[Jun8]

There's a common pattern in population dynamics that I've observed, which probably has a formal name: A given system with (to simplify) two types of behavior: good and bad will tolerate and self correct up to certain percentage of bad actors (call this the jerk threshold, JT) after which all participants switch to the bad behavior. Examples are line formation for a service, cutting people off in traffic, stealing from common areas, hoarding office supplies, etc.

Anecdata: When I travel back to the country I've grown up I resort back to being a jerk in traffic and cutting in lines because you have to.

I've always wondered what the JT for different situations is, before the system breaks down, e.g. what percentage of people in a line have to cut in before everybody abandons the concept of forming the line.

[bombcar]

They've actually had to try to make people more of a jerk in the midwest, and instruct them on why the "zipper merge" is NOT bad: https://www.dot.state.mn.us/zippermerge/

Otherwise if there's a sign that says "Left Lane Closed 10 Miles Ahead" everyone will get in the right lane for ten miles.

[sidewndr46]

It is plausibly a more efficient usage if both lanes of traffic are moving.

In practice drivers treat the "zipper merge" as acceptable when the left lane is closed ahead, there is 500 ft of road remaining and traffic in the right lane is already dead stopped. It isn't uncommon to see people approaching at 1.5x - 2x the posted speed limit and trying to "merge" then.

[Loudergood]

That is the correct behavior. If only one lane is stopped its far more likely to stretch back and block an exit that does not need to be blocked.

[NoMoreNicksLeft]

> everyone will get in the right lane for ten miles.

And have 10 miles to get up to speed. As opposed to the zipper merge where everyone waits til the last minute, and everyone has to come to a complete stop.

Supposing of course your observations are true. From what I can tell with the "lane closed in 10 miles", is some asshole invariably sees it as a chance to get ahead of everyone else, zooms through the left lane, almost causes an accident, and then other idiots who think he's being successful in zooming ahead imitate his behavior. You get your "zipper merge" anyway, where everyone has to come to a dead stop at the chokepoint anyway. We're not all one gigantic robotic hivemind, preparing early makes for a smoother experience for all and fewer hostile feelings.

For those who use the shoulder on the interstate who do not have genuine medical emergencies (why didn't you call the ambulance?) I propose that they bring in mobile car crushers. I'm even generous enough to let them exit the vehicle first.

[master-lincoln]

In my experience there is no full stop if zipper merge is done right, but lots of angry people and a an inefficient use of the road leading to traffic jam miles before the merge needs to happen if people try to get into the one lane early.

If people all zip merge, there is no zooming possible on the closed lane and both lanes that merge get down to around 20mph at the choke point.

This only works in countries with civilized traffic participants as mentioned by OP. I have personally only seen 2 countries out of maybe 30 where that was the case. In cars many people behave like animals for some reason

[CrispyKerosene]

Even odder when it comes to behavior, in my experience people in my current city routinely fail at navigating a zipper merge without significantly reducing their speed, even coming to a full stop.

But when it comes to freeway on ramps, even in busy interchanges, there s usually no issues, which is odd because they are more or less the same maneuver.

There must be some psychological factors at play.

[practicemaths]

So that sounds like a massive flaw in traffic management design.

It's more efficient to develop traffic rules that work in how people are not how they should ideally be.

[CrispyKerosene]

Brawndo! it's what plants crave.

Where do you draw the line though when catering to the uniformed? There is a reason we hire SME's.

[practicemaths]

You design systems for the way nature works not how you wish nature worked ideally.

[CrispyKerosene]

I understand that my argument will lead to the inevitable 'well if everyone one followed all the rules all the time' type response, but there if people are leaving the appropriate following distance with perhaps a bit extra at the as they approach the zipper merge, than there is no reason the merge cannot take place at near highway speeds, certainly with no need to stop.

Its no different than how a freeway entry works.

[NoMoreNicksLeft]

I can't prove it, but I suspect (strongly) that whichever method you believe to be the correct one, following distances must shorten dramatically, or speeds must drop so low as to threaten traffic jams (which, perversely, leave everyone bumper to bumper). It's all about how many cars are trying to use that stretch of road for a period of time, and if all are to fit through within that window, there is both a minimum speed and a maximum following distance. Wish I wasn't a math flunky.

