>The L train is also the only line that uses modern >communications-based train control (CBTC), which allows >trains to operate in a more automated fashion.
It's not "more automated fashion", it's fully automated, LIDAR and all - they are fully autonomous. L trains have operators because of unions throwing a fit and most of the times they sit there twiddling their thumbs.
I feel it's worth reminding that there's a difference between autonomous trains, driverless trains and unattended trains.
The L train is considered Grade of Automation 2 - it starts and stops itself, though it has someone in the driver's seat who controls the doors and gives control to the ATO system. Driverless (GoA3) trains move that role to a train attendant who controls the doors from within the train. Unattended (GoA4) trains function without needing an on-board operator at all.
The driver probably is just sitting with a hand near the emergency stop button most of the time, but they're still required to pay attention. I imagine it's harder work now that they don't have to necessarily concentrate to operate the train.
I live in London, where a lot of tube lines are already (or soon to be) GoA2, with the introduction of a part-GoA2 mainline service this year (Thameslink). I don't think we'll see a GoA 4 tube for many years; aside from the obvious union rows, you've got safety concerns (primarily platform-edge-doors, which can't be retrofitted to cramped stations), and there's mixed public opinion on the idea of being stuck on a broken down train in a tunnel without on-train staff. Then again, people don't seem to acknowledge that the airport terminal transfers are unattended GoA4.
> I don't think we'll see a GoA 4 tube for many years
Perhaps I'm missing something, but I used to ride a fully automatic (e.g. there was no driver, not even a cabin for one) subway line in Lyon back in 1994/1995.
Newer automated systems have the benefit of foresight, and usually contain walkways the length of a tunnel with evacuation cross passages.
In a legacy system like New York or London where these facilities don't exist and introducing them would be very disruptive, if not completely impossible, you probably need at least one person on a train to assist with evacuation.
Yes, each DLR train has a Passenger Service Agent, who is fully trained in the safe driving of a DLR train by a human, but during normal operation they're walking around in the train, closing the doors when it's ready to leave, and so on.
The PSA will drive the first train on each route every morning, since unlike the train's automation their human eyesight allows them to detect some very dramatic mistakes, for example if maintenance engineers left a metal trolley full of equipment halfway across the line, the train can't see that but the PSA can hit the "Stop" override. Normally they don't drive, allowing passengers to sit at the very front of the train, like on a roller-coaster, the manual controls are under a lockable panel.
The train also has a "full" manual mode in which the PSA is able to explicitly drive it wherever they want, rather than just stopping unexpectedly if they see a problem, but in this mode the train deliberately cannot reach its normal speed, because humans have poor reactions and so it would be too dangerous without the machine supervising.
The video is from 1988 (good find!), but the only accidents ever to have occurred were under manual override.
The Copenhagen Metro is an example of a level 4 underground system. It's fairly new, so the tunnels all have an adjacent walkway which can be used in an emergency.
I don't mind paying a man to sit at the head of the train even if they're autonomous. I also don't believe making all the lines CBTC are going to solve the core issues of breakdowns outside of an ancient signaling system.
Some of the lines in NYC are still using R32 cars[1]. They were built in 1965 that is ludicrous. Granted, they were refurbished in the late 80s but that's still 30 years ago.
That said most of the lines are running "newer" stock (either from the late 80s[2] or mid-2000s[3]) but the focus on signaling, while at the core of many delays, is a band-aid on a system that is fundamentally broken by way of funding EVERYTHING appropriately. While I do appreciate the wifi in stations I completely don't care about it if the train doesn't arrive. While I do appreciate the "wait time" clocks they're meaningless when they say "5 minutes" for 15 minutes then switch to DELAY. You might as well not have a sign at all.
There are fundamental bureaucratic issues going on causing many of these problems and technology is not a root solution. It's a patch. You won't be able to argue with me that one of the busiest transport system in one of the richest cities on the planet cannot manage to fund serious upgrades.
Also wary of politicians blaming the unions like they're the cause of everything. This is an ongoing tactic that makes sure that nothing continues to happen. They aren't the problem directly even if there's a lot of dead weight. They're, again, a symptom of higher level management problems.
As a person living in Osaka, Japan right now, I always find it funny reading people comment on NYC Subway. Not your specifically, but in general.
In Osaka, Hankyu Railway (granted, a commuter train operator, but most if not all commuter trains service in Japan runs like subway) still runs multiple 3300 series, manufactured in 1969. They (the 3300's) actually also run into the Osaka Metro system, and you cannot really tell it's built almost 50 years ago.
