Public Transportation in Megacities

I’ve been talking so much lately about integrated timed transfer in the context of Boston that people started asking me if it’s also applicable to New York. The answer is that the basic principles are not scale-dependent, but the implementation is, so in very large cities, public transport planning should not look like in Switzerland, a country whose largest metro area is staring at 2 million people from the bottom.

The one caveat here is that most cities are not huge. The developed world has seven megacities: Tokyo, Seoul, New York, Los Angeles, Osaka, London, Paris. And Los Angeles doesn’t really have public transportation, so we’re down to six. The middle-income world has a bunch more for sanity checking – Mexico City, São Paulo, Rio de Janeiro, Buenos Aires, Johannesburg, Moscow, Istanbul, Tehran, Beijing, Shanghai, Guangzhou, Shenzhen, Bangkok – but all are either still in convergence mode building up their networks or (mostly in Latin America) have given up. So much of this comes down to the idiosyncrasies of six cities, of which the largest three networks are substantially in the same planning tradition.

Demand is huge

Big cities have big centers, which can’t really be served by any mode except rapid transit. Even in Los Angeles, what passes for a central business district has around a 50% public transport modal split. This means that the transport network has to deliver high throughput to a relatively small city center. Even in a low-kurtosis city like Paris, most Métro lines converge on a narrow area ranging from Les Halles to Saint-Lazare; in a high-kurtosis one like New York or Tokyo, there are a few square kilometers with 200,000 jobs per km^2, which require an exceptionally dense network of rapid transit lines.

Some other network design principles follow from the need to amply serve city center. Specifically, high frequency is rarely a worry, because there’s so much demand even off-peak that usually megacity subway systems do not venture into the frequency range where long waits deter traffic; New York’s 10-minute midday gaps are bad, but that’s unusual and it comes from a combination of the legacy of postwar fear of subway crime suppressing demand and excessive branching.

But other principles require careful planning still.

Electronics before concrete, megacity version

The driverless lines in Paris support peak throughput of 42 trains per hour – a train every 85 seconds. CBTC on Line 13 without driverless operation supports 38 tph, and London’s CBTC-equipped lines support 36 tph when the branching isn’t too complex. It is imperative for other cities to learn from this and do whatever they can to reach similar headways. The difference between 21 tph, as in Shanghai, and Paris’s 42, is equivalent to building a brand new subway line. And what’s more, in a city in the size class we’re talking about, the primary concern is capacity – coverage is already good, so there really is no reason to build two 21 tph lines instead of one 42 tph one.

The situation in Paris is in a context with self-contained lines. That said, extremely busy self-contained lines do exist in other megacities – London has a bunch with near-Parisian levels of throughput, New York has some, Tokyo has a few, Seoul and Osaka are both more self-contained than Tokyo is.

Throughput and organization

The primacy of throughput means that it’s worthwhile to build small infrastructure upgrades, even with concrete, if they help with capacity. Right now the Northern line reverse-branches with the branches to the north recombining with those in the center, and Transport for London would like to split the line in two, reducing branching complexity, which would increase capacity. But doing so requires improving pedestrian circulation in the corridors of the branch point, Camden Town, where TfL expects very large transfer volumes if there’s a split and already there are circulation problems today without a split. Hence the plan in the medium term is to upgrade Camden Town and then split.

If there are bumper tracks at the end of a line, as at 8th Avenue on the L or Flushing-Main Street on the 7, then it’s useful to dig up the street for another block just to add some tail tracks. That way, trains could enter the station at full speed. This increases throughput, because the terminal interlocking has trains heading in opposite directions crossing each other at-grade, which imposes schedule constraints; it’s best if trains can go through the interlocking as fast as possible to reduce the time they’re in a constrained environment, but that in turn requires short tail tracks so that an overrun of a few meters is not catastrophic. Ideally the tail tracks should even extend a full train length past the platform to place the interlocking on the other side of it, as is done in Paris and Moscow; in that case, trains cross the interlocking out of service, when it’s easier to control their exact timings.

Such projects are disruptive, but the disruption is very localized, to just one transfer station for a deinterlining project as in London or one terminal as in New York, and the impact on capacity is very large, if not quite as large as the full suite of signaling and track upgrades that make the difference between a train every 3 minutes and a train every 1.5 minutes.

Network design

The ideal metro network is radial. Megacities already support that just because so many lines have to serve city center. However, it’s important to make sure every pair of lines intersects, with a transfer. No large metro network in the world achieves this ideal – Mexico City’s network is the largest without missed connections, but it is not radial and its only three radial lines are overburdened while the other lines have light ridership. Paris has just a single missed connection on the Métro proper, not counting the RER, but it has many pairs of lines that do not intersect at all, such as M1 and M3. London is more or less a pure radial, but there are a handful of misses, including one without any transfer between the two lines anywhere, namely the Metropolian line (including Hammersmith and City) and the Charing Cross branch of the Northern line.

