Hacker News new | comments | ask | show | jobs | submit login
‘The Trains Are Slower Because They Slowed the Trains Down’ (villagevoice.com)
376 points by jonathanjaeger 10 months ago | hide | past | web | favorite | 237 comments

This article has a lot of text before getting to describing the problem.

Summary: Train wreck 1995 and a installation of a safety-first, CYA response that slows down the trains and does not increase safety.

Longer Summary: Train wreck in 1995. Maybe driver was asleep. Start installing new signals that not only automatically stop the trains when there is a train ahead (good idea and already in place), but also when going over 45mph. Trains used to go up to 55mph with no systemic problems. Drivers go even slower than 45mph because the train is automatically stopped when going faster (as you would want to do if there was a train ahead) instead of just giving a warning and letting the train keep going.

As more of the system gets these new signals, more of the system has slower trains. Thus the increase of delays that are labeled generically as "insufficient capacity, excess dwell, unknown" from about 20% of delays in 2012 to about 60% in 2017.

>Trains used to go up to 55mph with no systemic problems.

During the age of steam, express passenger trains regularly exceeded 55 mph. For example, during the 1930s the Burlington Zephyr regularly exceeded 85 mph. For example it made the Chicago to Denver run in 1934 in 13 hours rather than the 18 hrs it takes today.

Source: http://www.slate.com/articles/life/transport/2009/05/stop_th...

Realistically speaking, you are talking about two different use cases here.

A long distance train benefits from very high top speeds, because it spends a lot of time at those top speeds. Conversely, increasing the high speed of a subway train is subject to the law of diminishing returns, since the train has to comfortably accelerate and decelerate within the span of half a mile. With subways, the main concern is that every single train operates in pretty much the exact same manner WRT acceleration, deceleration, and speed, since the headways on subways can be measured in minutes, or even seconds.

The "super-express" on the A-train is 5.3km and takes 8 minutes, that's 40km/h average (e.g. half of 55mph, with no intermediate stops).

This makes me sad about missed potential. I've taken the ski train from Central London to the French Alps - was great. With high speed rail it should be possible to get a 8hr overnight train from Chicago (or Dallas & Houston) to Winter Park CO. I'm guessing it'll never happen


Running the ski train in Europe makes sense because it's leveraging high speed rail that was built and is mostly used for other things; very little new stuff was needed.

Chicago-Denver, as a city pair, is so sparsely populated and so far apart that new HSR doesn't pencil out, nor does building it out incrementally (it's hard to justify KC to Denver or Omaha to Denver)

We built a rail system that spanned the country once and the highway system doesn't just maintain itself.

That rail system still exists. Asking for a national high-speed rail system is like giving everyone a Rolls Royce.

We un-deregaulte rail and force them to offer the passenger service they used to do an excellent job at delivering.

Imagine if the response to the invention of rail were "well we just built all those canals."

The original rail network was on par with airplanes in terms of relative improvement of speed of travel. The highway network did the same for travel to areas unlikely to ever get a rail connection. (There aren't that many sparsely populated countries with a rail connection to every town; even fairly dense countries don't manage this.)

In today's world, HSR fills a very specific market niche; planes are much faster for long distances due to higher speed and small requirements for land acquisition, and cars are just as competitive for short distances due to the last-mile problem. So I don't think it's reasonable to connect every possible city pair in the US with HSR, but rather it's much better to just build strong regional networks of HSR that work on their own. There's no reason to build a continent-spanning HSR network.

The parent is talking about subway "trains" (sets of coaches, with each car having its own small electric engines, all powered by a common third rail); not, er, train trains (vehicles that pull themselves along two rails using a locomotive.)

Nitpick: self-propelling cars are not a distinguishing property for light rail. Some modern high speed trains use self-propelling cars[0].

[0] https://en.wikipedia.org/wiki/Siemens_Velaro

The technical term is "electric multiple unit", and it's becoming the default for high-speed trains use them:



High speed aside, in the UK, i think pretty much all inner and outer suburban main-line trains are electric or diesel multiple units (eg if you get a train from London to Cambridge that takes 45 minutes, that's an EMU). According to some random report i found on wikipedia [1], in 2011 there were 1248 locomotive-hauled carriages, 2892 DMU carriages, and 8046 EMU carriages. I could tell you a great many more particulars but suppose that you are tired of it by this time.

[1] http://archive.nr.co.uk/browse%20documents/rus%20documents/r...

Loco hauled are very rare - HSTs from the 70s mainly being replaced on GWR mainline, a little use on WCML and cross country 2 or 3 sets?) and I think the east coast does them too. And Eurostar of course.

Aside from those it's freight and the sleepers isn't it?

If you're counting HSTs then everything on the east coast, plus great Eastern main line, the sleepers, East Midlands HSTs, Chilterns trains, 4 xc sets, the new transpennine sets that are being built, a few Cumbrian coast services, and within a few weeks the great western pocket rocket HSTs and the scotrail ones. Less popular than they used to be, but very rare is overstating it a bit

> Drivers go even slower than 45mph because the train is automatically stopped when going faster (as you would want to do if there was a train ahead) instead of just giving a warning and letting the train keep going.

Is there any protection system that allows the maximum speed to be exceeded with just a warning? Certainly all those I can think of (TPWS, TVM, PZB, or something more modern like ETCS) have a given maximum speed you're allowed to exceed by a small tolerance but nothing more (though, of course, the system may be isolated, but that's not allowed in passenger service).

The failing here sounds like it is setting the line speed to 45mph rather than anything higher.

There are... one of them is given in the article—which calls the current system "one-shot"—there exists k-shot variants which allow for at least k opportunities/warnings before automagically braking. The (apparent) reasoning behind the MTA not putting these (for some number k>1) was "safety," but the article states that, in general, the timing circuits are much more complicated therefore resulting in a higher cost.[0]

Additionally, a simple case would be to have circuits which do, after a certain point in time, force slow (but safe) braking into the required velocity---I could see problems with robustness here due to the specific application, but at least in theory this should be possible (we all know the difference between theory and practice, though).

On the general case, this all seems a little weird to me since it seems like such a trivial problem (increase error bands, etc), but may have some deeper reasons that aren't explained in the article(?). It's quite difficult to tell since the primary sources are not provided at any point in time.


[0] I'm reluctant to accept this claim since most of these things aren't done in an analog way, anymore. Timing is a matter of figuring out empirical constants and programming them in directly, which, while not totally trivial, seems to have roughly the same difficulty (amortized over all of the lights that need to be put in) as doing the timing for a 'one-shot' light.

Indeed, this is my understanding of how most non-US high-speed-rail controllers work - the driver sets a target speed, the physical location on the track has a maximum speed, and the traffic system looks at the cars ahead and sets its own maximum speed (e.g. zero when another train is right in front of you). Instead of stopping the train when the driver exceeds the speed limit, it just uses the minimum of all the target speeds.

TVM is the system I'm most familiar with when it comes to HSR (used on the LGV in France and HS1 in the UK); it has a variable overspeed allowance (5 to 15 km/h depending on current maximum speed). If you exceed that, an emergency brake application occurs.

Japanese ATC and ATS-P, used on all Shinkansen lines and most busy metro/commuter lines, just apply service break until the train is under back speed limit. AFAIK no overspeed allowance.

I think the reason that it's a single shot instead of multi shot and full emergency braking instead of moderate is that the stops are all track based not train based, at least for the 'train in circuit' stops. The original signals basically put up a metal arm that trips the emergency brakes on the train itself so if they piggy backed off this existing system by just adding a speed sensor there's no way to do a slow deceleration because the signals are designed to work without any real cooperation from the train beyond the emergency brake functioning.

That, and considering a speed limiter instead of stopping the train when traveling too fast. What good does it do to stop the train when going too fast instead of limiting it in the first place?

The general principle in railway movement authority is that there's an easy fail safe: stop moving. If nothing moves, there can be no collision.

If something goes wrong? Stop the train. You can then figure out what went wrong and address it with no risk. Things shouldn't be going wrong, speeds shouldn't be being exceeded, etc.

>If nothing moves, there can be no collision.

If no train on the line moves, sure. But stoping just the one train won't improve safety, and was historically the cause of many accidents. I remember just the one in Shanghai in 201x. One train stopped, the following train hit it.

That shouldn't happen under normal circumstances. That's whole idea behind railway signals.

Rail segments (blocks) are supposed to be automatically locked off while a train is on that segment of track. No two trains can be on the same block at the same time. If a train stops, the signal guarding that block will remain red, ordering the next train to stop before it gets anywhere close.

If stopping doesn't fix things, you're not using enough of it. Time for networked emergency brakes on all trains!

Europe may has that, with the GSM-R system. I remember reading of at least one minor accident that stayed minor because the driver to discover the landslide over the track at low speed passed the button, stopping the incoming faster train on the other track.