And CTBC... Most of the newslet says that one of the problem with old system is that it cannot accurately tell where the train is... Well, I know NYC signalling system is ancient, but not being able to tell accurately where the train is IS NOT the problem. Most of the Japanese lines (including various subways) still run on block system (though sometime with a block length of just 100m). Communication-based system is in place on some line, though, but mostly for lightly-traffic line to save on signalling equipment cost.
Yea I figured this may be mentioned, it's the same in the UK they were running very old train stock of the same vintage until (I think) this year? I know it's also not the only problem but both the Japanese and UK cartridges aren't complete shit and falling apart.
Three times in the past 2 years (since I moved to the G line) the train had to be evacuated because the brakes locked on a car and someone smelled the awful burning it thinking it was an actual fire in the car. THREE TIMES.
If you maintain the cartridges there's no problem at all but there are so many problems it's way beyond that.
I'm a little of a train nerd (if not obvious) so I went to the train stock as an example but other underground systems prove that proper, well funded and well managed maintenance will make this a non-issue.
"Proper. Well Funded. Well Managed."
- Not the MTA Subway System
London's oldest trains are from 1972, the 1969 stuff was withdrawn by the end of 2014.
But England runs older trains, and this has recently been newsworthy [1]. I think the underlying problem is the electrification of two lines has been delayed, so the newer diesel trains used on those lines are still needed, and can't (as originally planned) be used to increase capacity on the other route.
It also shows that London's transport, which is directly managed by a government body, works far better than the privatized mess in the rest of the country.
It's not a question of whether "one of the richest cities on the planet" can "fund serious upgrades." It's how much money are New Yorkers willing to throw into a system where it takes several times as much money to do the same thing as other countries? https://www.nytimes.com/2017/12/28/nyregion/new-york-subway-....
New York is rich, but so is London. Based on that wealth, there is a certain amount of money New Yorkers are willing to pay for subway service. If it costs twice as much to do the same thing, half as much service will get purchased. (Indeed, it's worse. Crappier subway service reduces demand, which decreases the money available to fund it.)
Costs are the fundamental problem with the MTA (and D.C.'s WMATA), and those are union problems, not management problems. If it cost Lenovo twice as much to build the same laptop as Acer, Lenovo would simply cease to exist. Transit systems generally cannot go bankrupt for political reasons, but they can decline to a state of government-funded life support where nobody uses them except people who have no other choice. (See, e.g., the transit systems in most cities outside NYC/Chicago/DC/Boston). Consider:
> France’s unions are powerful, but Mr. Probst said they did not control project staffing. Isabelle Brochard of RATP, a state-owned company that operates the Paris Metro and is coordinating the Line 14 project, estimated there were 200 total workers on the job, each earning $60 per hour. The Second Avenue subway project employed about 700 workers, many making double that (although that included health insurance).
This is a large-scale problem in the U.S. Our public services suck, which means that they turn into safety nets instead of something that are broadly used by the population. In turn, people have limited willingness to fund them (because people naturally are less willing to spend money on safety nets versus something they also use and benefit from). Unions aren't the only reason for our public services sucking, but to the extent they drive costs out of alignment with what is the case in other countries, they're a big part of the problem. If you can buy less service with the same amount of investment, that's a problem.
One of the things that's happened in the last 30 years that nobody talks about is that Europe became far more market oriented, and their unions adapted. Anti-union rhetoric in the U.S. yielded a very different result, with most private unions dying out, and the public sector unions that survived remaining a bulwark of the "old way" of doing things.
It's worth noting that the problem in the article is not that the MTA negotiates poor contracts with the construction unions, it's that they don't negotiate with them at all.
The contractors that do the negotiation are more than happy to oblige the union's demands, as long as their margins are fat, and so far the MTA has not cried uncle about the costs being passed on, although that looks to be (very slowly) changing.
Rich people buy bottled water (and our water system is a disaster that’s poisoning poor kids with lead). Rich people either send their kids to private schools, or carve out enclaves with public schools that have just other rich people (while the “safety net” inner city schools deteriorate). Electricity isn’t a “public service.” For the most part, private companies provide electric service in the US using private infrastructure. Nobody uses libraries.
That leaves our roads, which are much crappier than in say Germany or Japan.
2M library card holders, which represents about 10% of the number of people who are eligible for a library card, and fewer than 1/2 checked out a book.
So 5% of the total eligible population used one.
Don't get me wrong, I love libraries and I go every week with the kids, but I think OP is right -- few people actually use the library.
> So 5% of the total eligible population used one.
That's not a good metric to use because everyone who lives or works in New York State is eligible to join, but for most people it only really makes sense to join if you live in Manhattan, Brooklyn, or Staten Island. Otherwise you'd just join your local library.