Big cities that plan out a metro network have to make sure they do better. Missed connections reduce passenger ridership and lead riders to overload the lines that do get connections; for example, in Tokyo one reason cited for the high ridership of the Tozai Line is that until Fukutoshin opened it was the only one with a transfer to every other subway line, and in Shanghai, Line 1 was extremely congested as long as the alternatives going north either had critical missed connections (like Line 8) or avoided city center (like Line 3).

The role of regional rail

Regional rail as a basic concept is mostly scale-invariant. However, the design principles for trains that come every half hour are not the same as those for trains that come every 5 minutes. If trains come every half hour, they had better connect cities in a roundtrip time equal to an integer number of half hours minus turnaround times, so that they don’t have to loiter 25 minutes at a terminal collecting dust and depreciating. If they come every 5 minutes, they’re not going to loiter 25 minutes anyway, and the difference between a 5-minute turnaround and a 7-minute turnaround is not really relevant.

The design principles are then mostly about throughput, again. The most important thing is to build independent trunk lines for trains to serve city center. Even in a huge city, the finances of building a purely greenfield subway deep into suburbia are poor; Tokyo has done it with the Tsukuba Express but it’s mostly above-ground, and for the most part regional lines there and elsewhere come from taking existing suburban lines and linking them with city center tunnels.

Tokyo’s insistence on making these city center tunnels also form a coherent metro network is important. Only one non-Tokyo example is worth mentioning to add to all of this: this is Berlin, which is not a megacity but has three independent S-Bahn trunk lines. Berlin, unlike London and Paris, painstakingly made sure the S-Bahn lines would have transfers with the U-Bahn; its network has only one U-Bahn/S-Bahn missed connection, which is better than the situation in Tokyo, Paris, or (with Thameslink and Crossrail) London.

The role of development

All first-world megacities, and I believe also all megacities elsewhere, have high housing demand by domestic standards. All are very wealthy by domestic standards except Los Angeles, and Los Angeles is still incredibly expensive, it just doesn’t have the high wages to compensate that London and New York and Paris have. In such an environment, there’s no need to try to be clever with steering development to transit-oriented sites. Anywhere development is legal, developers will build, and the public transport system has a role to play in opening more land for more intense development through fast trips to the center.

A laissez-faire approach to zoning is useful in such an environment. This contrasts with smaller cities’ reliance on finger plans, like the original one in Copenhagen or the growing one in and around Berlin. No limits on development anywhere are required. The state’s planning role remains strong through transportation planning, and the suburbs may well form natural finger plans if developers are permitted to replace single-family houses with apartment buildings anywhere, since the highest-value land is near train stations. But state planning of where housing goes is counterproductive – high transit ridership comes from the impossibility of serving a large central business district by cars, and the risk of politicization and policy capture by homeowners is too great.

The advantage of this approach is also that because in a high-demand city public transport can to some extent shape and not just serve development, it’s okay to build lines that are good from the perspective of network coherence, even if the areas they serve are a bit light. This principle does not extend indefinitely – subway and regional rail lines should still go where people are – but for example building key transfer points in near-center neighborhoods that are not in high demand is fine, because demand will follow, as is building lines whose main purpose is to close some gap in the network.

Construction costs

The larger the city, the more important cost control is. This may sound counterintuitive, since larger cities have more demand – only in Manhattan could a $1.7 billion/km extension like Second Avenue Subway pencil out – but larger cities also have a bigger risk of cost blowouts. Already Tokyo has stopped building new rapid transit in the core despite very high crowding levels on the existing network, and London builds next to nothing as well. New York’s poor cost control led Philip Plotch to entitle his book about Second Avenue Subway The Last Subway. Even Paris builds mostly in the suburbs. Extensive city center and near-center construction continues in Seoul, in the context of very low construction costs.

The flip side is that a New York (or even London) that can build subways at the cost of Paris, let alone Seoul, is one that can rapidly solve all of its transport problems. My Assume Nordic Costs map fixates on a region of the world with small cities, but the construction costs in South Korea are if anything lower than in the Nordic countries. And even that map, given free reins for developers, is underbuilt – some lines would look ridiculous at current costs and zoning but reasonable given low costs and liberal zoning, for example something meandering through currently industrial parts of New Jersey.

Small cities designed their public transportation philosophy around scarcity: Switzerland really can’t just draw crayon and build it, because housing and transport demand there are finite and limited. Cities like New York and London, in contrast, should think in terms of abundance of infrastructure and housing, provided their regulations are set up in a way that permits the state to build infrastructure at low costs and private homebuilders to redevelop large swaths as they become easily accessible to city center.


  1. yuuka

    You might probably know this from elsewhere already, but for some reason, Chinese subways apparently require the driver to physically get out of his seat and watch the space between the train and the platform doors after two incidents where people were trapped between the train and the platform doors, adding about 15 seconds of time to each station stop. There was a 2018 TRB paper that studied this.

