Can Ridership Surges Disrupt Small, Frequent Driverless Metros?
A recent discussion about the Nuremberg U-Bahn got me thinking about the issue of transfers from infrequent to frequent vehicles and how they can disrupt service. The issue is that driverless metros like Nuremberg’s rely on very high frequency on relatively small vehicles in order to maintain adequate capacity; Nuremberg has the lowest U-Bahn construction costs in Germany, and Italian cities with even smaller vehicles use the combination of short stations and very high frequency to reduce costs even further. However, all of this assumes that passengers arrive at the station evenly; an uneven surge could in theory overwhelm the system. The topic of the forum discussion was precisely this, but it left me unconvinced that such surges could be real on a driverless urban metro (as opposed to a landside airport people mover). The upshot is that there should not be obstacles to pushing the Nuremberg U-Bahn and other driverless metros to their limit on frequency and capacity, which at this point means 85-second headways as on the driverless Parisian lines.
What is the issue with infrequent-to-frequent transfers?
Whenever there is a transfer from a large, infrequent vehicle to a small, frequent one, passengers overwhelm systems that are designed around a continuous arrival rate rather than surges. Real-world examples include all of the following:
- Transfers from the New Jersey Transit commuter trains at the Newark Airport station to the AirTrain.
- Transfers from OuiGo TGVs at Marne-la-Vallée to the RER.
- In 2009, transfers from intercity CR trains at Shanghai Station to the metro.
In the last two cases, the system that is being overwhelmed is not the trains themselves, which are very long. Rather, what’s being overwhelmed is the ticket vending machines: in Shanghai the TVMs frequently broke, and with only one of three machines at the station entrance in operation, there was a 20-minute queue. A similar queue was observed at Marne-la-Vallée. Locals have reusable farecards, but non-locals would not, overwhelming the TVM.
In the first case, I think the vehicles themselves are somewhat overwhelmed on the first train that the commuter train connects to, but that is not the primary system capacity issue either. Rather, the queues at the faregates between the two systems can get long (a few minutes, never 20 minutes).
In contrast, I have never seen the transfer from the TGV to the Métro break the system at Gare de Lyon. The TGV may be unloading 1,000 passengers at once, but it takes longer for all of them to disembark than the headway between Métro trains; I’ve observed the last stragglers take 10 minutes to clear a TGV Duplex in Paris, and between that, long walking paths from the train to the Métro platforms, and multiple entrances, the TGV cannot meaningfully be a surge. Nor have I seen an airplane overwhelm a frequent train, for essentially the same reason.
What about school trips?
The forum discussion brings up two surges that limit the capacity of the Nuremberg U-Bahn: the airport, and school trips. The airport can be directly dispensed with – individual planes don’t do this at airside people movers, and don’t even do this at low-capacity landside people movers like the JFK and Newark AirTrains. But school trips are a more intriguing possibility.
What is true is that school trips routinely overwhelm buses. Students quickly learn to take the last bus that lets them make school on time: this is the morning and they don’t want to be there, so they optimize for how to stay in bed for just a little longer. Large directional commuter volumes can therefore lead to surges on buses: in Vancouver, UBC-bound buses routinely have passups in the morning rush hour, because classes start at coordinated times and everyone times themselves to the last bus that reaches campus on time.
However, the UBC passups come from a combination of factors, none of which is relevant to Nuremberg:
- They’re on buses. SkyTrain handles surges just fine.
- UBC is a large university campus tucked at the edge of the built-up area.
- UBC has modular courses, as is common at American universities, and coordinated class start and end times (on the hour three days a week, every 1.5 hours two days a week).
It is notable that Vancouver does not have any serious surges coming from school trips, even with trainsets that are shorter than those of Nuremberg (40 meters on the Canada Line and 68-80 meters on the Expo and Millennium lines, compared with 76 meters). Schools are usually sited to draw students from multiple directions, and are usually not large enough to drive much train crowding on their own. A list on Wikipedia has the number of students per Gymnasium, and they’re typically high three figures with one at 1,167, none of which is enough to overwhelm a driverless 76 meter long train. Notably, school trips do not overwhelm the New York City Subway; New York City Subway rolling stock ranges from 150 to 180 meters long rather than 76 as in Nuremberg, but then the specialized high schools go as far up as 5,800 students, and one has 3,000 and is awkwardly located in the North Bronx.