[CrispyKerosene]

I think most people are just unaware of how much distance is required between to cars at highway speeds to considered safe. A proper buffer would leave lots of room to absorb the extra cars and loss of speed. That the point, its flexible and elastic.

[pixl97]

> if people are leaving the appropriate following distance

The general rule is, no, they are not. Unfortunately there is really no penalty in the US if you're following to close unless you get in an accident. I quite often drive on I-35 and so often it's in conditions I call a fast moving traffic jam. People are traveling nearly 80MPH with 2 car distances between each other. Quite often this turns into multi car pileups because someone has to emergency break.

[red_admiral]

Arguably, unless traffic is free-flowing at maximum speed, doing the zipper merge is in fact the non-jerk option.

[myko]

ugh tell that to the idiots who give me the finger for zipper merging in Columbus, OH

[ToValueFunfetti]

Probably should prioritize telling them that red lights mean stop and that the arrow on the one-way sign points in the only direction you're allowed to go. Maybe mention turn signals too.

Allegedly there are three cities with worse drivers, but I find that highly implausible

[master-lincoln]

What's the legal situation? Where I live in the EU merging lanes demands a zipper merge according to law.

[aidenn0]

Varies from state to state since traffic laws are per-state in the US, but in most states the traffic merging is supposed to yield to the traffic staying in their lane.

You also are supposed to maintain a "safe distance" (where the law might be literally "safe distance" or might give a specific units) which at freeway speeds would mean leaving enough room for zippers but I don't know anyone who has ever been cited for following too closely, except after an accident caused by following too closely.

[myko]

There are signs posted explaining that zipper is the correct way to go, but I still get run off the road for doing it when merging onto the outer belt in this city

[practicemaths]

Well if you're forcing them to break when you merge then that kind of requires a finger imo.

Regardless Columbus has a great lack of turn signals and understanding of basic intersection rules.

[jstanley]

You're not forced to brake if you can see that 2 lanes go down to 1 and you therefore anticipate that the cars in the other lane are going to have to merge into your lane and you leave a gap to allow them to do so.

[practicemaths]

If there is an appropriate gap it is fine. However in my experience many people in the zipper lane merge without regard to what the traffic in the lane they're merging into is doing. Thus forcing people to slam on breaks and slows down the overall flow of traffic.

[myko]

> if you're forcing them to break when you merge then that kind of requires a finger

i do my best to avoid forcing anyone to brake, ever

usually the people flipping me off are speeding ahead of me from the other lane as soon as they see my signal go on

[practicemaths]

Oh, so you end up with space to merge easily then if they're speeding up.

[incanus77]

Contrast that with Oregon, where everyone will stay in both lanes until one car length short of ten miles.

[sircastor]

This is a pet peeve of mine, but also isn't it the most practical? Don't we want the lanes consumed as much as possible so traffic disruptions aren't distributed further back down the road?

[SirSourdough]

Yes and no. If consuming the lanes as much as possible was the goal, we’d be parked bumper to bumper with no available space to move at all.

We want the throughput of the road to be as high as possible. Broadly, that means maintaining the optimal (minimum) spacing to make the merge safely without changing speed, while maximizing speed.

You may get larger areas where speeds are disrupted, but on the average be going faster. Stop-and-go traffic introduces tons of inefficiencies since drivers need to be much more cautious than in a consistent traffic flow.

The big challenge is that it’s basically impossible to know what the optimal speed is, since it relies on knowing the detailed state of the road and traffic volume miles ahead of what you can see.

Self-driving cars are interesting here since they will open up new opportunities for all the vehicles on the road to co-regulate their speed near to the optimum.

[incanus77]

Growing up learning to drive on the East Coast, my intuition finds a sweet spot between the two, typically merging while moving slowly when there is a natural gap between cars.

What I tend to see instead is hard braking ripple back down the line to a half-dozen cars or more for each car trying to merge right at the end because suddenly there are two cars right next to each other and one lane of car width left. Every time it appears that the lane-ending driver is like "oh wow, my lane went away!"