    That might be a reason why frequency is so restricted, and given how much the construction sector contributes to economic growth I daresay they may find it a better idea to drop money on building more subway lines. I would have said the same for Japan, but the private operators there have an interest in keeping frequency high so as to reduce capital investment, so you see things like 30tph on the Chuo Line.

  2. Gok (@Gok)

    LA is really more of a country with than a megacity. Greater Los Angeles is around 88,000 km^2, larger than the other six put together. (Greater Tokyo, New York Metropolitan Area, Île-de-France, Keihanshin, London Metropolitan Area, Seoul Capital Area together are around 68,000 km2). If LA is a megacity then so are Taiwan and Benelux, which are both smaller in area and larger GDP and population (and of course have much more coherent public transit).

    Is there science about improving throughput with shorter headways vs. larger/longer trainsets?

    • Fredrik Staxäng

      The Los Angeles metropolitan area is 12,562 km, which includes all of LA County and Orange County. The comparison with Benelux is interesting though. The train/bike transportation system and narrow row houses leave a lot space for other uses.

    • df1982

      Logically, shorter trainsets also have a higher bar for maximum possible frequency (since they take less time to pass through a given point on the line). So while the Paris metro’s relatively short trains can achieve 42 TPH, longer ones, such as on the New York Subway, would be more constrained. London can’t seem to get its Victoria line beyond 36tph (with mid-length trains at 133m per set), even though it is perfectly unbranched, and it’s not for want of trying.

      • Nilo

        Moscow hits 39 TPH according to Alon, and has roughly 154 meter trains on line 1 according to Wikipedia.

      • Alon Levy

        M14 has 8-car platforms, so 120 meters. And “London can’t do something Paris can do” is not that much of a man-bites-dog story.

        • df1982

          But they only run 5-car trains at the moment. I’d be interested to see if they can keep 42tph once they go to full-length trainsets, I suspect not. In any case, I have used the M14 quite frequently over the years and they never really seemed to get to headways of 90s or better in real life, even at the peak of peak. The best I saw was usually 100s (the PIDs are in 10s increments).

          The point still stands that all else being equal shorter train sets allow for tighter headways, so you can’t automatically assume that a system with longer trains can match the frequency of those with shorter trains.

      • Tonami Playman

        From my understanding London is yet to adopt the aggressive turnaround practice. Decelerating and Accelerating over the 1.3m/s limit past the terminal statitions before reentering service on the opposite tracks and faster door opening and closing. So still some room to shave time plus the extra dwell time having to do with the upwardly curved sliding doors over passengers heads. Where all other metros have flat vertical doors.

        • fjod

          The reason the Victoria line can’t run more trains is that the platforms at a couple of stations (mainly Kings Cross St Pancras) are poorly designed for egress in the peak. London’s tiny trains do contribute to slower egress as you say and this worsens the situation. Other solutions aren’t worthwhile until this is solved.

          • Matthew Hutton

            I think one of the other reasons for not extending it is that there are quite a few stations that are nearly as bad as Kings Cross, so fixing Kings Cross would get you to 95 seconds, but then you’d need to fix about 6 stations to get to 90 seconds so it isn’t worth it.

      • Kamoro

        That shouldn’t be confused with higher capacity though. A shorter train has a shorter headway, but it is the GOOD headway that is reduced. If a headway consists of X seconds that a train is passing, and Y seconds that a train is not passing, reducing the train length reduces X. You want to reduce Y, while keeping X large.

        • df1982

          Yeah, I’m not saying you SHOULD make trains shorter in order to increase maximum TPH, just that the two are linked. Once you get down to a train every two or three minutes, there is little advantage to the passenger of shorter headways. Maximum capacity at this point is more about ingress and egress, so door design and seating layout on the trains, and stair/escalator layout and platform/corridor space in the stations.

          • Alon Levy

            Just to put things in perspective… at M14 cruise speed, a train clears 1.5 car lengths per second. At lower cruise speed more typical of local trains, it clears a car length per second. Whatever impact train length has on minimum headway is second-order.

          • df1982

            Yes, but when you’re talking about getting headways down from 100 seconds to 90 seconds it does become a relevant factor. And you also have to take into consideration the added inertia of a longer and therefore heavier train, which means the minimum space between the two trains has to also be greater to prevent one from crashing into the other if there is an incident. I think it’s very clear there is a link between train length and maximum permissible frequency.

            Although it has never implemented it in practice, Vancouver apparently has a maximum theoretical throughput of 48 TPH (75 second headways), but again, the system runs with 2-3 car trains. If it had 8-car trains it would not be able to achieve that, even theoretically.

          • Matthew Hutton

            48tph with 3 car trains is worse than 25tph with 6 car trains.