Indeed, neither Vancouver nor New York schedules its trains based on whether school is in session. Both run additional buses on school days to avoid school surges, but SkyTrain and the subway do not run additional vehicles, and in both formal planning and informal railfan lore about crowding, school trips are not considered important. So school surges are absolutely real on buses, and university surges are real everywhere, but not enough to overwhelm trains. Nuremberg should not consider itself special on this regard, and can plan its U-Bahn systems as if it does not have special surges and passengers do arrive continuously at stations.
Pedestrian Observations 
what if instead of 75 meter trains they instead have 37 meter two car trains? Apparently for a while the concept was to run two car trains every 100 seconds…
Nukebro, die Kurzzüge waren ja das Problem. Mit Langzügen hätte der 100sec vielleicht geklappt. Mit doppelt so vielen Türen können die Menschen schneller Ein/Aussteigen.
Nukebro, die Kurzzüge waren ja das Problem. Mit Langzügen hätte der 100sec vielleicht geklappt. Mit doppelt so vielen Türen können die Menschen schneller Ein/Aussteigen.
That works on weaker lines or in smaller cities. Brescia does this – the trains are 40 m long, because it’s a small city and doesn’t need more. Vancouver has 40 m trains on the Canada Line but there are serious worries about long-term capacity (that said, the airport is on a Canada Line branch and isn’t the problem – in fact the airport branch is the weaker of the two). Vancouver relies on high-density zoning to support SkyTrain ridership, and the Canada Line took a while to be upzoned and is still much less developed than the Expo Line.
The problems caused by the limited capacities of ticket vending machines and fare gates have of course obvious solutions. For example, TGV tickets could come with a free trip on the Paris Metro/RER included.
In Paris it’s an issue of the fare media needed to go through the fare gate. The fare gate can’t read a QR code from an electronic ticket, only paper tickets and the Navigo card.
If you solved the fare gate issue, you wouldn’t even need an integrated ticket, you can sell them online separately. Once I bought a Brussels-Zurich trip (via Paris) in person, from the salesperson at the station and they offered to sell me the Paris metro/RER ticket too for the connection between Gare du Nord and Gare de Lyon. They could do this because they could just hand me the paper ticket.
I think Eurostar used to sell local transit tickets from their on-board bar?
I experienced a long queue to buy RER tickets when arriving at Massy by OuiGo in 2017, but nowadays it’s possible to make your phone act as a Navigo card so I expect this is much less of a problem today.
RENFE AVE includes a Cercanías ticket, but it requires you to go to the booth to buy the ticket with the code you’re given and find the option, very few people ends up doing it I think. In theory this could be solved with an integrated app that allowed NFC tap, but RENFE seems to have a bad relationship with IT in general
When I used this I just went to a ticket machine, I did not need to go to a ticket booth. And nowadays at the major stations they have QR scanners at the gates. So it has improved.
But RENFE still does not know how to run a proper railway…
Yes, booth/machine my linguistical mistake here. But wait, QR code… How do you get the QR code? Last time I tried the only way I found was buying the single ticket. Is there another way?
Both Eurostar and the TGV Bistro provide this. It’s a bit more expensive and only provides a basic ticket. Many want other tickets (or even personalized passes required for weekly tickets) and aren’t willing to buy an “unnecesary” ticket.
The benefit is avoiding the queues at the ticket vending machines. Might be worth it if you’re in a hurry or tired.
Also the navigo pass is available digitally at least on iPhone like Suica. That works well.
You do not need to solve the fare gate issue. Just have a manned gate at transfer stations. London solved this ages ago.
Or you can just do away with fare gates. People arriving in eg. Berlin can be on their way on the S or U Bahn within minutes of alighting from their train.
Solar system brain: do away with fare gates
Galaxy brain: do away with fare gates and the need for a local ticket
On a visit in 2000 to Lille, which has small, driverless metro trains, I saw a large queue at a station, filling the platform and reaching up the stairs. I don’t know what caused it, but I was amazed as several cars came in close succession, each of them taking 20 or 30 people, and in no time the station was empty again.
The biggest surges I can think for a transit system are people leaving a stadium. The tactics used to deal with stadiums are usually: have the transit at some walking distance, so that people arrive a bit less suddenly (like the famous Olympic Way in Wembley), have multiple transit options, or have transit options that start at the stop that serves the stadium (Munich, Brussels) so you can accumulate waiting vehicles ready to start boarding people.