AFAIK, this is just a Czech extension to GSM-R. That said, I recall various instances of GSM-R being used in the UK to broadcast a general message to stop, albeit one that the driver has to act upon.

In fact, the crash at the origin of this whol debacle was caused by a train being stopped on the tracks and being rear-ended by another.

That's an interesting way of looking at the cause and effect. Trains unexpectedly stopping in an expected and inevitable property of all networks, whereas trains exceeding their movement authority is not.

Stopping a train anywhere should therefore always be safe, so the safety failing is with the train behind, not the one in front.

Even then you should have some give aka Speed of X+3 MPH break, speed of x+1 MPH warning. You want fault tolerance not simply a hard limit without feedback.

Just a PID speed controller, maybe. Then, some boundary condition where when the controllers inputs mismatched from predicted outputs, then throw an error. But, if you're trying to drive 45, and you achieve 46 with a tail wind it doesn't seem necessary or safer to stop the train.

I think the idea is that if the operator is incapacitated or the control system is malfunctioning, stopping the train is the safest solution. While true, a multi-shot timer would be equally effective in detecting such a situation.

Wow. In the Netherlands, intercity trains can go up to 135 km/h (83 mph), and they’ve successfully increased the frequency of these trains on some lines this year. Suffice to say with Dutch population density, the train network is also quite dense. Quite an achievement I think.

The MTA subway isn't really intercity, unless you think of the boroughs as different cities. Although the express trains should be able to go much faster than they do, like the A from 59 St to 125 St, which takes ~8 minutes for 5.3 km even though there are no stops and no other trains on the track.

This kind of thing makes a huge impact on the quality of life for people living in the city.

For intercity trains in the Netherlands some stops are only kilometers apart as well. 50 kilometers between stops would be the absolute maximum.

Yeah. The larger point is still that you just can’t compare IC trains to subways. The TGV gets to 300km/h+, but it’s irrelevant to the Disneyland Monorail.

Many of the subway stops in Manhattan are only hundreds of meters apart, so it is still an order of magnitude of difference.

A closer parallel might be the London Underground, that manages 55-60mph.

The new Elizabeth line is apparently going to go up to 90 miles per hour, but that is probably only for when the train leaves London and heads out to neighbouring towns (it's essentially a subway line that also connects commuter towns to London).

I agree with most of your summary, but I don't believe the article provided evidence that installation of signals to enforce a lower max speed "does not increase safety".

I'd speculate, without evidence, that it probably has increased safety. But maybe it's just a bad tradeoff, compared to other options.

This is one of those comments that is annoying to respond to because it may be technically correct because the original comment wasn't worded ever so precisely, but we all know that "does not increase safety" is not an absolute statement. You could strap a pillow to your chest and get ready to jump out of an airplane, and I'll say "that won't cushion your fall," but "technically, it will," you argue.

Besides--reduced capacity causes platform crowding which is dangerous and displaces passengers into other modes of transportation like foot or automotive traffic, both of which are more deadly per passenger-mile than the subway.

Enforcing the speed limit the system is designed for intuitively sounds like very reasonable and effective safety measure, not comparable in any way to your fantasy scenario.

The whole point of having an analogy in the first place is to sacrifice accuracy in order to illustrate a point in a more obvious way. Analogies are often fantasy scenarios. And making arguments about how things “intuitively sound” is a big step into a world of unscientific thinking.

The question here is, “What is the relationship between train speed and safety?” My guess is that the risk has a bathtub curve to it. At low speeds, certain factors dominate, the dangerous effects of reduced capacity. At high speed, other factors dominate, like crashes. This type of curve is found in many different types of systems. This curve exists in automotive traffic, where both high-speed and low-speed drivers are known to be dangerous, and speed limits that are either too high or too low make the road more dangerous. With an actual study, you could try to measure the curve. Perhaps you would find a large, flat area where changes in speed have small or even negligible effects on safety.

Kind of like the negligible effect on safety that wearing a pillow on an airplane would have.

But we don’t know the answer to the actual interesting question, do we? We don’t know what the exact effect of speed on train safety is. We know that train safety is very different from car safety—I was just looking at some statistics comparing them. But here we are, dodging the real question and complaining that an analogy isn’t realistic enough.

The article glosses over that there were three train-on-train collisions in 1995, and a couple earlier in the 1990s. There was another incident when an IRT sped too fast through Union Square (there's a significant bend in Union Square for the 4/5/6 trains) and derailed.

This may confirm my memory of 2/3 trains hurtling at what I recall as much faster speeds in the early 90s.

This would not be the first time that the MTA has willfully deceived the public by maintaining two sets of facts - the private set which reflects the truth and another set of facts that are used in discussion with the public.

15 years ago the MTA was caught maintaining two different sets of books - one internal that showed the agency had a surplus of cash and another set of books for the public which showed the agency as being "cash strapped" and needing to raise fares as a result. See:


The culture of this agency is rotten and its corruption endemic. It's hard to believe it's capable of producing anything other than it's current broken state.

About 8 years ago there were a few house ads in the cars that were effectively gaslighting riders. It was one of the craziest things I'd ever seen as a consumer -- borderline antagonistic. The agency, and the workers to some extent, can be pugilistic when criticized.


For those of us who aren't familiar with the price history what's the lie here?

The MTA is comparing single-ride fares with "effective" cost per ride of a then-$81 30 day pass, which they amortize over around 70 rides, which is a bit high for the average person in my experience (though not ridiculous).

It might be more accurate to compare the single-ride fare then with the single-ride fare at the time the photo was taken, which was either $2.00 or $2.25 (the price changed in 2009, I think).

I surprised that they don't have a commuter ticket system. In my city the commuter ticket price is calculated to be somewhat cheaper than return fare*20 as well as being a tax deductible.

Did they have multi-day passes back then? It might be reasonable to compare the cheapest options available then.

They did not. You actually bought metal subway tokens that you dropped into a coin slot on the turnstile.



They're comparing a single-ride fare to a monthly-pass fare. A single ride fare cost $2.00 in 2008[1]

[1] https://en.wikipedia.org/wiki/New_York_City_transit_fares#Fa...

The obvious part of the lie to me is that they would raise fares to increase revenue, and claim that because of the 30-day and 7-day discounts you wouldn't be paying any more.

Huh? How are you going to increase revenue if no one is paying more.

Tourists. We have quite a few of them here.

Right and if they ran that ad today, 10 years later the single ride fare has more than doubled @ $2.75 a ride on a $131 dollar monthly pass.

Twice as much for worse service.

How is this gaslighting? Transportation costs have done a reasonable job of not growing faster than inflation...

> The culture of this agency is rotten and its corruption endemic

How does one realistically reform something like this?

>"How does one realistically reform something like this?"

Slowly and methodically.

For starters:

Create legislation that prevents construction firms from making campaign contribution and MTA employees from taking jobs as political consultants. This is probably the largest source of corruption, graft and waste.[1]

The MTA's Board of Directors is predominantly controlled by the Governor of the state. The Board of Directors has long been a political dumping ground. Residents - the people who depend on the train have almost no say. Residents should have representation on the board of directors even if its a only a handful of rotating seats. Having people on the board who actually use and depend on the train would provide an instant feedback loop.

I think you also need to implement a rotating independent oversight committee whom the MTA must be beholden too. One that is democratically elected. The oversight committee should have real teeth though perhaps with veto ability in certain instances.

And yes as someone else mentioned you probably need to fire a whole bunch of people.

[1] https://www.nytimes.com/2017/12/28/nyregion/new-york-subway-...

The main issue is that the MTA and its predecessors have always been dependent on one person (now the Governor, formerly the Mayor). The actual board members have no credibility and no accountability, hence why we are here today.

The MTA would be much better off if the appointed representative of the county was replaced with the actual elected executive of said county (so county executives, borough presidents, mayor/council speaker/what have you, etc.)

It's never been "entirely" dependent of one person, although the majority control seems to be the Governor. Either through his direct 5 seat appointments or indirectly through the 7 other non-NYC seats.

I don't believe it was ever under Mayoral control. The current governance structure came under then Gvernor Nelson Rockefeller.

But anyway Nassau, Suffolk, Westcheste, Dutchess, Orange, Rockland, and Putnam counties each get to appoint a board member. That's nearly a third of the 17 total seats. Beside Wescheter and Nassau counties none of those counties even border any one of the 5 boroughs.

That's 5 votes by people who most assuredly do not ride the subway to work and maybe not at all. As such it's hard to believe they would ever have NYC residents best interests at heart.

I am referring to the state of the subway and buses under the NYC Board of Transportation, which was similarly a giant incompetent mess.

The distinction to make is that the counties and the City get to recommend the appointees for their slots to the Governor. It's a formality most of the time, but the Governor or State Legislature could tell the mayor or county executives to "try again".