          • michaelrjames

            JM’s cri-de-coeur and your post provoke me to post this that I posted here a few years back:

            Seats per hour:
            ……………………………30tph ……..25tph ……..15tph
            bi-level MI09 trains: …..28,440 ……23,700…….14,220
            1-level MI84 trains: …..18,720……15,600……….9,360
            CrossRail(short train):..13,500……..11,250……….6,750

            Total pax per hour (standing + seated):
            ……………………………30tph ……..25tph ……..15tph
            RER-A MI09 trains: …..78,000.……65,000……..39,000
            CrossRail(short train):..45,000……..37,500……..22,500

            This was in the context of line capacity conditioned on old 1-level versus newer bi-level trains on Paris RER-A which carries >300m/yr with a peak rate of >60,000/h each direction. I believe these big trains can be run as twin-sets of total 10 cars and length 224 metres. With the new bi-levels RER-A had dropped its frequency from 30 tph peak/24 tph reverse peak, to 25/24, due to increasing pax numbers causing occasional incidents (presumably while egress/boarding) that can cause network disruption when attempting the higher train frequency. The point being that even when forced to slow the service down the bi-levels still carry 5,000 seated and up to 20,000 total more pax/h than the single-levels trains. This has nothing to do with the train’s capability in acceleration or deceleration but purely the human factor in orderly access; I suppose with foresight they might have made the inner-city stations with Spanish platforms though they are already gigantic.

            I was also speculating on how London CrossRail would cope with their proposed 15tph and short trains for their debut service (obviously the can run higher frequency and run longer trains but at first they won’t have the trains or carriages).

          • yuuka

            Added inertia with longer cars is an especially bad take since logically you’d expect there to be a corresponding increase in braking power; trains with different lengths should have largely similar dynamic characteristics. In any case that is not the main issue; technically the issue with most CBTC solutions today, from what I know, is that they don’t really give you the benefit of moving block through interlockings, so shorter trains would naturally spend less time in interlockings. But I don’t think it’s *the* biggest problem.

            The bottleneck here is not just ingress and egress, but passenger behavior – an idiot holding the doors is more likely to ruin headways irreparably in 42tph Paris than in 21tph Shanghai, and in 39tph Moscow this issue is moot because they have no qualms slamming the doors on people. In any case both are extremes (my personal sweet spot hovers around 36tph), and while there is little advantage to the passenger in going that high, the advantages are to your capacity goals. Short trains and high frequency might actually be attractive to Alon and other bean counters, since short trains mean shorter stations, smaller pits, and lower construction costs.

            Also, Vancouver pairs two married pair Mk IIs to create 4-car trains in addition to the 4-car Mk IIIs on the Expo Line, and it’s on the Expo Line where 90 second headways are frequently seen. It appears to me that the problem with RER A, much like NYC, might be that it can’t go any lower due to all the schedule constraints imposed by turning off to the branches – 36tph is a reasonable demand of a somewhat segregated metro line, not the trunk section of three headed regional rail.

          • Nilo

            Yukka, I believe there’s a long discussion on this blog of the RER A’s issues and the lower frequency has to do with the use of bilevel trains. Whether this is a good idea or not is debated pretty thoroughly in the comments between Michael James and Alon.

            Longer trains are good if your city is big, and probably a waste of money if your city is small. Vancouver isn’t likely to have 20 million people any time soon so the reliance on short hyper frequent trains definitely makes sense from a capital costs perspective. On the other hand São Paulo’s Line 4 is entirely automated, but seems to have made this mistake opting for thinner trains then the unautomated older lines, and thus giving up a quarter of the capacity per train of those lines.

    • adirondacker12800

      Smells like you are comparing Metropolitan Statistical Area to Combined Statistical Area.

    • Nilo

      Are you including the entirety of San Bernardino County in that number? That just isn’t an accurate description of LA’s urban form if so.

  3. Tonami Playman

    The issue of picking between shorter headways vs. longer trainsets is very evident in Shanghai with Line 6. A line that was royally messed up from day with the designed undercapacity. The line uses 19.5m long and 2.6m wide trainsets at 4 cars long for a total length of 78m. Platforms are 80m long and they can’t use wider A(3.1m) or B(2.9m) size trains because the loading was designed to accommodate no more than 2.6m. This on a corridor that is the only North South trunk on the Pudong side of the Huangpu river. It gets so crowed they have to close several stations along the line during rush hour.

    The opening of Lines 9 and 12 have done nothing to relieve the crowding on the line. Shanghai can decide to go the shorter headway route or the longer train route. The line currently runs 30tph (Maybe because the line is yet to have platform screen doors that’s why they can achieve such a high frequency relative to other lines in Shanghai) on the busiest sections at peak. So going up to Paris M1 frequency of 42tph will increase capacity by 40%. They could also choose to keep the current frequency, but lengthen platforms to 6 car trains which will increase capacity by 50% or go even further with 8 car long trains which will increase capacity by 100% and bring it in line with Line 1 running 6 car 140m long 3.1m wide trains at 24tph.