I wonder how Madrid does with the L10 stop next to Santiago Bernabéu, but I suspect it may be similar to Camp Nou, where a lot of people walks the extra mile (literally) to reach another option that is more direct and comfortable.
At some point Barcelona will open the driverless L9 Camp Nou stop, which will be the closest to the stadium, and I expect that the bottleneck will be the elevators instead of the trains (if you want to use Google translate, there’s some interesting figures here: https://x.com/volemL9/status/1783186027412119704)
26 meter long VAL trains are good for about 10 000 (VAL 206) or 12600 (VAL 208) passengers per hour and direction.
Lille spreads the crowds between Hotel de Ville and Stade P Mauroy stations on L1, and Les Pres on L2 reached by a 5 km bus shuttle. For the 50000 seat arena, the free buses run for 1 hour after big games and concerts. With the same trains, the shuttle toward Barriere de Paris Toulouse metro B station runs for 1 hour after games at Ernest Wallon. This rugby stadium has a capacity of 19500 and will have its dedicated station on line C. Toulouse stadium, capacity 33000, is 100 m from a tram stop and 10 minutes walking distance from two metro B stops.
Turin VAL trains are 52 m long and its Juventus/Allianz Stadium has 41500 seats. The bus shuttle to Bernini metro station only runs for 45′ after the games.
Automated metros schedule is flexible. More trains are injected by pushing a few buttons. This is easier than scheduling drivers for a short surge of uncertain timing as concerts can run late and football go in extra time. Capacity may still be an issue. Milan M5 last station is only 6 minutes from San Siro stadium. The arena can welcome 78 000 during concerts but the 50m long trains can only take at most 24 000 in one hour.
The weakest part of the UK rail network is the inability to react well to issues. I am not sure driverless metros fix that.
I do agree with you both that the biggest surge in general is stadiums or large music events. Though there are other things like the Bath Christmas market that fall into the same camp.
Setting up the queues accordingly is important.
Driverless metro can react well to some issues. If a station is getting crowded (which we can sense) we can add a few more trains to the line. In the case of sports you will have no problem finding staff who will watch the game as part of their job just so they can hit a “game ending soon” button so the system can adjust. Or just run enough trains to handle the surge all the time – 100k people leaving a game isn’t much different from all the people who leave work downtown at 5pm, and the additional service will be enjoyed by everyone who doesn’t have to wait.
Of course if a switch breaks driver and driver less cannot do anything about that. (maybe the driver can get out and hit the broken part to get it to move, but that is a maybe at best and even if it works I wouldn’t want to be on the train)
I really don’t think it is that sophisticated.
I do think it is sophisticated. I’ve been in subway after 100k stadium lots of times, in million people demonstration, and downtown 5pm. For the downtown thing to be equivalent to the others, it should be a really really centralized metropolis, though. Stadium should theoretically be the ultimate stress test, but it’s something you can plan around, you know the exact number of people, you see on TV the moment when they’re told to start moving to the transport, and if you don’t get it right one week, you can fine tune it and try again in one or two weeks. I do have the sensation that they put extra trains even in the automated lines. A million people event on the street looks a harder problem, because it’s a one-off, so the people behavior is more unpredictable, the finish time is often not so sudden, so you don’t know when and where to run the extra trains, and it’s easy to miss details like how many turnstiles should be open to enter or leave at each moment and end up closing a station because there’s risk of somebody falling on the tracks. In general, though, for stadiums or huge demonstrations, I’ve always felt that people are aware that this is an exceptional situation, and take it easy, walk looking for alternative routes, and if it’s an one-off thing, they can just decide to linger longer at the place because there wasn’t a plan to hurry back home anyway.
What we CAN do and what we SHOULD do are often different. It isn’t technically hard to do the things I stated. However that doesn’t mean we should, it means it is a tool that could be used if needed. However before you pull that lever ask if you should.
You should be able to run a train every 90 seconds all day (that is the most frequent I’m aware of anyone running a train so I will put it as what your goals should be, but I suspect it would be possible to get down to a train every 30 seconds if you want to put $$$ into it!). That will handle a lot of people without any other effort on your part. This should be the first lever you pull anytime there are capacity problems since high frequency trains are so much better for the common rider in off peak times – nobody misses a high frequency train and nobody complains about the wait.