So really ~60% of all residents (i.e. including young children) have a library card, and the overwhelming majority take advantage of library services and events.
> Rich people buy bottled water (and our water system is a disaster that’s poisoning poor kids with lead).
Rich people consume way more public water than poor people. How many poor people have huge single family homes with lawns and swimming pools? How many buy tons of consumer products, own several cars, and are buying large amounts of organic food?
And given that basic filtration removes 99.9% of lead, you'd die of hypernatremia long before it would be possible to get lead poisoning from drinking filtered tap water.
I drove around Munich for a week, then from there through Austria for a couple weeks and back into Bavaria.
The roads are good. The autobahns are probably better than our interstates --- though not by that much. Traffic congestion was as much a problem there as in the American midwest. Their surface streets are on average not as good as Chicago metro area surface streets (but are probably better than Baltimore's). And I feel a good deal safer on US rural highways than I did on German and Austrian rural highways, which are beautiful, well-maintained, nightmarish death traps.
I think you'd be setting yourself up for a pretty tough argument if you wanted to claim that the US doesn't do public roads well. Do we do them better than Germany? No. But we do public roads anomalously well, especially for the degree of difficulty involved in providing them across a whole continent.
I'll concede that I don't have data to back up my claim, and that anecdotally roads might be the area of infrastructure where we're least bad compared to Western Europe and Japan.
>Costs are the fundamental problem with the MTA (and D.C.'s WMATA), and those are union problems, not management problems.
Managerial corruption causing cost overruns like this does not even tangentially involve unions:
"The New York City subway system has continued to deteriorate, rankled by massive delays, misspent funds, and widespread claims of corruption. On Friday, the feds confirmed the corruption part by securing a 46 month sentence for a former employee who promised subway contractors future work in return for bribes."
What is the acale of that in comparison to paying more worksrs than you need more money than you need to? Most of MTA’s expenses are labor costs. Contracting corruption is bad and illegal, but is not what’s driving the costs.
> That's certainly MTA management's narrative when asked by, say, the New York Times.
If you're saying that the MTA is trying to misrepresent their role in their own dysfunction when speaking to the news in order to generate coverage that makes them look more favorable, then the MTA should fire their whole PR team, because they're clearly incompetent.
It's pretty clear to anybody who's following the issue that the MTA and TWU have an unethically incestuous relationship, in which the TWU and MTA essentially collude to extract as much money from the state as they can while literally not even doing the bare minimum to keep things running.
The relationship between the MTA management and TWU is fractious at best.
The relationship between MTA management and the scores of construction companies that curiously increased their costs 50% in recent years is curiously pretty good, but I'm sure that had nothing to do with the increased costs.
Of course tried to spin a corruption problem as a union problem. Are they really going to blame themselves?
As the NYT article explains, the companies’ construction costs are driven by labor costs. The construction companies negotiate with the unions, and pass the higher labor costs onto the MTA, which does not negotiate directly with the unions.
Construction is a labor-intensive industry, and labor costs dominate. If you look up the public companies that do civil construction work, you’ll see that the profit margins are razor thin (2-5%). These are not companies like Facebook making 20-30% profit margins.
This is not a moral issue. It's not about evil construction company CEOs. Reducing executive pay at these companies wouldn't amount to a drop in the bucket for these multi-billion projects. Nor is it about evil union workers. We're just talking about hard-working folks who are getting paid a rather more comfortable wage than their counterparts in Europe. Which is good for them, but wrecks the economics of the MTA, an entity where payroll, health, and pension make up more than half of expenditures.
Both of those things can be true simultaneously. Anyone insisting on a single, isolated cause of badness in such a complex system is pushing an agenda.
1. Each employee on a train costs well over $100k/year: going from 2 employees per train (driver & conductor) to 1 would save on the order of $200+ million dollars a year.
2. Wifi in stations is entirely paid for by the advertising portal, so that upgrade was "free" for the MTA.
Dismissing Unions assumes you really know all the reasons Unions might object. You go directly to 'taaaking our jooooobs' without even reflecting once on the other concerns.
BTW, if people don't have jobs what is the economy? we were told a few years back that with the decline of manufacturing the jobs were in the service industry. Service, like y'know, staff on the underground?
Maybe, the path out here is to keep them working and paid, but have them do other things?
Unions are not bad. Sloppy critique of unions because 'story' is bad.
Well, they're not always twiddling their thumbs. Every time the train comes in to the station, they have to open the window and point up at the zebra-striped sign.
That's not literally twiddling your thumbs, but I'll leave it as an exercise to the reader to determine whether that serves any purpose on a fully-automated train that couldn't be eliminated with no loss of functionality, safety, or convenience.