    • yuuka

      One thing I’m not sure that I entirely got right was why that is a problem on Shanghai lines 5 and 6 – 8 can’t be helped because it already has platform doors and presumably there is lesser headroom there (still no excuse for not running 8-car trains on line 8 especially if platforms are built that long). Beijing lines 1 and 2 were built to the same type-C specifications in the 70s albeit with 6 cars, but now use modified B types that are wider at the waist than they are at the floor for increased passenger capacity.

      As for Paris, M14 is getting 8-car trains as well with the GPE-related extensions to the north and south. I’d like to see M14 retain the same frequency with the 8-car trains, to see if interlocking occupation times are that skewed in favour of shorter trains.

    • Tonami Playman

      Correction on my earlier post. Shanghai Line 1 uses 8 car trainsets and not 6 cars so the total length is 186m. A hypothetical type C 8 car trainset at 30tph will only approach 90% of the capacity of an 8 car long type A trainset at 24tph.

    • michaelrjames

      It seems no matter how they approach the problem Shanghai is going to incur very substantial costs and probably lots of disruption to existing service to slowly improve service.
      What interests me is how they got here. It seems they have made a whole series of poor decisions with entirely predictable terrible consequences down the track a bit. Is this just retrospective wisdom in the eyes of we observers, or is it simpler in that back in the 80s when they began building their system they weren’t nearly as wealthy as we today tend to think of Shanghai, and were “forced” into these compromises? It seems to contrast so much with some other huge infrastructure China has undertaken so successfully, such as the HSR system where they set out to trial and learn (ahem, steal) the best from the rest of the world. OTOH this was long before their HSR program so maybe they learnt the hard way? Also I would have thought Shanghai took their cues from Hong Kong, rather than Beijing, and is much more pragmatic versus ideological than Beijing. I suppose it is too early to have a (English) history of this story? (Funny enough, the book Shanghai on the Metro is about White Russians on the Paris Metro!)

      One also wonders about these true mega-mega-cities (of which there is really only one in the rich west, Tokyo) whether transit will ever be able to cope. Like roads, the only factor limiting use/overuse will be congestion.

      • yuuka

        Shanghai is mainly doing things properly with big type-A trains like HK (unlike Beijing still sticking to Soviet-sized B types except for 14 and 16), except for lines 5, 6 and 8 – 5 is understandable as a light metro (and de facto line 1 extension) proposed by a district-level authority.

        6 and 8 is more tricky – from Chinese Wikipedia it looks like the business cases were rigged to reduce carrying capacity and so justify allowing Shanghai Electric-Alstom to produce type-C vehicles from their local manufacturing plant. Lines 3 and 4 might also be a bit touchy, but there apparently exist plans to provide a new terminating facility for line 4 (so they don’t loop) so that the shared section can take more line 3 trains.

  4. Matthew Hutton

    To be fair some of the missed connections in London are very short – Warren Street to Euston Square is like 100m. That said maybe it’s worth the disruption of building a new tunnel to link some things like that together.

    • df1982

      By the same token, some of the S-Bahn/U-Bahn connections in Berlin are pretty bad, either in involving long underground walks (Alexanderplatz) or even having to traverse street space (Charlottenburg-Wilmersdorfer Straße). It helps that Berlin doesn’t have fare gates.

      • Henry

        There are OSIs in London (Euston Square to Warren Street actually is one) but many of them are not publicized, partially because of lack of space on the tube map and partially as a means to not overwhelm certain stations.

        I believe that is the reason Crossrail doesn’t have a transfer to the Victoria Line, for instance.

        • Eric2

          It would also be a little silly for Crossrail to have two stops 400m apart. Maybe when there is a Crossrail 2 to take the load off the Victoria Line, they can add a walking transfer here. (That seems to be what Alon drew here)

          • Alon Levy

            London’s recent innovation of deliberately missing connections to make people not ride their system is charming, yes. What’s especially charming is that they’re not even trying to build a system that can be retrofitted later – e.g. the Battersea extension crosses the Vic line far from any station, even though it’s designed around the same time as Crossrail 2.

          • barbarian2000

            building a single station at Oxford Circus with pedestrian tunnels to Bond Street at one end and TCR at the other would probably work out, but Crossrail likely decided that the demand at a single station would be too high and cause severe overcrowding

          • Nilo

            I mean have you seen how much crossrail stations cost? For what they’re paying I suspect they could have made a station that could handle all the pedestrian flow. It’s not like London’s stations are going to have much higher passsenger throughput than Tokyo’s or São Paulo’s

          • barbarian2000

            Tokyo and São Paulo both have exceptionally overcrowded systems though, albeit largely caused by bad operational practices in the latter case

          • Eric2

            Sao Paulo has bad operational practices? All I know is it has the second highest ridership per km in the world (after Cairo). That sounds like overcrowding even if operational practices were great.