[DANmode]

Why look around and anticipate the movements of others when you can finish your tweet and jab the brakes on everyone's day behind you?

Duh.

[bena]

Basically, you want both lanes to effectively spread out to where those in one could be placed in the other without overlapping vehicles.

The goal is to have both lanes move slowly, but constantly. To basically have a "1 and 1" progression through the chokepoint: 1 from the left, 1 from the right.

It is a rough balance because you have those who don't want to let others merge at all. And there are those who want to merge at all costs. Both of which violate the "1 and 1" guideline which gives everyone at least the illusion of progress.

[yurodivuie]

For years, I thought early merging was the morally correct choice, until a friend in my car yelled at me for doing it. He was a very conscientious person, but he was also a bus driver in the city (and thus much better trained in driving through congested areas). I think it's just a matter of education.

Another way to think of it: if you merge early, then the actual correct time to merge becomes indeterminate. Do you merge when you see the sign? Wait till you see a good gap? What if the person behind you doesn't have a gap, and they drive right past to keep looking? It becomes chaotic, and everyone thinks they are getting picked on when someone decides to merge in front of them or passes them. So much wasted anxiety and anger. It's a lot easier (in congestion) to wait until you need to merge, then merge.

In free-flowing traffic, it's a bit different, but the Minnesota page on zipper merging acknowledges that at the end of the article.

[JohnFen]

> I thought early merging was the morally correct choice

I never thought of it as a moral thing, but I merge as early as possible (ideally, I set myself up in the lane I need for the entire trip and avoid merging at all) simply because it's the safest thing to do.

When I don't do this, I will either miss the exit I want, or I'll get stuck at a standstill waiting for the rare kind soul in the lane next to me to let me in.

[radlad]

> Otherwise if there's a sign that says "Left Lane Closed 10 Miles Ahead" everyone will get in the right lane for ten miles.

One reason for this is that it only takes one person "policing" from the right lane (i.e. driving down the middle of the road, or worse: swerving out in front of the left lane) to shut down zipper merging.

[sircastor]

The thing that really bothers me about this is that the Jerk Threshold pushes the Overton window on Jerkiness - now if you're not participating in Jerk behavior you're disrupting "The way things work" which on its own might be jerk behavior.

I grew up and live in Oregon. I've generally thought of our drivers as relatively non-aggressive. But I've seen Californians† who are aggressively merging and weaving (And if you've ever driven down in LA, you know the lay of the land). Then the Oregonians who follow suit, and then everyone is doing it.

†Oregonians have been complaining about Californians since time immemorial. It's just pure tribalism. We blame any negative change on our state as "Californians moving in". My apologies to Californians for unfairly dished blame.

[aidenn0]

The population of California drivers seems to have this weird bimodality of weaving lanes or never changing lanes at all; the other annoying "California Driver" thing is "If you are going slower than I want, how dare you believe that I might move left one lane to pass you. I'll only do that after tailgating you for a few minutes and then flipping you off"

Even when going 65 in the right lane I've seen this happen.

The other thing that seems more specific to SoCal that pisses me off is: "35MPH is a perfectly reasonable speed to merge onto the freeway at. Not just on short ramps, but on ramps that are 1/4 a mile downhill where even a Geo Metro could hit 55. Yes I'm going to go 80MPH eventually, but not until a half mile after merging."

[Lanzaa]

Speaking of population dynamics, I enjoyed the "Parable of the Polygons"[1]. It shows with simple simulations an example of individual biases being different from collective biases. This difference leads to different collective behavior depending on environmental conditions. Similar to your example of individuals having a certain "jerk"ness, but the collective can pass a "jerk" threshold, which leads to an increase in jerk behavior.

[1]: https://ncase.me/polygons/

[olah_1]

> after which all participants switch to the bad behavior

This is how it has become with job applications. So many people started lying on resumes that the job reqs starting raising the requirements for a position, which causes more people to need to lie. If you don't lie, you just don't get a job and starve/die.

[rachofsunshine]

The Jerk Threshold is particularly low in low-trust, low-verifiability, usually-not-iterated games, and there's almost no setting that fits those conditions more than a job search.

[olah_1]

Good point. I think the difference is that the society as a whole was much higher-trust than it is today. America is back to being the Wild West, take what you can and don’t look back.

[red_admiral]

The opposite can happen too. Scott Alexander's old post "In favour of niceness, community and civilisation" quotes an Atlantic article [1] for an example of a model where all of a country's officials can flip from mostly-corrupt to honest at short notice, once corruption falls below a threshold. A lot of other threshold situations like the one you described feature in that article too - including, sadly, a "Rwanda" model for how when inter-group animosity crosses a threshold you can end up with genocide.

As far as I can tell, the formal name is "population dynamics".

[1] https://archive.is/BXq66

[gojomo]

See also: 'broken windows theory': https://en.wikipedia.org/wiki/Broken_windows_theory

Though this concept isn't as pat and reliable as in its most-simplistic formulations – you can't fix all crime with aesthetic enforcement – it captures real human tendencies to 'flock' in the space of norms-of-behavior, using visible cues of what will or won't be tolerated. And, it has applicability outside of just literal 'policing'.

Keeping spaces/communities far from any chaotic boundary where people start to wonder - "what can I get away with? does anyone confront/correct problems?" - can save a lot on overall defection/enforcement losses in the long run.

[aidenn0]

Yeah, I think Broken Windows Theory takes it a step too far. I believe that if you have an area without graffiti and broken windows, fewer people will graffiti on walls and break windows, but the evidence for that preventing violent crime is flimsy at best.

[michaelcampbell]

It feels like this is happening in US politics, and has been for some time. I'm in the US so I can only speak to it, but perhaps the world over?

[m463]

How you drive displays your true personality.

[rachofsunshine]

You can think of the Jerk Threshold problem as one of a prisoner's dilemma with a (sufficiently large) extra term, which you can think of as some combination of personal guilt, societal punishment, and whatever positive prosocial instincts humans have.

In the classic prisoner's dilemma, your payoff is +3 if you defect and the other prisoner doesn't, +2 if neither of you defects, +1 if both of you defect, and 0 if you don't defect and they do. There's no Jerk Threshold in this problem - regardless of your opponent's behavior, in isolation, you are always better off defecting.

But let's add the extra Antijerk Term - call it A - that you pay to defect. This could be you feeling bad about defecting, or it could represent the chance that you face retaliation for defecting later, whatever - there's some slight cost to being a jerk. We can see how the Jerk Threshold changes.

Your payoff matrix is now [[2, 3-A], [0, 1-A]]. If your opponent has a probability p of defecting, you can compute:

E[cooperate] = p * 0 + (1-p) * 2 = 2 - 2p, with the former term corresponding to you getting screwed and the latter corresponding to cooperation.

E[defect] = p * (1-A) + (1-p) * (3-A) = 3 - 2p - A, with the former term corresponding to the defect-defect state and the latter one being you screwing them.

For A < 1, 2 - 2p < 3 - 2p - A, so you're still in a prisoner's dilemma. But for A > 1, the problem abruptly shifts into a cooperation game, because 2 - 2p > 3 - 2p - A for A > 1.

-----

In this case, the Jerk Threshold abruptly shifts from 0 to 1 (that is, nothing could make you cooperate -> nothing could make you defect). But in the real world, the A term varies depending on context. The A term with a friend is very high, because you have a lot of opportunity for retaliation and they'll feel especially bad screwing you over. The A term with a stranger you're somewhat hostile to is low (maybe even negative). And A varies from actor to actor - some of us have stronger consciences than others. So you end up with bubbles where there's a stable local equilibrium (because the A values are high internally and that maintains stable cooperate-cooperate equilibria) despite differences with the outside world.

Properly modeling this probably looks something like the Ising model [1] on some complicated social graph. Which explains why we see these kinds of phase transitions - most real graphs are dense enough to have them. The bubbles we just described correspond to magnetic domains, and the incentives not to cooperate while in contact with a defect-bubble (or vice-versa) correspond to the high potential energy of domain walls.

[1] https://en.wikipedia.org/wiki/Ising_model