Well this isn't correct either. It's not fully automated in the sense that doors are still operator controlled. LIDAR is completely unrelated and is used for asset management rather than signalling.
...if the train is on happy path. And believe me, there are some really ugly unhappy paths where underground passenger rail is concerned; we definitely have no technology autonomous enough for that.
The line to Moorgate was owned by the Metropolitan Railway, which became part of the amalgamated London Passenger Transport Board, using the Underground brand, and branded as part of the Northern Line.
The line up to Holloway Junction (where it joins the ECML) was owned and maintained as part of the Underground, operated by Underground drivers and Underground owned trains, and it was only transferred near the end of 1975 to British Rail.
Perhaps most tellingly, it was London Underground 1938 Stock involved in the 1975 crash.
More pertinently, a good deal of these seems to to driver error, specifically missing signals. If there's anything automatic drivers are good at, it's not missing signals.
The author does briefly mention this, but the countdown clocks are not every accurate. I often see a wait time of "3 min" and then it's there in the next 10 seconds. Or you can see the opposite where it just hangs on "3 min" for a minute or two and was really 5 minutes away. In reality it should just report at its accuracy, e.g., "<5 min" "5-10 min", "10-15 min", etc.
PS, why do people stick their necks over the tracks to look if a train is coming?
> PS, why do people stick their necks over the tracks to look if a train is coming?
You can often see further down the tunnel from an angle close to the tracks than you can standing behind the yellow strip. If someone ahead of you on the platform is already doing this, then they're probably blocking your view down the tunnel, so you have to lean to see past them. It seems pretty silly, but sometimes it's useful, for example, in determining whether the express or local will come first, and you can then switch platforms at the last minute if it makes sense for you.
If countdown timers were reliable and accurate, fewer people would feel compelled to lean out.
Accurate countdown clocks are hard, they by definition involve predicting the future which is widely accepted to be a Hard Problem :)
More specifically there are two major problems one of which is expensive to solve, and the other is impossible to solve.
The expensive one is providing and making use of lots of sensors so that we can have the best possible evidence on which to base our prediction. Most obviously we need to sense where the train is - the more accurately the better. But it's also valuable to know if the train is moving, and even if the previous platform is full of passengers who will board it before it can leave, or empty.
The estimates may prove wrong because of something as dramatic as a fault, or as trivial as somebody's coat trapped in the door.
London's system has always seemed fine to me, but it's noticeable that buses (for which London also provides countdown clocks) have much poorer accuracy, because as well as the vagaries of passengers they must contend with variable traffic. It only takes one idiot trying to reverse a lorry onto a major street to add 3-4 minutes delay to your bus and of course if you can't see the lorry you have no idea that's where the time went.
I don't agree that it makes sense to reduce the displayed precision as you suggest. The problem isn't with precision, it's accuracy, and you can't really fix that by reducing precision.
Yeah, I've heard of one vendor needing to use both the real-time and static-time data in such a way to determine how off-schedule the trains are, and using that information to discard the unreliable real-time data bits.
Also, many days the system is outright not running at all, at least as far as displaying arrival times in the station, even on the L. This seems to be the case many weekends for some reason.
That feed archive haven't been updated in years; you have to scrape and save the feed yourself or scrounge together different sources to look back in time.
I notice the numbered trains are all coming out ahead of the lettered trains. Is this taking into account the bug in the MTA data where the countdown clocks for the lettered trains weren't accurate at the terminal stations until a week or so ago?
I found it interesting to contemplate how much harder it would be for someone to do this level of analysis and draw conclusions from it if they weren't physically proximate.
I love remote work, and I hope to sustain it for the rest of my career, but there is something to be said for direct exposure to the systems you study.
(It also makes me appreciate all the more how hard it must be for archaeologists et al to glean details of the past from fragmentary evidence of a world long gone.)
This is interesting analysis, but it's based on NYC MTA countdown clock data which is just rampantly inaccurate on lines other than the L. It is regularly, customarily, up to 4 minutes off in either direction, in my experience.
Since trains are supposed to be spaced out 4-5 min or so, that means the clock is only really useful to tell if there's a huge delay.
He's completely right that the data is messy and sometimes makes no sense at all. Even a passenger waiting for a train, if paying attention, will notice trains appear, disappear and "jump around" on the countdown clocks.
Moving due to the L tunnel shutdown next year, I've been asking if any other train comes nearly as often (I rarely wait more than 5 minutes). This is a very timely list, thanks!
It's not "more automated fashion", it's fully automated, LIDAR and all - they are fully autonomous. L trains have operators because of unions throwing a fit and most of the times they sit there twiddling their thumbs.