          • barbarian2000

            the suburban rail system – which has metro-like frequencies – has a 4-track corridor running through the city centre but insists on terminating all services there

          • michaelrjames

            @Eric2: “Sao Paulo … has the second highest ridership per km in the world ”

            Surely that is the wrong metric to gauge “success”. The metro has a puny 101km of route so in the biggest city in south America and one of the biggest in the world, of course it was going to be crowded beyond tolerance. It’s a bit like saying the massive road congestion in LA/Sao Paulo/Djakarta etc indicates a successful road policy. I suppose a better metric would be the percentage of city trips is by transit.

          • Matthew Hutton

            With regards to the lack of interchange with the battersea branch at Vauxhall. Sure on the map it’s disappointing but which station pairs does it really make a difference for? Brixton is the only one in theory where it might be worse but to Waterloo you’d change at Vauxhall for a mainline train and everywhere else you can change at Green Park/Oxford Circus

          • fjod

            There are a few more negatives:
            – Byzantine fares mean you’re unnecessarily penalised for taking tube and mainline rail in one journey (e.g. Brixton>Vauxhall>Waterloo); this is a fare of £2.40 vs. £4.10. This also affects a lot of journeys if/when the line is extended to Clapham Junction.
            – Brixton also will have less easy access to Embankment/Charing Cross/Leicester Square. For the latter, the Green Park change is so hideous that people actively avoid doing it if they can.
            – For Nine Elms and Battersea, it will be harder to reach all destinations north on the Victoria line (the Charing Cross branch takes longer). Yes you can change at more central stations but there is a significant time penalty for this.
            – Brixton to Battersea is also ignored.

            Ultimately this is quite a lot of negatives for a two-stop extension.

          • Matthew Hutton

            I think it’d be difficult to make the Vauxhall interchange good. I think it’d be more like Green Park which is poor for London.

            – For nine elms for destinations north on the Victoria line you’d walk to Vauxhall.
            – From Battersea for Victoria you’d take a mainline train so it’s only Battersea to Green Park and Battersea to Pimlico that’s difficult as for Oxford Circus you could change at Waterloo or Embankment onto the Bakerloo line and otherwise change at Warren Street.
            – For Brixton to Battersea the best solution would be to add an overground station at Brixton – which would also improve a whole bunch of station paired journeys.

            I think there are significant missed connections in London, the central line and Piccadilly at Park Royal, the south London line at Brixton, the Metropolitan line and bakerloo where they cross at Northwick Park and West Hampstead with lots of lines come to mind. But I don’t think the northern line extension is one – unless it is extended beyond Clapham junction or the Victoria line beyond Herne Hill

          • barbarian2000

            that Bakerloo/Metropolitan line connection has stations close enough that it really just needs a protected walkway between Kenton and Northwick Park

          • Eric2

            “the suburban rail system – which has metro-like frequencies – has a 4-track corridor running through the city centre but insists on terminating all services there”

            True, I had in mind the metro which is plenty crowded without that bad practice, but you are correct about the suburban rail.

            “Surely that is the wrong metric to gauge “success”. ”

            I didn’t say it was a measure of success. I said it was enough to prove that there can be crowding there without incompetence.

        • Nio

          São Paulo has more urban rail transit trips than NYC or Delhi. Alexander Rapp finds it to be top ten world wide. The system may be 1/3 the length of Delhi, but is significantly more useful with over twice the ridership. The suburban rail’s lack of through running is frankly rather bizarre given how modern the rest of the suburban rail is (open gangways, all electrified, high frequencies.

          • Eric2

            Agreed. Compared to the general Brazilian dysfunction, Sao Paulo is very refreshing with its steady construction of high-productivity well-run lines in recent years. The lack of regional rail through-running is the one mystifying point.

          • Nilo

            Construction costs are far too high, but it doesn’t seem like they build bad lines. Rio is probably the most extreme version of this, costs on the last metro extension were very high for the non-anglo sphere (The governor ended up going to jail over this), but I don’t think anyone has ever said the project route wasn’t a good choice.

          • Eric2

            Re Rio: the stub end extension to Gavea is a head scratcher, the division of corridors between metro and BRT is probably not optimal, and I would think the downtown streetcars are a bad choice. But yes other Brazilian cities are worse.

          • Nilo

            Well they got rid of the Gavea stub (the partially completed station was supposed to be filled in). And any metro/BRT Split weirdness down Avenida das Americas comes from a combo of corruption and high costs IIRC.