As far as that last point, it's important to understand why the MTA was created in the first place; its primary mission is to dole out the goodies from the State and the tolled crossings to its subsidiaries. What they know about the system is at best, secondary to the job, since they are, by and large, not the people who run the system. That's why they appoint chairmen and agency presidents and all that.

The MTA isn't just the subways it's the regional rail and transit system too, including Metro North, LIRR, and so on. It's not obviously irrational for all the constituent counties to be involved.

Yes and that's also a huge structural problem.

The LIRR, the Metro North and NYC Subway system are completely independent lines, with completely separate payment systems, completely separate citizens as their majority riderships and they don't share any tracks. They are completely separate concerns - the LIRR and Metro North serves the commuter belt people that most likely have a car or two in the driveway. The NYC Subway's majority ridership are people that have neither a car or a driveway at home.

They are non-contiguous in every way. The only thing they share is a bloated org chart.

So yes, the idea that someone from Duchess County who drives a car to work and has most likely never ridden the NYC Subway to get to work is making "informed" decision about whether or not to address the problem of overcrowding on say the 456 line is perhaps the height of "irrationality."

This comment doesn't really make sense. There are people that never fly on airplanes but still vote for Federal officials who manage the FAA. There are people in NYC who vote for Federal officials that administer cattle grazing. So what?

If anything the problem with regional transit systems in the US is that they are far too provincial and federated, leading to all sorts of overlapping and wasteful things. For example the fact that the Port Authority and the MTA don't get along means I can't take a fucking train to any airport in the most important and densest city in the country.

Or stuff like how they didn't bother to connect the 4/5/6 line to the PATH system during WTC reconstruction despite the fact that they are compatible, and they apparently had over three billion dollars to light on fire.

Or stuff like how the PATH train runs to Newark and inexplicably stops a mile or two short of the airport despite the fact that NJ Transit/Amtrak right of way runs right there down the exact same line.

The idea that separating LIRR from Metro North from the Subway and creating yet another set of competing bureaucracies is going to somehow help things is delusional, and a fundamental misunderstanding of how NYC area politics works.

>"There are people that never fly on airplanes but still vote for Federal officials who manage the FAA. There are people in NYC who vote for Federal officials that administer cattle grazing. So what?"

Yes exactly those people vote in those instances. The key word being vote - i.e they have a say and can research the candidate before casting a ballot. The context here is positions where people CAN'T vote because those MTA board positions are political appointees. Ir'a apples and oranges. Your comparison makes no sense.

>"Or stuff like how they didn't bother to connect the 4/5/6 line to the PATH system during WTC reconstruction despite the fact that they are compatible"

Except you are wrong those are actually connected via the Dey Street Concourse. See:


>"The idea that separating LIRR from Metro North from the Subway and creating yet another set of competing bureaucracies is going to somehow help things is delusional, and a fundamental misunderstanding of how NYC area politics works."

No its quite simple, if infrastructure passes through multiple counties its under the purview of a state agency. I its infrastructure that falls within a single county or just the 5 boroughs then it should be a city specific agency.

Lastly I have actually worked in NYC politics so I'm pretty sure I have a firm grasp of how NYC politics works.

It's interesting that you choose to make statements like that without knowing anything about a person's professional background. So an opinion contrary to your own constitutes delusion and indicates that someone misunderstand how things work? Is that correct? That seems like a pretty fool-proof prescription for an echo chamber.

> Yes exactly those people vote in those instances. The key word being vote - i.e they have a say and can research the candidate before casting a ballot. The context here is positions where people CAN'T vote because those MTA board positions are political appointees. Ir'a apples and oranges. Your comparison makes no sense.

That's flatly untrue. People don't vote for FAA administrators or USDA policymakers, they vote for representatives who pick civil service people and so on. The MTA governance isn't strange or unusual, that's my point.

> Except you are wrong those are actually connected via the Dey Street Concourse.

Well yes, the lines are connected if you get out of the train and walk. By that definition all transit lines in the country are connected.

But we could have had a subway that runs from Grand Central to Newark: https://newyorkyimby.com/2014/09/the-port-authoritys-missed-...

I mean in Europe they can figure out how to run fast trains through three countries in an hour or two. But I have to transfer three times to get to the airport. Our way of doing this sucks.

> No its quite simple, if infrastructure passes through multiple counties its under the purview of a state agency. I its infrastructure that falls within a single county or just the 5 boroughs then it should be a city specific agency.

But.... what if you don't want the infrastructure to just fall within a single jurisdiction. That's the entire point I'm making. I want transportation infrastructure to be planned out and integrated on a larger scale. I want to consolidate the various agencies. Having MTA and PA and Amtrak and NYC DOT and NYS DOT and NJ DOT and so on makes for insane schizophrenic policy. Your idea that the subways and regional rails should be split into separate agencies makes that worse not better.

The question is whether or not it has to be this way. Short of politics there is no reason that the LIRR and Metro-North could not also work for intra-city or intra-regional travel; in fact, many umbrella organizations contain merged separate railroads for exactly this purpose (STIF operates the Paris RER, Berlin's BVG operates the S-Bahn, TfL operates the Overground and TfL Rail, etc.)

Without the regional integration, we would probably never get things like the Freedom Ticket; utilizing commuter rail lines as de facto subways, even premium ones, is much preferable to the much more expensive option of duplicating existing corridors.

You fire everyone.

However, if I remember correctly, they're unionized, which is why this hasn't happened yet.

Is management unionized? The article makes it sound like the engineers themselves are frustrated with this system.

You can't just stop the gravy train by removing management. Bad culture is at all levels.

Bad culture is at all levels, but management control the factors that turn culture bad.

If you'd like to see a fine example of how changing the management and the management philosophy can turn a place around, look at NUMMI. There's a This American Life episode with extensive worker interviews: https://www.thisamericanlife.org/561/nummi-2015

They are not.

Fire everyone at Agency A for shitty behavior. Then turn around and realize everyone at Agency B is cooking the books, so fire them. And so on.

Then turn around and realize basically every large institution humans have ever built relies on shady, selfish, behavior and stop caring about reforming individuals since thousands of years of policing have done jack and squat to remove the behavior from the species.

Seriously, this forum is the worst for overlooking relevant variables (human biology and contemporary culture as a whole) for their trite commentary, usually rooted in nothing more than culturally acceptable bullet points.

Hundreds of years of the same sort of thinking. Still happens.

So why does Tokyo’s system work so much better? They’re humans too, but apparently their transit system works much better.

I don't know, but Japanese companies are notorious for never firing people, so it's very unlikely to be because they "fired everyone" as DiabloD3 was suggesting.

They're privatized.

It seems like you're saying "it's impossible to achieve perfection, therefore you should never even try to improve."

Not saying I disagree with you - but assuming what you say is true, what would you propose?

I don't think it's the workers that are the problem, from the article they sound just as frustrated (and indeed seem to be the source of this article). The management is not in a union. How does firing the workers fix anything?

If you fire everyone, wouldn't that stop the trains? I mean, you have to hire new people and get them trained on what to do?

Doesn't a union make that easier? That's the point of unions.

A union makes it easier to fire workers? What?

It makes it easier to collectively fire workers.

Which will never happen. Yet another of China's many advantages: ease of optimization.

As long as you agree with the optimization that the man in charge is implementing.

If you don't, well, in China, sucks to be you.

It depends. You may not enjoy certain micro-optimizations, but to me it seems perfectly possible that even though living in a communist dictatorship has many negatives, the net benefits under a benevolent dictator could easily surpass the drawbacks, especially when compared to a free but largely corrupt Western society where the benefits may not make it outside the top 10%.

And this isn't even getting into the longer term prospects of the very real danger of the efficiency and scale of China starting to eat into the core economies of other nations as they move away from simple manufacturing and beginning moving up the value chain, as we are now seeing.

> And this isn't even getting into the longer term prospects of the very real danger of the efficiency and scale of China starting to eat into the core economies of other nations as they move away from simple manufacturing and beginning moving up the value chain, as we are now seeing.

I wasn't really that worried about China as they are automating extremely quickly and have shown no better ability to utilize the displaced people any better than anybody in the West--see the recent layoffs in steel and coal in China along with the corresponding strikes and protests.

Now with Xi Jinping being President for Life, I'm even less worried. Xi Jingping will regard striking and protesting workers as a threat to his power and he will suppress them. This will break what little feedback there is in China between the working class and the ruling class and the decisions will become increasingly unmoored from external reality.

If I were TSMC, I would start moving fab lines out of Taiwan.

I think I'm missing something - TSMC should move their fabs off the island because of some goings-on in the mainland? The countries are obviously connected in various ways, but they're definitely not yet united so I'm struggling to see what the reasoning here is

> The countries are obviously connected in various ways, but they're definitely not yet united so I'm struggling to see what the reasoning here is

China doesn't see it this way. And the only thing keeping China from steamrolling both Taiwan and Hong Kong are the fact that it would decimate the economies and businesses there.

