You ever see a dog that’s got its leash tangled the long way round a table leg, and it just cannot grasp what the problem is or how to fix it? It can see all the components laid out in front of it, but it’s never going to make the connection.
Obviously some dog breeds are smarter than others, ditto individual dogs - but you get the concept.
Is there an equivalent for humans? What ridiculously simple concept would have aliens facetentacling as they see us stumble around and utterly fail to reason about it?
Is this the root pathology behind traffic? Like, I never understood traffic, is there someone at the front refusing to go fast enough or is it the result of some distributed error like this that everyone mis-optimizes for that in aggregate results in traffic?
It may be helpful to think of it as a stream or a river, and not a collection of individual drivers. We can only control ourselves, not the stream. People working so hard to put themselves and others at risk are maybe shaving a minute or so off of their commute. Just not worth the risk.
Based on a game* I think that the root issue is that there are multiple bottlenecks, unavoidable for the drivers, like turning or entering/leaving lanes, forcing them to slow down to avoid crashing. Not a biggie if there are only a few cars, as they’ll be distant enough from each other to allow one to slow down a bit without the following needing to do the same; but once the road is close to the carrying capacity, that has a chain effect:
- A slows down because it’ll turn
- B is too close to A, so it slows down to avoid crashing with A
- C is too close to B, so it slows down to avoid crashing with B
- […]
There are solutions for that, such as building some structure to handle those bottlenecks, but they’re often spacious and space is precious in a city. Or alternatively you reduce the amount of cars by discouraging people from using them willy-nilly, with a good mass transport system and making cities not so shitty for pedestrians.
*The game in question is OpenTTD. This is easy to test with trains: create some big transport route with multiple trains per rail, then keep adding trains to that route, while watching the time that they take to go from the start to the end. The time will stay roughly constant up to a certain point (the carrying capacity), then each train makes all the others move slower.
Laughs in good public transit(rail based is based, but buses are good too), where it can achieve 10~100x the capacity in the same footprint
With rail, as long as you have a good timetable and a robust signaling system, 27tpdph with multiple service patterns is achievable, and >33tpdph if you run just one service pattern, all while having a top speed of 120km/h and an average speed of >50km/h
Railway in general (excluding Line-of-sight based light rail and trams) can move stupendous amounts of people at full speed really quickly due to signaling and mass transit inherently being more efficient in general
Most of the time and places a city doesn’t need that capacity. Since your rail cannot get the garbage from my house, or my new bed to the house, we need roads as well. Thus for most a bus running in mixed traffic (remember most roads do not have heavy traffic!) is good enough and a lot cheaper. Where you need capacity a train is really good, but you don’t need it.
That said I support trains in a lot more places because trains can run fully automated and thus in the real world can achieve the high frequency people need to choose transit even when a car isn’t a problem to own (they can afford it and there is no traffic). This is however just a stop gap since self driving buses don’t exist (yet?). In most “first world” countries cost of labor is high and automated trains are thus useful in places where a bus could do the job.
You forgot one solution.
Teaching people how to drive safe and smart. Way too many people focus on the car in front of them instead of the traffic ahead. If you watch for brake lights as far up as you can see and let off the gas when appropriate, not only will you be less likely to be in or cause a wreck, you will also save wear and tear on your brakes and use less gas (even more pronounced with regenerative braking).
In addition to the above. When you are driving a route you know well, get the fuck over from which ever side is more likely to be used to turn off. For most highways this means moving left before you near an onramp. Plan ahead and get over before you need to do so you don’t have to speed up or slow down to let people in.
Driving safely and smart is essential for other reasons, it does prevent additional bottlenecks (you mentioned one, wreckages), and it reduces the impact of the unavoidable bottlenecks (because the cars won’t waste so much time re-accelerating after them). But if my reasoning is correct, most of the time there isn’t much that drivers can do against traffic besides “don’t use the car”.
That won’t help much. By the time anyone notices the roads are slowing down there are six times as many cars on it as it can safely handle. Driving skills will help on backroads, but that isn’t where most people are driving. No amount of training can make heavy traffic safe.
Traffic is a numbers game. I’ve often observed that in free flowing traffic where I live (a tiny city with only about 700k people in the entire metro) that if you take two cars that are a safe following distance apart there will be 5 cars in between. If we put in 6 times as many lanes (already a 3-4 lane freeway each way, so we are talk 20 lanes for my tiny city!) traffic wouldn’t go any faster, but they would space out to most maintaining a safe following distance. (if you put in 7 times as many lanes they would get farther apart yet, but still not go faster)
There is research showing that adding lanes only helps for maybe six months. Then people realize that the route is better and change the routes they take, which leads to more congestion again. Fewer lanes can actually decrease congestion.
https://smv.org/learn/blog/how-does-roadway-expansion-cause-more-traffic/
That research is useless! Sure they measured it, so it isn’t wrong. However it is useless. What it is really saying is your city was so bad that people were not taking advantage of living in the city because they couldn’t conveniently get places. Those people could have lived in rural Montana for all the good a city did. Cities are about all the things you can do by living in it, so if people change because of new roads then you are a city were not meeting their ideals.
Also note that they measured one lane. I already asserted that by the time a city is thinking about adding one more lane they already need to add 6 times as many lanes (not 6 more lanes, 6 times!) IF your city needs 6 times more lanes than it has, no wonder people are choosing alternates, and once a lane exists they will start using it.
Again, the moral is build transit in cities.
If I’m understanding correctly, your example wouldn’t apply to a highway that is experiencing heavy congestion.
It would, but worse. Both are a case of more cars than there is space. Heavy congestion would just need a lot more lanes to fix - maybe 10x as many. (don’t ask me to pay for that or where those lanes go)
Or in short, support better transit for your city. For that cost of miles of 15 lane highways you can put in a lot of transit.