          • barbarian2000

            São Paulo has a lot of dysfunction of its own: Lines 4 and 5 are built to a seperate gauge than the rest of the system, each individual line is slowly being parcelled out seperately to different private companies, Line 6 construction is a mess that keeps stopping and starting, the Line 18 monorail was abrubtly cancelled right as it was about to start construction, and the Line 13 route is just horrible. Going by current plans, future lines will also be focused almost entirely on being radial lines going into the city centre, and the city looks like it’s just allergic to through-running.

            The last metro extension in Rio had a very controversial route since it was touted as an entirely seperate Line 4* even though it was practically just an extension of Line 1, to the point that a few years after the line was built they started through-running trains there, making ”Line 4” this weird undefined thing that’s technically a line but also isn’t really. The original plan for Line 4 had it going north of the Lagoon, across the western part of the South Side and into the city centre via a seperate route.

            Downtown light rail is a very good choice but expensive fares combined with a lack of fare integration make it have a much lower ridership than it could otherwise have. The light rail distributes people from higher-capacity transit throughout the city centre, integrates the port region to the rest of the city, and connects to both the long-distance bus station and the domestic airport, both of which were very difficult and dangerous to access before it was built

            The Gávea stub is still weirdly undefined rn, the partially completed station has been threatening to collapse for years now and take a huge private university with it, but it’s also forgotten about rn because the governor (who didn’t really give a shit about transport anyway) just got impeached

            *Line 3 doesn’t exist yet and has been on the drawing board since decades before the Metro even existed

          • Eric2

            “Lines 4 and 5 are built to a seperate gauge than the rest of the system, each individual line is slowly being parcelled out seperately to different private companies”

            Nothing wrong with this. Each line will be filled to capacity with riders (if it’s not already) so through-running will not be appropriate

            “Line 13 route is just horrible.”

            The dead end is stupid, but I assume it can someday turned into a branch of Line 12?

          • barbarian2000

            the different gauge doesn’t just prevent through running, it also prevents the use of the same rolling stock as the other lines

            Line 13 already sort of acts as a branch of Line 12, they use the same tracks from Brás to Engenheiro Goulart

          • barbarian2000

            ”Nothing wrong with this. Each line will be filled to capacity with riders (if it’s not already) so through-running will not be appropriate”

            They aren’t filled to capacity right now and won’t be unless the lines are extended. Having Line 4 terminate in the city centre is especially wasteful of the potential capacity of the line, *especially* considering plans for Line 19 extending north.

          • Eric2

            “it also prevents the use of the same rolling stock as the other lines”

            I don’t know why this is a priority…

            “they use the same tracks from Brás to Engenheiro Goulart”

            Ah, that’s good. On the map, line 13 just goes from airport so Engenheiro Goulart, so it’s good that it goes all the way downtown. Given that, though, what exactly is problematic about its route?

            “They aren’t filled to capacity right now and won’t be unless the lines are extended. ”

            That is hard to believe – given that their ridership per km is almost triple that of Beijing, more than double that of Moscow or Paris, six times higher than Madrid, etc.

            ” Having Line 4 terminate in the city centre is especially wasteful of the potential capacity of the line”

            Agreed. There’s no reason why they shouldn’t be able to extend it to the north later, though.

          • barbarian2000

            Using the same rolling stock on all lines eases maintenance and allows for all the same equipment to be used on the entire system.

            Line 13 only runs to Brás (and occasionally Luz) at rush hours, and infrequently at that (20-30 minute intervals). Brás is also on the edge of a very large city centre and the connections it offers aren’t *that* good considering the lack of through-running. On the airport side, the station labelled ”Aeroporto” is, somewhat confusingly, not actually at the airport, but rather about a kilometre away from Terminal 1 (which is small and little-used), and quite a bit further away from the other terminals. It’s a bit like calling Jamaica station in NYC ”JFK Airport”, but without the AirTrain.

            I’m not sure if your figures include just the Metro or the Metro+Suburban Rail, but regardless, the figures are inflated by Lines 1, 2 and 3 (especially Line 3), which immensly overcrowded. Look up ”Estação da Sé” on google images and you’ll get an idea. If your figures do include the suburban rail, then they’ll get inflated by that, too: even Line 9, which follows a perimetral route through the west side of the city on an old cargo alignment, gets overcrowded at times.

          • Nilo

            Line 4 was pretty full when I took it at the beginning of the year. Though obviously it deserves an extension northward. It feels difficult to criticize the routing like it’s complete though. The line is still being actively being extended to the south! Some numbers: in 2019 Line 4 had 229 million riders. If we assume totally flat distribution for all seven days that’s 627 riders a day, which means weekday ridership is probably significantly higher. For comparison Paris M1 (obviously significantly lower capacity due to smaller trains) 181.2 million riders.