However, once you have a leader more worried about his power rather than being beholden to the working class, that is no longer an obstacle.

Well, I guess, one could argument than after losing contact with working class and solidifying their rule, China's ruling class will what every non-democratic regime did --- try to expand territoriality. And they have grudges with Taiwan.

How do you know what individuals will do?

> benevolent dictator

First of all, https://www.theguardian.com/environment/blog/2010/nov/19/cra...

Second of all, and this is probably something that was brought up in your high school civics class, the advantage of a democracy is that the overall quality of a government will be the average of all of the people in it. Speaking metaphorically, you can roll N dice and average the results together; doing this repeatedly, you'll notice that as N goes up, the resulting averages will follow a normal distribution more and more closely. If you have enough dice, you can assume that you'll hit the middle of the curve most of the time, instead of having to worry that you might roll a 1.

No benevolent dictatorship has ever existed. It is hard to imagine how it could exist; it seems that corruption is such a natural and apparently reasonable behavior that we will always see it.

Your a theoretical should be discarded.

Do you believe it is possible for outcomes to vary under different forms of government?

It's interesting how many people on HN seem to believe they are fortune tellers and mind readers now. Personally, I am speculating, but it sounds like all you guys know that I'm wrong.

I've continued to be mostly right for the last decade, I think I'm going to stick with my predictions rather than those of people who don't take too kindly to disagreement.

I think that under all forms of government powerful people are going to try to accrete more power to themselves. Difference of outcomes therefore are based in the ability to restrict that.

A benevolent dictatorship cannot exist et alia because a person cannot live for ever and I simply don't know how you could assure that the transition from one benevolent dictator to another when the first one ceases to be competent. For instance, let's presume that Xi Jinping is the perfect president of China. When he dies or retires or becomes too ill to govern, will the PRC be able to select a new perfect president to replace him? or will they choose someone who is inclined to serve their own interests and was able to benefit from Xi's time? This is the ultimate problem of a benevolent dictorship: it cannot assure state continuity and stability beyond the effective lifespan of the dictatorship.

People also tend to become more habitual as they do something for a long time. Politeness and friendships limit them from being able to introduce policies that will benefit the country at the expense of those who have their ears. So it's possible that a governing style which served the state and the country well for many years will need hard revisions. Periodic changes in governor (as democracies have) mean it's more likely the incoming team will be able to retain what is still fresh and discard what is now stale, and they will be connected with different people and therefore not too worried about upsetting some of them. (I think this is a good part of what's wrong at the moment: major parties around the world are connected too much to the same interests, so a change of government doesn't result in as much upset as needs to happen. Votes will gradually transfer from the establishment to the anti-establishment parties. They will be able to upset the entrenchment, but it remains to be seen whether they will do so in the interests of the nation or the interests of an alternative elite.)

As for your last sentence, I think I misunderstand it. I disagreed with you and straightforwardly stated my opinion, and then it seems that you attacked me: it reads like you're accusing me of not being able to handle disagreement. I don't think I attacked you. Perhaps you would clarify it, or highlight the place where you think I "don't take too kindly to disagreement".

> A benevolent dictatorship cannot exist et alia because a person cannot live for ever and I simply don't know how you could assure that the transition from one benevolent dictator to another when the first one ceases to be competent. For instance, let's presume that Xi Jinping is the perfect president of China. When he dies or retires or becomes too ill to govern, will the PRC be able to select a new perfect president to replace him? or will they choose someone who is inclined to serve their own interests and was able to benefit from Xi's time? This is the ultimate problem of a benevolent dictorship: it cannot assure state continuity and stability beyond the effective lifespan of the dictatorship.

Agreed, this is why a dictatorship is generally not a wise idea. I certainly cannot assure a transition to another benevolent dictator, but what's interesting if you peruse the comments here, the overwhelming sentiment of people is that they can assure that a successful transition can not and will not happen, ever, full stop. As you may have noticed, I often have a bit of an issue with self-proclaimed mind-reading, future-telling, and other supernatural capabilities. You would expect this sort of thing among the general public or on Reddit, is it too much to ask for a higher standard on HN? It would seem so.

As for my last sentence, admittedly I'm guilty of taking out my frustrations with HNers in general (for the above stated reasons) on you, in response to your absolutist stance (No benevolent dictatorship has ever existed); perhaps no major nation in recorded history been led by a purely benevolent dictator, but if we had access to a truly omniscient being to settle this disagreement, I'd happily throw down a $100 wager that many truly benevolent (to the best of their abilities) dictators have led smaller, less famous groups of people on many occasions throughout history. Perhaps I have too much faith in humanity at its best.

> No benevolent dictatorship has ever existed.

I can quibble with this just a little. A truly benevolent dictator will recognize how harmful dictatorship is and will immediately remove himself from it.


I think if someone can foresee that they won't be a benevolent dictator and refuses the office, I don't think that's the same as being a benevolent dictator.

Maybe we're playing semantic games, but Americans worship their leaders in a way that makes me feel uncomfortable. I don't think it's necessary for me to limit my statement in order to participate in your religion.

> Yet another of China's many advantages

Yes, you can fire everyone in China, throw them in a work camp, take their property for the greater good, etc.

Funny thing happens when you walk all over people that way for the "greater good" ... a few people at the top, who define what the "greater good" is, tend to become extremely wealthy and powerful. The society becomes lopsided. The peasants become frustrated. Then there's violence.

Not that the people at the top are idiots. They understand this.


Incentives work. Figuring out the correct incentives to achieve good results is hard. Keeping ahead of the people who will game them is hard. This requires competent management, which is obviously lacking. Fire the incompetent managers and fire the political leadership who appointed them (term limits might help here).

Since there are credible allegations of crime, is there an inspector general? There should be. If not, bring in the fraud squad and investigate the hell out of them. They'll leave.

IANAL, but keeping double books should land you in prison, as simple as that.

We, collectively, seem to accept car accidents because we think we can avoid them with good driving skills. Train accidents leave us helpless, and we can’t cope with that.

The whole concept of ”bad things happen to good people for no reason” is extremely hard to bear. Somehow this reminds me of people who believe the right diet will protect them from cancer, the right prenatal care from birth defects, etc.

This will be the hardest thing to accept about self-driving cars. If the US goes from 37k auto deaths per year to 1k deaths, but some of those are "my Tesla sped into a brick wall at 200mph with little Susie in it," how many people grab the steering wheel?

For Susie parents it's basically: "A single death is a tragedy; a million deaths is a statistic."

It's not just that.

Look, let's say a Tesla, oh, just as a random example, careens directly into a clearly visible truck because its cameras can't distinguish white trucks from bright sunny skies, killing "my daughter Susie." In this hypothetical, this is a truly autonomous Tesla, and overall, these truly autonomous Teslas are less crash-prone than the human average.

As Susie's dad, my argument is that this accident was caused by Tesla. Yes, Tesla may have also saved some other people, under different circumstances, circumstances which are harder for humans to deal with and easier for cars to deal with. So what? That doesn't change the fact that Susie was killed by their product misbehaving in a way that a basically competent human driver wouldn't have. It's not like Tesla previously saved Susie's life, and at least she got to live longer than she would have in the counterfactual -- it's not like all the Susies out there in the world are all essentially doomed to die in auto accidents today.

If Susie went to a doctor with a mild infection and he gave her an inappropriate drug and it killed her, you had better believe that I would sue the shit out of him and also want him prosecuted, even if in the year before that he saved 37 other kids' lives from other illnesses.

I don't think you should be able to win that lawsuit against the doctor unless there was actual negligence on doctor's part.

But that's the difference of being against self driving cars, and a specific model of a car (for which it's manufacturer is liable). The OP was saying, to use your analogy, we still use doctors in the U.S. because it changed deaths from medical problems from like 90% to 400,000 people killed every year by accident.

It’s not about the manafacturer. There isn’t a ledger where lives saved can be balanced against lives taken.

Lives saved is effectively transparent. If you are only acknowledged for negative outcomes then you’re incentivised to not play the game.

I don't think that is the crucial difference. In a train or plane crash, hundreds die in a single event. But when a hundred people die in a hundred car crashes every day, the individual event is barely worth a mention on page 6 of the news.

It's the same mentality that means we have root cause analysis for plane and train crashes but for a car crash, SOP is pretty much to blame 1) whoever died or 2) noone at all, and 3) never the infrastructure.

Not only are car crashes less salient than train crashes, but also the converse effect: the inconvenience of car speed limits is more salient than the inconvenience of train speed limits, because you're the one driving.