While you can put more trains on the line just for events, this is rarely useful. Most cities put their stations near downtown, and so you need capacity on other parts of the line for rush hour and even off peak there is still a need for the trains. Putting more trains on a line for events also begs the question of where do they come from: why store them in a yard when you can just run them (see previous paragraph) and save the cost of a storage yard. You can take trains from a different line, but that means people not at the event get worse service which isn’t a polite thing to do – typically 95% or more of the population is not attending the event and they need to do something.
As such the only thing I think you should normally do is schedule line maintenance to not happen until events are over.
At high load stations are likely a bottleneck – this can be used to your advantage (the platforms are not overcrowded if your gates limit how fast people can get to the platforms so consider this as a safety feature in station plans!). However you need to get as many people through the stations as possible, and that means stations near large events need to be designed for the event even if they rarely see that surge.
@henrymiller74, the Jubilee line in London does 24tph off peak until 11pm and 30tph peak. Would it make sense to do 30tph between North Greenwich and West Hampstead after an event at the O2? Maybe, but you are making the service more complex.
The tube is the only metro outside Japan that makes a profit so I am sure the ultra high off peak schedule makes sense commercially.
As per https://content.tfl.gov.uk/wtt-18-jub-16-may-2022.pdf Looks like they have one spare train on the Jubilee line for Wembley events.
one of the things the U9 will do in Munich is add an additional station at Theresienwiese which is only busy during Oktoberfest. Whether or not that’s wise is for others to judge…
Just noticed a typo – Cities tend to be Stadiums near downtown. Stations as I wrote also go near downtown, but that is obvious. Stadiums near downtown mean that your event lines are also useful for normal traffic. Stadiums not near downtown (which are also common!) are much harder for transit to serve as there are not a lot of lines in the area anyway. (at best you have a radial and a circular line near, but often just highways).
A station should be cheap enough to build so lone as we assume the line is already going there, so a station just for one event shouldn’t be a big deal. Of course should is doing a lot of heavy lifting here. Alon has found stations are a large part of the cost explosion in many parts of the world. I’m not sure how Munich and cost explosion is playing out so I will withhold judgement, but if they are keeping costs reasonable a station just for Oktoberfest seems wise.
A reminder — the 2026 FIFA World Cup will have its final game at MetLife Stadium in East Rutherford, NJ. When Beyonce had her COWBOY CARTER concert series, special trains (called the BetMGM Meadowlands) from Secaucus Junction, and special buses, were run
They run event trains whenever there is an event. They even run trains allllllllllllllll the way from Connecticut when there are events they can service. It will likely be a flustercuck like the Superbowl was.
https://en.wikipedia.org/wiki/Mass_Transit_Super_Bowl
Crowds at any US venue of the 2026 World Cup? Not likely to be an issue.
https://de.m.wikipedia.org/wiki/U9-Spange_(U-Bahn_M%C3%BCnchen)
This speaks of four billion euros and over half a billion just for the station under Hauptbahnhof. But it doesn’t really say how many km of new line it’ll be. Overall it seems quite pricey for a line in non-anglophone Europe but it is in the heart of the city and has to dig under existing lines quite a bit…
Yes… and that doesn’t mean you should do it all day. It’s expensive to run so frequent. In my city in peak hour, interval between trains is around 2:30 – 3:00. Rest of the day it’s around 3:00 to 4:00. Late at night the frequence drops more. Not even Moscow keeps the super high frequencies all day. But, when you have an event, that’s another peak (even late at night), so you go back to 2:30 – 3:00 interval for a lap. In some cases they can even delay the closing time of the subway. Since depots aren’t exactly at the end of the line, and you need workers in special working timetables, all this requires some organizing and prevision.
they could run them 24 hours a day. Many railfans forget that the point of passenger trains is passengers and get all frothy over runnnnnnnnnnnnnnnning trains.
@jordi the subject here is automated metros. if you have the train the cost to run it is minimal. There is a little more wear and cleaning, but the majority of the costs are fixed. Driver for non automated train mean running an empty train is too expensive, but the ecconomics of automated trains are different. You should be running peak service all day because high frequency almost always gets enough additional riders as to make up the costs of running those trains.
the only reason not to run that service 24×7 is you need to reserve some time for track maintenance.
@henry, clearly TfL don’t think the cost of the driver is particularly high given they run the full off peak schedule at up to 11pm on a Tuesday 😛
Crowds at any US venue of the 2026 World Cup? Not likely to be an issue.
Nonsense. There is enormous demand to attend matches domestically, both from immigrants from Central and South America and from the grown up children from those “soccer moms” that were so popular in the news a few decades ago. The World Cup with the highest ever attendance records (both highest overall and highest average per game) was held in . . . the United States, and that was in 1994 when soccer was a lot less popular than it is now.