            Same equipment is generally good yes, but I’d point out that they almost certainly weighed that against the cost premium of 1600 mm rolling stock. Alon often then justifies that premium in terms of use of legacy routes, but the extant legacy routes already— sans line 9— have good CBD access. I’m not saying they made the right choice, but I think their choice is actually pretty defensible. I think for most other Brazilian cities though 1600 mm/1000 mm gauge routes would probably be more preferable due to the extant legacy infrastructure. My real criticism of São Paulo’s standard gauge stock would be that it is significantly narrower at 2.8 meters than the 3.1 meters used in the original lines, which seems like basically trading out a lot of the capacity benefits of automation in for no good reason.

            Parceling out can work! The fares are still integrated, the transfers free. The biggest issue I found while riding it was that Lines 1, 2, and 3 didn’t show Line 4 and 5 on their in train maps (only as small transfer dots), same for line 4 with respect to the other lines.

            Yeah 13 being just an airport connector instead of actually trying to provide transit to the surrounding city of a million+ that has no rapid transit.

  5. Tonami Playman

    Alon, That Kurtosis word blew my brains out. Is there a less convoluted explanation of what it means as described in the Wikipedia article for people that are not math inclined?

    • Alon Levy

      See graphic here:

      High kurtosis means very high density at the mean, fast dropoff near the mean, and slower dropoff farther away from the mean. Low kurtosis means more gradual dropoff near the mean but no (or almost no) density far from the mean. Note: all the distributions in the link have the same standard deviation – it’s just a question of whether the contribution to the standard deviation comes mostly from a handful of points very far from the mean (high kurtosis) or many that close-ish to the mean (low kurtosis).

      • Tonami Playman

        Thanks, now I get it. That graphic flew over my head when I looked at it initially. But with the density distribution explanation, it now makes sense.

  6. barbarian2000

    ”but all are either still in convergence mode building up their networks or (mostly in Latin America) have given up”

    it’s not quite as simple as that, rail transit in Brazil at least tends to be scuppered by lobbying from private bus companies – to the point that a former governor of Rio de Janeiro state is currently in prison for both soliciting and accepting bribes from them – but there is still a lot of effort that’s put into getting transit infrastructure built

    • Nilo

      I’d say giving up probably requires a level of intentionality its difficult to attribute to the government in a state where it looks like 5 straight governors are going to go to prison on corruption charges. It’s just completely dysfunctional governance. Transport at this point is a small and incidental part of that.

  7. Herbert

    A few days ago the German EU presidency has announced plans for a new pan European night train network with trains possibly reaching up to 250 km/h.

    What’s your take on this and the announcements by various railroads and governments in Europe to increase night train services?

  8. adirondacker12800

    My Assume Nordic Costs map

    Has too much capacity in the suburbs. And has too many stops in the center.

    for example something meandering through currently industrial parts of New Jersey.

    New Jersey didn’t ask you and instead is redeveloping formerly industrial land right by the side of existing tracks that already have passenger service. Silly them. Once you get west of Secaucus there’s excess capacity all over the place. Without carving something new.

  9. Frederick

    Many non-European cities (like NY or Osaka) have significant geographical constraint that makes a hub-and-spokes system difficult. It is also more difficult to directly fit the suburban line into the city center, and compromise has to be made.

    • Nilo

      There’s almost zero non radial service west of the Hudson (I guess Newark and Hudson Bergen Light Rail count) , which is by far NYC’s biggest geographic constraint. When the Subway was being built The builders had no problem crossing the east river.

      • Henry

        The East River was less of a navigable channel than the Hudson. And Brooklyn and Queens do not have 300-ft high cliffs shortly after the riverside, or really anywhere for that matter.

        Politics was a lot of it, but in general the Palisades and the wider width of the Hudson were never going to make it easy. The original plans for a 57th St bridge had to be much larger than the East River bridges:

      • adirondacker12800

        There are things to do in New Jersey other than go to Manhattan. The the light rail lines in New Jersey go downtown.

      • Nilo

        Sure the Hudson is harder to cross, but from 1908-1910 three railroad tunnels were opened across the Hudson. For a brief moment there there were two subway tunnels between Manhattan and New Jersey, but only one between Manhattan and Brooklyn! Though of course the Manhattan and Brooklyn bridges provided rail crossings, but zero rail crossings have been opened since then. Three additional road crossings have been built in the Lincoln tunnel, Holland tunnel, and George Washington Bridge, the last of those opening in 1957/62 whether you count the last Lincoln tunnel bore or the double decking of the GW Bridge as the final crossing added. The Problem then seems not technical, given the early tunneling across the Hudson, but political and due to the various political priorities at the time of mass subway construction that favored NYC over non-NYC tunnels, which makes sense for a city built system.

        Adirondacker, obviously there are places to go in New Jersey. Obviously you know when I say radial I mean a line that hits the regional central business district, because you’ve been reading this blog longer than I have. I have no idea why you’re deciding to be a truculent asshole about this.

        • adirondacker12800

          Newark is the regional center. 9 million people in New Jersey have a lot of things to do that don’t involve Manhattan.

Leave a Reply to Herbert Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.