I think the cycle feeds into itself though: part of the reason planes and trains are so safe is because of the extensive root-cause analysis of crashes, which routinely result in suggestions for improvement that make the rest of travel safer. Remove the investigations and you'd end up with more plane and train crashes.

When a car crashes, we just scrape up the debris and move on, after usually assigning blame to one of the drivers who ends up with higher insurance premiums or a lawsuit, assuming they survived. There's little effort to learn from the crashes.

Different risk profile. Trains generally don't have belts, crumple zones, padded surfaces, etc. Their risk profile dictates they avoid accidents. Cars are more chaotic, even in automated form. They are designed to crash safely as possible first, then anti-crash measures are added. Because trains depend soley on the anti crash measures, they must be stricter. The issue of harming dozens rather than a few humans also factors. Its an uneviable position for those responsible imho. Every move is going to be a little wrong and costly on some level.

You're absolutely correct about fear combined with lack of control.

It's the same thing that powers fears over mass shootings even though, statistically, they are a tiny fraction of overall gun violence. Most of the rest, however, can be avoided by keeping yourself out of bad situations with gangs or drugs...so even though the stats are higher people feel more in control...vs just being in the wrong place at the wrong time.

It's worth looking at "for no reason" in greater detail for cars, which is different from those other medical cases in significant ways. We collectively as a society decided that motorization of roads was worth pursuing at the expense of human health and safety, if you look at rhetoric at the time of cars' introduction, in no small part due to input from car manufacturers. This wasn't about driving skills; it was about restructuring all of society to make cars close to the bottom rung, not infrequently at the expense of minority groups. More recently, car-related deaths occur from bad design and usually refusal to fix it until sufficiently many lives have been taken. Any kind of "Vision Zero" which is focused on enforcement rather than infrastructure fixes is bound to fail.

So if "bad things happen to good people for no reason" is unpalatable and it's the narrative for mass transit, and "bad things happen to drivers for reasons in their control" is the easy narrative for cars, I'd say "bad things happen to good people for reasons that are in our control but we still do nothing" is closer to the true narrative for cars, but it is also unpalatable. There is an element of this in many risk-related scenarios, but it is especially clear here.

The "Just World phenomenon." You also see it when people blame the victim in a lot of cases. "I wonder what they did to deserve that!"

I'm not lack of control is the deciding factor here. People still accept car accidents even if they're not the ones driving the car, such as when you're riding a bus.

Busses are really heavy, so they tend to 'win' in a crash by sheer inertia. A better example might be taxis, but there you kind of have the control of screaming "I want to get out NOW" if the driver seems incompetent.

But there is one type of delay that’s gotten exponentially worse during that time: a catchall category blandly titled “insufficient capacity, excess dwell, unknown,” which captures every delay without an obvious cause.

This smacks of corruption or of elements of society deciding that systems aren't going to simply work anymore. The population of New York City when I lived in Queens was 7.3 million when I lived there in 1989. The subways pretty much worked back then. The figure for population of New York City in 2017 I found was 8.5 million. Population increase can't account for the difference. There must have been considerable systemic decay.

NYC's transit mode share for commute trips is above 50% [1]. Of the additional 1.2 million people, this means that this population growth probably produced at least 1.2 million additional daily rides, since 50% of the 1.2 million rides the subway twice a day, to and from work.

1.2 million additional daily riders is about double the entire Washington Metro daily ridership and triple that of BART's daily ridership! Indeed, 1.2 million daily riders is about equal to ridership on the Lex! So I disagree with your conclusion that "Population increase can't account for the difference". This sort of ridership growth would have destroyed any other comparable system in the US. If anything, it's amazing that the NYC subway hasn't fallen apart more than it already has.

[1] https://www.thetransportpolitic.com/databook/travel-mode-sha...

The ridership was higher in 1948 and there was no crisis of sardine packed trains and crowded platforms then. MTA is intentionally running the system below capacity to cut costs.

That is very true. I rarely commute during peak hours, and the trains are just as packed at these off-hours because the headway drops off like crazy after the peak time. That's just cost-cutting. The system can run with rush hour frequencies at 2 in the morning if someone is willing to pay for it. (Yes, at some point you have to do maintenance; for that a reduction of frequency is perfectly reasonable.)

There are also plenty of places on the system where longer trains could be run. A trains are 600ft long. C trains are run on the exact same tracks with 480ft trains. Nothing feels worse than boarding a C train that's packed because you know there's nothing stopping the MTA from running longer trains there. They claim they don't have the equipment for this, which may be true during rush hours, but it's a complete lie during the rest of the day. There are plenty of 600ft trains ready to go just sitting in the yard.

For some context in case anyone was curious: according to old news articles, the average weekday ridership in 1989 was ~3.7 million rides per day. It is ~5.7 million rides per day now. Ridership in the past few decades has increased faster than the population. One factor cited in some analyses of this phenomenon is the significant decrease in the crime rate relative to the 1980s and 1990s, but that is likely not the only contributor.

> one type of delay that’s gotten exponentially worse

C'mon Village Voice, that's not a log chart!

Growing exponentially has a real meaning and it doesn't mean growing fast.

Yes. Even if you think the complaint is mathematical pedantry, the original sentence still uses a big fancy word wrongly when it could have used simple ones correctly.

Both "...quickly gotten worse..." and "...gotten much worse..." are sharper.

Do take congestion into account; something that can run perfectly smoothly at 90% of capacity may severely struggle at 96% and basically grind to a halt by 99%. It’s not linear.

(I am not stating that congestion plays any part in it, but you can’t simply write it off.)

There's an attempt to make this all sound like a conspiracy, but it all makes sense to me.

There were accidents. They slowed the train to improve safety.

The train is about as safe as ever. Why? Many, many more people are using than the trains than before. It's harder to keep that many people safe, so more caution is warranted.

The only problem I see, if it's true, is that they classified delays stemming from maintenance under the wrong heading. And so far as I can tell, there's no real proof of that.

The story I got was rather different; first that the safety measures were implemented in a ham-fisted manner and secondly that they failed to model or recognise correctly the overall effect that the measures would have on carrying capacity, and they are still failing to recognise it.

I don’t know how fair these accusations are, of course.

But beyond a certain point slowing the trains gains little safety but (eg) pushes people back to the road where (in the UK for example) we spend 10x less to save a life than on the trains.

Thus, overcaution on the trains will kill more people.

Wonder how much further you can take it. If you lose a little safety (trains) and take more people off the road, you increase the risk of a train accident. How many people go back to the roads after a train accident? Is the backlash greater (my 0 data thoughts would be that more people would return to the road then just if the train was slow). I heard somewhere that more people died through increased car accidents after 9/11 (forget time span) that would have flown but didn't.

You don't really need to propagate it all the way to actual lives lost. If there is an hour of life lost on a too safe train, that's a fraction of a life lost.

You can only take it so far as you're willing to hold accountable the people making the decisions.

More importantly, they used the overcrowding excuse as a means to raise $800mm+ from New York taxpayers. If this capital wasn't necessary given the problems they cited were non-existent, this amounts to fraud.

That's not a really fair way to put it as the signaling system is still the root cause of the slowdown.

1. Outdated signaling system cannot handle current load at high safety threshold

2. Network could not upgrade signaling system due to lack of money

3. Network responses by slowing down critical paths and putting in more locks

4. System capacity drops

So yeah, you still need the money to fix slow things.

That rationale implies that your assumption in #1 is true, which seems not to be the case.

Yes, the signals system needs to be upgraded if they want to increase the number of trains (and their speed) on the lines, but even without the signals system upgrades, they can still run the current loads at acceptably-high levels of safety.

Put another way, they are slowing the trains down and increasing delays while not actually increasing the capacity of the system.

I'm rather familiar with the L. Signal upgrades wont create a large improvement - it is already a fully automated line ( there's a sign at the end of every platform that spells out that the trains of the L are under automated traffic control ). It was just designed in the bone-headed way - just look at the 8th ave termination point or realize where the switches are in Manhattan and Brooklyn.

Yup - need to pay to get the signals sorted. I guess the you might unearth a bunch of other things, that need sorted... but it just needs doing.

Well, fraud only if it were deliberately misleading. It seems highly likely that the MTA believes what it is saying, even if it happens to be misguided, and driven by the right motivations.

It is fraud though: they have/had an "other" category, but they called it "overcrowding" and then used the confusion on what overcrowding means to get funds (I'll bet when you hear overcrowding you think too many people not any of the many random things that can go wrong). Maybe it really is overcrowding, but since the category is really other we don't know.

Overcrowding is NOT a misc category, it is a specific situation with a specific definition (the exact definition is something they use should be published so we can decide if we agree with it), with specific training on how to tell if this is really overcrowding as opposed to a failure to get a normal crowd through in time since the solution to the two problems is different.