As to your sensationalistic article, the total number of Swiss citizens denied entry into the US this year is . . . two. Not two thousand, or two hundred . . . just two. My source: the article you posted. 31.8M people entered the US internationally last month – no one is coming here!
@henry, clearly TfL don’t think the cost of the driver is particularly high given they run the full off peak schedule at up to 11pm on a Tuesday
Clearly they do think the cost of the driver is an issue because they don’t run the full PEAK schedule that late. With driverless, you could run the peak schedule until 9pm, say, and the full off peak schedule until 1am, or whatever you need.
Note that despite what many boosters of driverless metros say, the driver is not the only cost to running a train, and you cannot exactly run the same service you would at peak all day for free. Maintenance costs scale with distance travelled, hours of operation, or cycles performed. If you double those because formerly peak hour trains that were in service 10 hours a day are now in service 20 hours a day, your maintenance cost will double. Every driverless metro I am aware of still reduces services at least during late night hours if not also off peak, as it doesn’t make sense to incur maintenance costs on a train that could carry 1000 people if it only has 3 passengers on board.
I am not actually sure you could run the full peak service even if you wanted to that late.
The off peak riders won’t be as efficient at getting on and off the trains for starters.
That may be why the O2 doesn’t get better service.
And 24tph on the Jubilee line until 11pm is pretty strong
The soccer stadium a few miles away in Harrison, sells out fairly frequently. There will enough demand.
https://en.wikipedia.org/wiki/Sports_Illustrated_Stadium
The owners of the stadium and mall don’t want to spend money to upgrade the train station and it will likely be a flustercuck like the SuperBowl was. Cue clueless transit fans that NJTransit should have had 1,000 buses and drivers standing by just in case.
If you’re even vaguely Hispanic-appearing, you’d be insane to make yourself a target (ie appear in public) anywhere near a mass concentration of “law enforcement” stormtroopers.
Reminder too that the US-hosted bullshit made-up “Club World Cup” of the last month had such shit sales that they pretty much had to pay people to attend.
Unless you have urgent fascist business to conduct, the least dangerous course is stay well the fuck away from the fascists. Don’t visit the USA. It’s not safe. You have less than no rights. You can be fucked over arbitrarily, you can be disappeared, you can be deported anywhere (South Sudan? Guatemala? Heard and McDonald islands? Whatever! The US Supreme Court is just fine with any of this. Good luck with having your embassy summon their ambassador!)
You have better things to do with your life and with your money than support fascism.
Stay away from the USA.
@Onux 3 passangers per trains at 45 tph = 135 passengers per hour. That is one direction, so 270 for the line. You can find infrequent buses in mixed traffic that does more than that (most do not, but you can find them). And as an infrequent line, that doesn’t even get the bonus that frequent service gets. Which is to say if your only get that many people on your trains you shouldn’t be building a train in the first place, the capital costs are too high for the ridership.
You are correct that running trains isn’t free. I said the cost is minimal, but it isn’t zero. You will need twice as much maintenance staff and such – but you have much less of them than train drivers. I always assume that you greatly reduce service most nights for track maintenance purposes (this appears not as true as it should be – and when things wear out they will pay a price), but otherwise running frequent service should be your goal.
I always advocate for frequent service because I have missed my bus in the past, and that long wait for the next was enough to make me want to drive. That is the point is to serve people first! Cost is a required budget item that you must watch of course, but if there is money run more service. With automated metros I believe that running more service will bring in enough more riders to pay for the marginal additional costs – remembering the cost of the line itself is a sunk cost.
@Onux 3 passangers per trains at 45 tph = 135 passengers per hour. That is one direction, so 270 for the line. . . . Which is to say if your only get that many people on your trains you shouldn’t be building a train in the first place, the capital costs are too high for the ridership.
@henrymiller, my analysis was in reference to your comment that the cost of running driverless trains is minimal, so you should run them as frequently as possible all of the time, and relating to late night service. The line that gets 270pph at 3AM could be getting 27,000pph at rush hour, so you absolutely should build the line. But if you agree that 3 pax/train isn’t high enough ridership, then you agree that you shouldn’t be running 45tph in the off hours, even if you can, because the cost is too high for the benefit. 15 or 10 or even 5tph would be outstanding service for very late at night, and would all be fast enough that people wouldn’t feel bad with a long wait after missing one train.
if you have the train the cost to run it is minimal. There is a little more wear and cleaning, but the majority of the costs are fixed.
The key point is that the cost to run additional driverless service isn’t really minimal, from one perspective. If you run the train twice as much the operations and maintenance cost is double. The lack of driver salary means that operations cost is much lower than driven trains, meaning you can run them more for a given budget, but it doesn’t mean you don’t pay more when you run them more than that. The high fixed cost means that overall cost per hour or cost per km goes down as you run a train more, but that accounting fact doesn’t mean that if you spend $10M on maintenance this year that you will have $20M for maintenance next year when you run everything twice as often. Budgeting is as much about cash on hand as it is about total cost of ownership.
High capital cost can also work against running more frequently. Trains have a certain lifespan based on km travelled or hours operated. If you run your service more often, you will wear the vehicles out sooner, and have to replace your fleet sooner, which is a very high cost.
With automated metros I believe that running more service will bring in enough more riders to pay for the marginal additional costs
In general, this is a great mindset, however, there is a limit. At a certain point it doesn’t matter how much service you run, there are a fixed number of people willing to travel/for whom the transit line makes sense, and after you capture all of them more service won’t generate more riders. Freeways and airports are empty at night, because most people are at home not out shopping or going to work. At a certain point in the off hours running full service doesn’t make sense because you cannot get more riders no matter how frequent or convenient the service is, and you are just incurring a higher electric bill and maintenance costs for the extra trains.
There are exceptions, Friday and Saturday night travel patterns are different, plus special events like sports or concerts.
I think honestly after midnight the people who won’t wait 12 minutes for a train would pretty much always get a taxi regardless.
I think it’s worse than that. I think a big part of why New York’s running costs are so much higher than other places is because there isn’t an overnight shutdown most nights of the week to complete maintenance tasks.
It is much easier to plan e.g. litter picking in London than New York because you can just do it Sunday-Thursday nights without any fuss. Hell you can just have people paid to work a fixed pattern 5 nights a week as full time employees for tasks like that and they are on a fixed working week.
Overall it almost certainly means that literally billions of dollars a year of taxpayers money is being spent keeping the New York Subway open nights from Sunday to Thursday and 99%+ of taxpayers never use it. Its crazy.
Yeah fully agree there.
@Onux I see one of our failures to communicate. You are using “off peak” service to mean/include late night, while I consider 3 levels: peak, off peak, and late night. Peak is “rush hour” and generally from 6:30-8:30am, and 4-6pm weekdays (and sometimes lunch). Offpeak is 8:30am-4pm, and 6pm-(9 to 11pm). Each city is different of course, so you may have to adjust if the local culture is different. Sports and other special events may also change late service into off-peak service.
You must reduce service late nights because it is critical that every section of tracks be closed at least 1 time per week for at least 6 continuous hours so that track maintenance can happen. In those 6 hours the only trains allowed on a closed section of track are specifically for track maintenance and are not carrying any passengers (track workers would not count as passengers when riding to their work assignment). If your system is built correctly you can provide half hour service late nights while every section of track is closed once per week, but this does require planning a head and adding switches/cross overs that are not used in normal service (there is a good argument that you should close the whole line and run night buses instead – not my preferred solution, but not a bad solution that you should consider)
When I’m talking about frequency service I mean things like 2pm on a Sunday when there will be a lot more than 3 riders, even if the train isn’t full it should still have a lot of passengers.
I think when you understand that I have already rejected frequent service late at night and I’m referring to daytime off peak service instead we are a lot closer to the same.
Which dank nether region did you pull that vague number from? A moment or two of asking “NYC subway budget” didn’t find a quick answer to that question. It did find the total budget for all of the Metropolitan Transportation Agency’s operations. Which includes the commuter trains, the money sink of buses and the bridges and tunnels. In a very very round number, 20 billion.
Buses cost a lot of money to operate. How much will it cost to run the replacement buses?
@henrymiller74
Thank you for the clarification. I had seen some of your earlier comments about 24×7 and assumed you were arguing for running full service at almost all hours of the day. Yes, we are much closer, I can agree to running a single high-frequency pattern during the middle 12 hours of the day every day. Although I still think it would be beneficial to drop frequency in the later evening hours, before the overnight shutdown times. Outside of certain nightlife spots there isn’t the demand for truly frequent service at 9 or 10 pm (although with driverless you could still run far more frequent service than the owl/overnight services most places provide, and do so on every line rather than a separate late-night network).
@adiron, my understanding from Alon was that the New York Subway had operating costs that were about double of e.g TfL’s or Paris’s? If that isn’t a fair comparison then my mistake.
The London Underground costs £2.7bn a year to run (p109), or $4bn PPP. So if 1/4 of the excess was down to the lines not closing overnight then that would amount to billions.
The very comprehensive bus network (and streets) in London costs £3.6bn or $5bn PPP – it’s difficult to imagine the night bus network as a whole costs more than $1bn a year to run – but I guess if New York doesn’t have any night buses at all perhaps it breaks even roughly.
Bus drivers aren’t volunteers.
Marginal cost handwaving works both ways. If it costs almost nothing to run driverless trains around it costs labor to run the staffed trains overnight. Which is less staff than substituting buses.
@henry, when talking about automated trains, I think you’re right, my fault for picking the wrong context here. Even though…
I remember reading somebody complaining that some security guards weren’t letting people inside an automated L9 train which incidentally starts close to Camp Nou, on a matchday. And it made me suspect that they may do some trick to accumulate trains at the head for the time the match finishes, on the account of an unidirectional and sudden passenger surge. We won’t be able to confirm until the stadium reform finishes, sadly.
One of the biggest such surges had to have been when France won the World Cup in 2018 and a million fans immediately headed to the Arc de Triomphe/Champs Elysée, where the team would do a ceremonial procession drive-thru. The Metro closed the closest stations making fans walk further. Despite all the fears it handled the crowds, perhaps aided by the M1 platform doors and 85 second headways.
A million people in central Paris isn’t that massive given the metro handles 4 million rides a day.
It could easily be planned both with and without drivers.
Speaking as someone who was there: it didn’t handle the crowds well. The M1 trains themselves were not particularly crowded, but all the passengers were trying to get off at the same few stations, which didn’t have egress capacity for them because the fare barriers are designed around security and counter-tailgating first, not around capacity. So the station passageways were overcrowded and I couldn’t get out, and had to get back on the next train. That’s what led to cascading station closures. (You may recall, I had to get off at a station outside the city and walk back east.)
Maybe Micheal has reasonable expectations. And doesn’t expect it to be set up for once a generation events.
It may not have been perfect but it did ‘handle’ the massive crowds in the sense of delivering them without major incident, you know like a complete closedown or the Itaewon Halloween crush.
Line 1’s new driverless and platform door system also came thru another singular event. As described by Clive Kessell in Railengineer journal:
My experience taking transit to and from baseball games in the US is that transit demand evens out because big stadiums are big and the walk from your seat to the platform could be anywhere from 2 to 20 minutes. In addition, some people linger to celebrate after a game and others want to get to bed early and leave as soon as possible, sometimes before the game ends. (Though I wonder how the new shorter game lengths affect this.) So probably the crowd’s arrival on the platform is spread across 60 minutes.
Hello Alon Levy,
I think your article contains some misunderstandings and misinformation.
The number of students: there are more types of schools than Gymnasium. Your list of high schools for the “Wörder Wiese” subway station would only include Maria Ward Gymnasium with 556 students. In fact, there are more than twice as many because it also has an elementary and middle school.
The Insel Schütt school with approximately 450 students is not on the list. And especially not the GSO University of Applied Sciences with approximately 8,000 students at this location. So, approximately 10,000 students for this subway station.
The previous stop, “Rathenauplatz” also has several thousand students in its catchment area.
The U2 and U3 lines run on the same tracks between Rathenauplatz and the main station. Normally, a long train with a length of 75 meters always runs on the U2. A short train with a length of 37 meters often runs on the U3.
The following situation: After school, the 75-meter-long platform is full of students. Then a short U3 train arrives. This one doesn’t stop in the middle of the platform, but at one end.
All the students want to board the train. The students from the other end run to the train and squeeze through the front doors. This takes longer and caused problems because the automatic train couldn’t close the doors.
Different schools are not normally coordinated with one another, so they wouldn’t all be crowding the same station within the span of a few minutes. If it’s not a few minutes but a longer period, let’s say 30 minutes, then these numbers aren’t so scary – the capacity of a G1 is 604 passengers, and if the students are split across more than a few trains then they’re not going to stress the capacity.
Universities are a bigger deal, yes. I was thinking of also mentioning Stockholm in the post. Both Stockholm University (33,000 students) and KTH (14,000) are on the same branch of the Red Line. The trains are long (140 m) and frequency-limited (5′) because of the branching. Courses at KTH are not modular, but there is still peak crowding in the direction of KTH and SU (north) on the branch in the morning, it’s just not enough to overwhelm the system.
Virtually all schools in Germany start at 8:00 AM and a large share of those that don’t start at 8:00 AM sharp start within ten minutes to either side of 8:00 AM….
Yeah, but the walk distances from the train station to the school building aren’t always the same, and that gives some additional wiggle room.
Define coordination. Perhaps they are bright enough to consult with the transit provider and decide the school on the north side will go “first” and the school on the south side will go “last” and the transit provider schedules a few extra trains.
It seems to me here the main issue is running trains shorter than the platforms, rather than the platform length being limited to 75 meters.
Thanks for reposting.
(To explain to everyone else: Stanze double-posted because of comment moderation problems and then people replied to both posts instead of just one; I asked Sean to repost to consolidate.)
On the point about trains being shorter than the platforms: it’s possibly to indicate with electronic signs where the train will stop, as is done in Berlin when the S-Bahn runs shorter than full-length trains.
People don’t always pay attention to that stuff to be fair.
A similar issue occurs with cable cars, but worse due to the smaller capacity of a cable car. Unfortunately, there is a current trend to build cable cars to universities.
I mean this potential dilemma was behind the adoption of subway through-running in Tokyo, as they found that on the Ginza/Marunouchi/Yamanote having to absorb the combined weight of JR/Private lines in the big terminals like Shinjuku or Ikebukuro, instead better use one seat rides that distribute passenger load over a bunch of within Yamanote zone stations. Eat a lower frequency to prevent timetable cascades.
I think its also why when they did the late 2010 upgrades to the Ginza line they didn’t push it towards higher (above 30 tph) frequencies. Ditto with the recently CBTC Marunouchi line. If you’re moving 1 million people per day and your stations were built 100-70 years ago, you don’t actually have the platform space to funnel people though.
Snowdon station in Montreal provides another comparative example. There, the blue line is often overwhelmed at peak school times by students coming off the orange line. Here, the mismatched headways (about every 3 and every 5 minutes respectively on the late peak) between the two lines contributes. If two orange line trains in each direction arrive between two blue line departures, people get left on the platform. This almost always clears on the following blue line train.
This is a simple maths question.
If you have a small frequent system that can do 85 seconds headway (which in fact will be worse when crowded as masses take time to enter/leave vehicles), and let say each vehicles can pack 500 people in, then if your relative infrequent system can dump 5000 people into it every 5 minutes then it will be overcrowded
But the actual number needed to overcrowd could be much less, since not just because the headway will be less than ideal when everything are packed, but also the trains might not come empty and there could be other people on the train before reaching the transfer station which the system still need to transport
Bahninfo is a more obscure forum compared to Drehscheibe or ICE-Treff… How did you come across it?
Gadgetbahns have problems. Airtrain’s main problem is that it’s a gadgetbahn. That breaks down. The Port Authority is, according to Wikipedia, replacing it with another gadgetbahn. At a billion, with a b like in ballooning costs, dollars a kilometer. Lets hope they move the new Terminal A station from the far reaches of the parking garage to someplace closer to the terminal.
Another classic “overwhelm” is the Waterloo and city line in London.
In practice this has been largely ticketless at the Waterloo end (ticket included in season tickets, enforced by suburban rail gateline) but some queuing at touch in/out posts for pay as you go riders.
Bank end has large shared gatelines
Hey Alon can I get an invite to your discord please. Long time reader and follower.
“The airport can be directly dispensed with – individual planes don’t do this at airside people movers, and don’t even do this at low-capacity landside people movers like the JFK and Newark AirTrains.”
Well, this doesn’t happen with plane-to-airside people mover transfers for the same reason the TGV transfers don’t overload the Metro: egrees time from planes is long, and variable dependent on luggage. Unless your connection is very tight – which does happen, but a) you still have the long egress time mitigating the situation and b) tight scheduled connections at least tend to happen at gates very close to each other – there’s less time pressure on the disembarking end than on the arriving end.
So this could still be a problem with transit-to-landside people mover transfers, even without fare barriers like at JFK and EWR today. Ideally, trains should make at least one stop within the terminal area to thin the crowds out. If this isn’t possible, a couple alternate solutions come to mind: