas like those mentioned in the essay. Link to book: http://www.amazon.com/Traffic-Drive-What-Says-About/dp/0307277194/ref=sr_1_1?ie=UTF8&s=books&qid=1251250678&sr=8-1 http://www.amazon.com/Traffic-Drive-What-Says-About/dp/03072...

mattmcknight 17y ago

Makes me think of Drum-Buffer- Rope from Theory of Constraints. Maybe we need a pull system for roads. There are a few onramps on criminally undersized (due to nimby nutsos) highways in my area where a traffic light limits flow onto the road at busy times- smart. Then the same people put in stoplights at the ends of the offramps, backing traffic up onto the highway.

nalbyuites 17y ago

This essay should appear as a Public Service Broadcast through all media outlets.

c0nsilience 17y ago

I've been doing this type of driving for years in L.A. I noticed when I was a teenager that, no matter what the cause for stoppage, people always hit the brakes when they see brake lights ahead of them. So whenever drivers ahead of me hit the brakes, I decide instantly whether I also need to brake or whether I can just let off the gas. This way I don't cause drivers behind me to brake and cause that long ripple effect.

devin 17y ago

I've been driving like this ever since I heard a show on Wisconsin Public Radio where they interviewed a traffic researcher. One thing that is not true, however, is that "cheaters" are always wrong. For instance, there is significant research that suggests in a 2 lane situation where both lanes must merge into one single lane (road work ahead, for example), it makes more sense for both lanes to proceed uniformly toward the point where they must merge. Most people would assume this would cause a deadlock at the point where the merge occurs, but it doesn't. When both lanes are allowed to proceed without switching lanes towards the point where they both must merge, they will achieve far better results than if they were to organize into a single lane prior to that point. This makes people /really/ angry. Once, when there was approximately 2 miles notice for road work, I stuck with the right lane and did not merge into the left lane like everyone else. I actually had a large pickup truck attempt to cut me off from proceeding in an EMPTY lane. Of course, this only holds when there is an inevitable traffic jam. When it comes to /avoiding/ traffic jams, you are better off maintaining distance to allow other drivers to merge.

ThomPete 17y ago

One has to wonder whether same principle can be applied to economics. I.e. that continous and moderate spending instead of savings somehow keeps the wheels moving in the economy.

aoeu 17y ago

Kinda related and interesting/amusing - "This Guy Can Get 59 MPG in a Plain Old Accord. Beat That, Punk.": http://www.motherjones.com/politics/2007/01/guy-can-get-59-mpg-plain-old-accord-beat-punk http://www.motherjones.com/politics/2007/01/guy-can-get-59-m...

bvanevery 17y ago

I completely agree with the moving wave! I have pondered this very same pattern myself. I'm excited to see it documented and shared on the web. Taking this "zen" driving a step further and imagining an ideal world of cars that never break down on the highway and where the process of merging onto and exiting from a highway could be scientifically mastered such that the speed of the nearby cars is at most nominally affected, it is not inconceivable that cars could travel at speeds far beyond what is considered safe today.

poltergeist 17y ago

My friend and I have designed, in our 6th semesters, a 8085 microprocessor based, congestion-aware traffic control for crossroads. The system uses switches on the roads ( replaced using RFID or other sensors placed on the ground or at the sides of the roads ), that detect the load on each path and route traffic accordingly. Not much, but it worked like a charm !!