Other/misc is an import category to have - every system of collecting data needs to have a none of the above because there is always something you didn't think of. What is normally missing is that when other is used there needs to be a text explanation of what happened, and then an investigation to see what should be done. You collect data for things like "medical emergency" because from the category alone everybody understands that the delay is beyond control. When something is misc that means you didn't expect it and need to redesign your procedures so it never happens again. Thus the misc category should be something all operators hate to use: they get sucked into meetings on how do we ensure that it doesn't happen again.

It's either fraud (deliberate) or incompetence bordering on fraud. They have declining ridership, yet are claiming increasing overcrowding.

In either case, probably time to clean house and get rid of the "better safe than sorry" crowd.

If they’re running fewer trains slower, then declining ridership is not incompatible with increased overcrowding.

That's not what I got from reading it.

My reading is that an accident that was most likely caused by a driver falling asleep resulted in cover-your-ass style changes that have little to no measurable effect on safety, but cause massive real problems with the system that negatively affect millions of riders. And beyond that, they're shuffling blame around and waving their hands to distract people from systemic problems.

I mean, if the delays are caused by slowing the trains down, they should say "we've slowed the trains down due to safety reasons, and that's why there are delays". Then people can either accept that and move on, or do actual studies into the problem to determine if their course of action is correct, and if the trade-offs made are worth it.

I don't think it's some big hush-hush conspiracy; it's just garden-variety corruption and a desire to avoid fixing root causes of problems because they're complex and expensive. Easier to just keep collecting that paycheck while using "overcrowding" as an excuse to get more public money diverted to your agency.

Risking asking the obvious: What is stopping us from having autonomous subways, considering we are almost gonna have autonomous cars?

I've said it before [1] and I'll say it again, even if I get downvoted to hell again: Unions

NYC has some of the strongest unions in the country. Everything from hotels to the MTA are plagued by really, really strong unions that only care about their interests -- leaving the city with a shameful transit system, in this case.

Unions make up nearly 25% of workers in NYC[2], that's more than the double the national average. The corruption is rampant[3] and the city is literally powerless against them[4].

[1] https://news.ycombinator.com/item?id=14948921

[2] http://www.crainsnewyork.com/article/20170910/BLOGS01/170909...

[3] https://www.nytimes.com/2017/12/28/nyregion/new-york-subway-...

[4] https://www.nytimes.com/2018/02/16/nyregion/carpenters-nyc-u...

Saying this is because of unions is like saying America has political problems because it is a democratic nation state. Democracies and unions are modes of people collectively exerting power with inherent.

The actual criticism to be made here is the pathology of all New York institutions and the people entrenched among them - the unions that serve the MTA, the people who own, build and rent the real estate, the politicians who effective rule by plutocracy.

The corruption of New York's unions is simply the end state of a city whose underlying mechanisms have all been captured by a small group of people interested only in enriching themselves. Many of those institutions are unions.

There's a difference, though: democracies (in theory) are designed to allow the majority to exert power, while unions are there to allow minorities to exert power. In this case, assuming the parent's assertion is true, you have a union that is exerting power for its own benefit, at a huge detriment to the majority. We can collectively decide that this is ok, if we want: if we value the continued employment of union members over a more efficient subway system, then that's fine. If not, then we're in trouble.

Hell, I'd be fine paying the unionized subway employees for the rest of their lives to sit at home doing literally nothing if that was the only way to get them to agree to automating their jobs away. At least then we'd have the outcome we want, and not have to pay the next several generations of subway operators to do a job that holds the system back.

Paris has been able to convert an existing line to fully automated (GOA4) and is currently working on a second one.

Unions sound more like your personal bogeyman than an actual cause.

Excellent counterpoint, and deserving of greater visibility. London is another example - the transit workers are heavily unionized and regularly go on strike, and yet there've been massive improvements to the Tube, which predates NYC's subway by 50 some odd years. Ever notice how narrow those tubes are, and how deep down they're dug?

To add insult to injury, construction costs for transit are much lower in Paris and London than NYC.

But let's attack the unions again, because that's less complex (and matches a popular and old political narrative) than pointing to systemic incompetence and short term thinking at the administrative level.

Although some lines are automated in London, they still have a driver to push the button.

The DLR is fully automated, but has a member of staff on board, and the 4 times I used it this year the staff member was always in manual control.

If rail doesn't get automated before cars, I don't see branch lines surviving. But then I don't see them surviving automated cars either. Main lines, sure, but I'll be getting an auto-taxi to the mainline station rather than the hourly connecting train.

> The DLR is fully automated, but has a member of staff on board

I believe the rules require a member of staff to check the doors and permit the train to depart, but that can be done from the platform instead (and is, at peak times).

Full automation will be permitted on lines with platform edge doors; the intention is to do this for the deep tubes at least for the central sections, but it's a 3-step process (replace the trains with ones that work with PEDs, fit the PEDs, then remove the driving cabs) that wasn't expected to be done before the late 2020s, and that was before the severe underfunding of TfL with the loss of central government support and the mayor's fare freeze.

So really the answer isn't unions. It's corruption. Unions work in many places in the world.

Per the linked articles, unions are intertwined with the corruption. This wouldn’t be the case if unions were politically touchable. As long as there enough voters who will always say, “unions aren’t the problem”, they will be part of the problem. They get money for nothing and nobody blames them.

It is curious how unions in America behave so very differently to unions in other places. American unions almost universally seem to be short-sighted and completely protectionist. This isn't the case in western Europe, so it doesn't seem to be inherent to a union.

This creates a communication hurdle between Europe and America. The expectations of what a union does differ so much, you could say Europeans and Americans mean a totally different thing when using the same word.

Unions get money to fund lawyers to protect their members, and to be able to go on strike. This protection and feedback is essential for preventing companies from taking advantage of their employees.

The increasing inequality in the US can in large part be attributed to corporations winning the war against Unions. Corporations have been ruthless towards this end, including funding think-tanks that gaslight Americans.

Massachusetts just eliminated all toll plazas in the state and those workers were highly compensated and unionized. Many of those workers were retrained for other positions.

There are ways to make the union happy and the workers happy it just takes longer term thinking which governments and voters are frightently bad at.

Denmark has one of the highest union participation rates in the world and they were able to a build a fully automated metro system in Copenhagen.

Denmark is also considerably different than the U.S. along many dimensions: https://www.bloomberg.com/view/articles/2018-02-23/you-can-t... and not just this one.

So it's not the unions then. It's how they function in NYC.

Yes, exactly my point. Blaming 'unions' for this is simplistic to the extreme. As oyhers have already mentioned many places around the world with strong unions still maintain far better, more advanced metros at far lower costs.

that's weird, the best subway in america is the most unionized one?

Perhaps in spite of? NYC's subway is the way it is because of many factors unrelated to unionization. And the worse transit systems in the US aren't really worse because of their level of unionization.

There was an automatic light rail to Docklands in London when I last visited. In 1991.

Helsinki metro was planning to have automated trains when the western extension opens in 2014. It opened in late 2017, without automation because of technical difficulties. I have a hard time believing there will be much automated traffic on the roads here (with uneven surfaces, unclear markings, snow, and other traffic) before the metro can be automated (in a dedicated tunnel where there is no other traffic, movement is technically only possible on the rails and there is no snow).

Docklands opened in 1987 but is only GOA3 (the train runs automatically but requires a staff member responsible for departure & emergency situations).

Port Island Line (Kobe) opened as GOA4 (fully automated) in 1981 followed by Lille Metro in 1983 and Vancouver's SkyTrain in 1985.

The US has several GOA4 systems, though they're mostly airport shuttles and people movers (AeroTrain at Dulles, Monorail at Tampa, STS at Tacoma) a few serve actual communities (DPM, Metromover, Morgantown, Las Vegas Monorail).

> Helsinki metro was planning to have automated trains when the western extension opens in 2014. It opened in late 2017, without automation because of technical difficulties.

Wow, that's a bit ridiculous.

The main cause of the cancellation was political: there were difficulties with the supplier (Siemens) and the price spiralled up to the point where it would never have paid off.


The Helsinki Metro is also a bit unusual in that it uses Russian wide gauge track (for compatability with the rest of the rail system) but Western tech otherwise, meaning basically everything has to be custom made and is priced accordingly.

I don't really think the rail gauge causes any difficulties for automating the system, though. Tram gauges vary a lot and what manufacturers do is just use the same design and put underneath a different width bogie (wheelset).

(The Finnish rail gauge is in fact nominally different from the current one in Russia, because USSR changed from "imperialist" 1524 mm to "metric" 1520 mm between 1970 and 1990; the trains are in practice compatible, though.)

> Wow, that's a bit ridiculous.

Less than what's happened with the London Underground Sub-Surface Railway, which is now on the third resignalling contract to get them to GoA Level 3.

And here we are approaching the crux of the issue. Building a new line so it's fully automatic is relatively straightforward. Re-engineering an existing line to make it automatic (without shutting it down for months, which understandably is rarely an option) is like performing open heart surgery on a patient while he is running a marathon.

Well, and the SSR is particularly hard (given National Rail services run on parts of the LU-owned SSR infrastructure); the Northern Line automation was nowhere near as painful.

Nothing. Hong Kong, Singapore, London, Paris, Copenhagen as well all have lines that run in driverless mode. Heck, even SF Muni runs in autonomous mode in the subway, it's only run in manual mode on the streets.

> Heck, even SF Muni runs in autonomous mode in the subway, it's only run in manual mode on the streets.

That's a bit of a stretch. Muni and BART use automated train control. SFMTA went with Alcatel/Thales' awful SelTrac system that the Docklands Light Rail uses while BART uses its own monstrosity. Both systems have human operators at all times because neither system is reliable enough to run unattended.

SFMTA used to show which modes trains inside the tunnel were running in. They also used to publish detailed daily service reports that revealed just how bad things were. At one point, > 50% of trains were unable to enter the tunnel in auto mode as the VETAG transponders had a roughly 100% failure rate. For many years the trains themselves would destroy the trackside inductive loops as well, periodically disabling the train control system. Instead the MTA chose to sanitize the reports, and eventually stopped publishing the information at all.

The great irony is that Muni's train control system was so inefficient that drivers would routinely drop into manual mode upon approach to the last underground system without even so much as a heads up to the dispatchers. And then, in 2009, one of the drivers passed out and ran his train into the train parked at the platform. Instead of requiring drivers to pass a health check like the FAA does for pilots, the MTA simply forbade manual operation in the tunnel. From the pictures, that driver was easily 400 lbs and Muni only cares about drugs and color blindness. Net result: 5-10 minutes got added to each outbound trip.

BART, of course, has its own maladies -- if you've ever had to wait for the train to be repositioned before the doors open you know why they still keep their drivers around.

At this point, IMO, neither humans nor automation alone can solve the transit problem.

Everyone is blaming unions, but it is really culture. NYC is very corrupt, has been for a very long time (part of its charm), and this corruption is at all level of society. It just happens that the effects of corruption at the MTA is very visible.

Money. The technology is not hard to create, but it does take a bunch of engineers time and effort to design it all (probably a mix of mechanical and computer work), and then more time to buy the equipment and install it. There is nothing off the shelf, even if some other subway gave them all the design (unlikely), it would still be hundreds of thousands of dollars per train to make it fit the New York trains.

> There is nothing off the shelf, even if some other subway gave them all the design (unlikely), it would still be hundreds of thousands of dollars per train to make it fit the New York trains.

What? Alstom, Ansaldo STS, Bombardier, Siemens, and Thales all have off-the-shelf designs that they will happily sell. And I've probably forgotten some here. I realise none of these are American companies, and "Buy American" is often a constraint on acquisitions in the US, but it's fundamentally untrue that there are no off-the-shelf designs.

See, e.g., https://en.wikipedia.org/wiki/Communications-based_train_con... for various systems using various suppliers' products.

That's the com system, but there are fully automated off-the-shelf systems (com, rolling stock, …). Two examples are Siemens's VAL (Lille, Taipei's Brown, Seoul's U-line, VAL tech has also been used in Lyon's D and Paris's 14 and 1) and Hitachi's AnsaldoBreda Driverless (Copenhagen, Taipei's Yellow, Milan 5, …).

The latter is very popular and getting deployed in several places right now: Honolulu, Thessaloniki, Lima 2, Rome C, …

> Hitachi's AnsaldoBreda Driverless

To nitpick, this is actually Hitachi (the ex-AnsaldoBreda part) and Ansaldo STS (which Hitachi only bought 40% of, and became its own company when Hitachi bought AnsaldoBreda).

Depending on rolling stock and space for equipment, replacing the rolling stock when replacing the signalling equipment might not be required (see, for example, the the London Underground 1995 Stock on the Northern Line).

> The latter is very popular and getting deployed in several places right now: Honolulu, Thessaloniki, Lima 2, Rome C, …

Is Breda really popular? SFMTA forbade Breda from bidding on this latest round of vehicles because the current ones have been so problematic. I don't think Chicago or Seattle had a much better experience with Breda either.

The former AnsaldoBreda… had issues (and that's really an understatement). Now owned by Hitachi, it seems like we aren't seeing the same build quality issues getting into service (though esp. in the earlier days of Hitachi running it, some issues were only found after they left the factory).

I semi stand corrected. Only semi because even if you buy those system isn't isn't a plug and play effort to make them work.

> There is nothing off the shelf, even if some other subway gave them all the design (unlikely)

VAL[0] has been a thing since the Lille Metro opened in 1983 and AnsaldoBreda[1] is being deployed all over the world (it's the platform for Honolulu's upcoming HART).

[0] https://en.wikipedia.org/wiki/Véhicule_Automatique_Léger

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

In principal Bart[0] is supposed to be operated almost automatically and the train operator is there for announcements, closing doors and watching the track for hazards. In practice, the operator does a lot of speeding up/down manually. I can't find it now, but I remember reading an article that claimed that Bart trains were operated in manual mode 90% of the time, not sure the reason.

[0] https://en.wikipedia.org/wiki/Bay_Area_Rapid_Transit#Automat...

Probably because otherwise the operator gets really bored and feels the machine is driving him instead of the other way around.

I doubt BART runs manually 90% of the time. When run manually the speed limit is so much lower you'd know. However BART operators often have to reposition the trains manually within a station and crank the switches manually.

Vancouver SkyTrain opened in 1985 and was fully automated from day one. The rolling stock used on the majority of the system (Bombardier ART) was specifically designed for driverless operation.

Interestingly, at the time the Vancouver SkyTrain was undergoing construction, the same tech was being installed in Toronto for the Scarborough rapid transit line. The transit union in Toronto opposed automation, so the trains had to be specially modified to add operator cabins so the trains could be driven manually. Vancouver was able to get away with it because the system was a greenfield project, there was no union of subway workers around already to oppose automating the trains.

The Las Vegas monorail is entirely automated, but that doesn't stop it from being horrendously expensive and not terribly useful.

We have that in Hungary. Line 4 in Budapest is fully automated. Initially it was manually operated, then the drivers were just a fallback (or to soothe the nerves..), then they pretty much removed the driver's cabin, and nowadays anyone can stand next to and look through the front window. It's cool.

"The line operates using fully automated [..]" https://en.wikipedia.org/wiki/Line_4_(Budapest_Metro)

Poor initial implementation combined with fear.

The DC Metro used to have a semi-autonomous mode that's been turned off for years since there was a glitch that caused a crash.

The "glitch" was decades of deferred maintenance that left sensors inoperable along the line. Theoretically now that the SafeTrack initiative is complete they should be able to return to semi-autonomous operation, but I doubt it.

I'm not expecting the DC Metro to go autonomous anytime soon either. It's a huge jobs program for inner city residents. Basically a commuter tax for a social program.

L (Manhattan/Brooklyn) is automated. The drive is there to simply signal "I'm here" to automated system. The second person on is there to mumble something that is difficult to anyone understand.

It's already a thing, first example that comes to mind is line 14 in the Paris metro: https://en.wikipedia.org/wiki/Paris_M%C3%A9tro_Line_14

Line 1 is even more interesting as it was converted from manually driven operation to fully automated operation without without major interruption to passenger traffic. (https://en.wikipedia.org/wiki/Paris_M%C3%A9tro_Line_1)

And it was considered successful enough that they've started work on converting Line 4. After line 14 was put in service in 1998, RATP actually started drafting plans to convert lines 1, 4, 6, 7, 9 and 13.

The Skytrain in Vancouver does not have human conductors either.

Although I would assume there are conductors in a control centre somewhere?

There are no conductors, there's a control room with system overview and links to metros, stations and service personnel.

See http://www.ansaldo-sts.com/sites/ansaldosts/files/brochures/... page 21.

Interestingly enough, all transit staff are trained to be conductors in cases when the automated system doesn't work. It recently snowed which incorrectly triggered the "something is on the track" sensors causing the entire system to stop. They just took all the people normally doing other tasks and made them manually drive the trains for a day until the snow melted.

I didn't know that, that's pretty damn awesome and excellent foresight.

Never underestimate the power of unions. They can make or break a technology. Automated driver less subway cars create a precedent which can make unions feel threatened.

Of course the reality of it is more complicated, but they are a powerful force (for better or worse).

Nothing. Viennas subway has semi autonomous trains since 1960 or so. The driver just presses a “leave station” button after verifying nobody is stuck in the doors.

Stockholm metro has this as well but I wouldn't call it "autonomous", it's just a departure button to get the train moving faster. The driver still has to grip the vigilance device after a certain amount of time.

According to https://en.wikipedia.org/wiki/Vienna_U-Bahn#Fifth_expansion_... it seems that vienna isn't driverless yet.

In Vienna it's fully automated between stations. The main reason it's not fully automated is because there are no automatically closing doors in the stations. U5 is going to get it though: https://www.youtube.com/watch?v=VBDqq7rwLE0

You can quite often see the trains operate fully automatic though when the drivers are changing shifts at the end of the line. The driver typically leaves and locks the train, lets it reverse itself and then the next driver continues: https://www.youtube.com/watch?v=0eyjqS9lvNA


There are countries with way stronger unions and automated trains. I doubt unions are the deciding factor.

In some European cities (Munich?) autonomous trains often have "drivers" who do nothing because that preferred to fighting the union.

Also, as far as I can tell US public sector unions are as strong as those in any other rich country.

> In some European cities (Munich?) autonomous trains often have "drivers" who do nothing because that preferred to fighting the union.

Many cities with autonomous trains indeed have drivers. But they don't do nothing, they sit there for emergencies and to set the go signal when leaving the station. In Austria for instance that's not because of unions but because the job of the driver is at this point considered important. For fully autonomous operation you need extra security features on the track which were not employed (fully sealed off track in stations, better emergency corridors, more reliable remote door controls etc.).

Where fully automated trains are in operation there are never any drivers.

> For fully autonomous operation you need extra security features on the track which were not employed (fully sealed off track in stations, better emergency corridors, more reliable remote door controls etc.).

FWIW Lyon's Line D is fully automated (GOA4) and doesn't use enclosed track. It may have changed since but it used to not even have turnstiles, you could walk up to the track without any pause.

The only incident I can remember is a drunk who literally fell on a train from one of the elevated passages over the track.

Union culture can be significant though. From my understanding, in the US I'd expect a union to refuse any form of automation indefinitely; in other countries, they might accept it if the redundant staff are provided training for another job with good employment prospects.

> Union culture can be significant though.

Union culture is as antagonistic in France as in the US (for the same reasons that corporations are antagonistic) and it not only has multiple GOA4 metro systems running right now across the country[0], Paris is actually converting existing lines to full automation.

All of Germany (which has a much more cooperative union culture) has two GOA4 lines both in Nuremberg, meanwhile Copenhagen Metro (in highly unionised Denmark) was created fully automated back in 2002.

I see no evidence that unions have anything to do with it.

[0] quite literally: Lille, Toulouse, Lyon, Rennes, Paris

I'd be surprised if the unions representing the train drivers in France were just all "ok, cool, automate away and fire the drivers when you're done". Did they require that drivers be employed in other positions, or receive some sort of training for different jobs?

Unions in the US are annoyingly different: most (at least those that I'm even passingly familiar with) seem to have one major goal: keeping the status quo (with regular pay and benefits increases for its members, of course). They generally do not go for "hey, we're going to eliminate your jobs, but we'll compensate you in such a way that you'll continue to be gainfully employed elsewhere".

> I'd be surprised if the unions representing the train drivers in France were just all "ok, cool, automate away and fire the drivers when you're done". Did they require that drivers be employed in other positions, or receive some sort of training for different jobs?

Often that topic does not even come up because companies are not firing people to begin with like they do in the US. They just transition into a new role (for instance they could become light rail drivers in the same network where automation is not yet achievable).

Interesting. Is there that much slack in employment to cover that? I mean, if you automate a transit line, presumably you're displacing dozens of now-former train operators. It would surprise me if they always have productive, useful jobs to move people to in these situations.

AFAIK all the Paris lines moving to GoA4 have been done with no compulsory redundancies.

DC Metro use to but then there was a major accident that resulted in them reversing the system. They plan to eventually make them driverless in the future.

> considering we are almost gonna have autonomous cars

10-20 years from now isn't exactly "almost"


Because of jobs?

Perhaps unions?

I spent 5 years taking the Toronto subway (Yonge line, North York Center to Union) every day. I noticed this interesting pattern.

During rush hour, any small delay on one train will almost certain impact every train down the line- there's little time buffer between trains. The bigger the delay, the more trains effected by it. The more passengers per train, the more likely that train will have a delay- loading and unloading taking too long, sure, that can cause a small delay. But consider events like heart attacks, seizures, a fight breaking out- all kinds of major-delay-causing-events that are roughly speaking a linear function from 'time passenger is on the train' to 'likelihood of major delay event'.

If you have twice as many passengers, you have twice the odds of a major delay. If a passenger spends twice a long on a train, you have twice the odds of them causing a major a delay. Delays cause more passengers per train and cause longer time-on-train for each passenger.

It's all non-linear. Any one tiny delay can spark a total breakdown and the longer that delay is the more likely it will cause more in turn.

How on earth does the system ever work at all?

How on earth does the system ever work at all?

By promising a schedule that doesn't assume nothing goes wrong. Time padding means the inevitable minor delays can be absorbed quickly, and infrequent longer delays are recoverable.

Also, I don't know about Toronto, but most of the big delays in Boston are issues with equipment rather than passengers. Derailment, signal failure, etc. should scale with vehicle traffic rather than vehicle occupancy/passenger count, and big delays decrease how much traffic is going around at the moment.

Single shot timers? Exclusion locks on fixed sections of the track?

Modern control systems such as ETCS3 can allocate virtual track segments (moving blocks) to individual trains and adjust the reserved length based on the current speed.

Far easier to install in greenfield (fresh) projects.

ETCS is designed for incremental rollout and upgrade of the levels. And they require less equipment on the track too since things are mostly done wirelessly, level 2 doesn't require signals, level 3 doesn't even require balises.

But the 2014 study — the first time the authority had attempted to analyze the impact of any of the revamped signals, using its improved data system — found 2,851 lost total passenger hours per weekday could be attributed to thirteen modified signals alone.

That's 84 person-years per year, just considering the weekday impact of those thirteen signals. Over a lifetime a year, to save (speculatively) one life per decade? Call me callous, but I don't think it's worth it.

In Minneapolis, the mob bought and dismantled a fully functioning streetcar system.

Same forces seem alive and well today. They just chose to invest hundreds of millions of dollars into building a new stadium.

I am sure those forces still exist, but the will to fix the subways in NYC and expand it in other cities (LA especially) is stronger than ever. White flight is either reversing or wants to reverse, and a group with increasing power wants public transportation.

I wish we built functioning public transportation because its the right thing to do, I'll accept it happening because young professionals demand it.

I think the topic of a declining/slowly collapsing society needs serious research. How do you definitively measure if a society is progressing or regressing? If you measure by infrastructure the US probably hit its peak in the 60s and has been declining since then.


Step one is defining a target condition. The problem is that people have different target conditions. The enormous rise in inequality since the 1960s [1] makes clear that some people have a target condition of "I personally own everything".

[1] http://economistsview.typepad.com/.a/6a00d83451b33869e201a51...

"Progress" seems subjective and political. I posit no definitive measure of progress exists.

its pretty objective if you measure it in terms of infrastructure.

- rate of new miles of roads - rate of increase/decrease in potholes per 10 miles - number of bridges that need fixing - number of people on government assistance - percentage of population in jail - peoples test scores on standardized international tests

its not like things are unmeasureable... just that politics awards those who ignore these things

Not objective at all. There are a lot of values and politics embedded in the things you chose and the implicit direction that is healthy for infrastructure. And it should be, because “objective” metrics are inevitable gamed and their value is gone.

they must have installed one of these single-shot timers on the uptown A between 50th and 59th. it always crawls through 50th, stops in the tunnel, then crawls into 59th. didn't used to until a couple years ago.

Does anyone know of a source that competently compares costs across different metro systems? Total operating cost per mile, per passenger-mile, et cetera.

Guys like this will be running the self driving automobile infrastructure in 50 years.

the abysmal state of their technology is completely to blame. problems described point to an inability to monitor trains or operators in ways that don't create massive slowdowns as knock-on effects.

wouldn't trains be the easiest transport system to add self driving to? all it would need is throttle, and if they have access to the position of other trains too that's a major advantage.

could it be that more time spent inside a train increases profit? maybe they just want passengers looking at advertising. they are a captive audience.

I live in NYC. For a couple of years I'd be reading NYTimes articles about the system is having problems from overcrowding, yet you'd see articles that more people took subways in the 1950's - 1960's.

Within the past few months there was a NYTimes article that stated that after an accident in the mid 1990s (?) that they slowed the system down. Thus, the problem wasn't overcrowding, but slowing the system down.

The system has been underfunded for maintenance. When the city went broke in the 1970s (?), the financing was transferred to the state from the city. NY State taxes the city but does not returned the taxed funds to the city for the MTA. Transferring management of the MTA back to the city would help with holding the Mayor accountable, something to think about if they want to be re-elected.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact