What Europe Can Learn From Asia

Most of what I write about is what North America can learn from Europe, but the rich countries of Asia are extremely important as well. But what’s more interesting is knowledge sharing between Western Europe and the rich countries of East Asia. These two centers of passenger rail technology have some reciprocal exchange programs, but still learn less from each other than they should.

The ongoing coronavirus outbreak made the topic of Western learning from East Asia especially important. To be clear, none of the examples I’m going to talk about in this post is about the virus itself or at all about public health. But the sort of reaction in democratic East Asia that’s staved off the infection, compared with the failure of the West to do much in time, is instructive. When the virus was just in China, nobody in the West cared. I went to a comedy night in Berlin a month ago and it was the Asian comic who joked about how all they needed was to cough and the white people gave them space; it was still viewed as an exclusively Asian epidemic. By the same token, Korea’s success in reducing infections has made it to parts of Western media, but implementation still lags, leading to an explosion of deaths in Italy and perhaps soon France and the US. Hong Kong (from the bottom up) and Taiwan (with government assistance) have limited infection through social distancing and mask wearing, and the West refuses to adopt either.

If it’s Asian, Europeans as well as Americans view it as automatically either inferior or irredeemably foreign. Whatever the reasoning is, it’s an excuse not to learn. Unlike the United States, Europe has pretty good public transportation in the main cities, and a lot of domestic innovations that are genuinely better than what Japan, South Korea, and Taiwan do; thus, it can keep going on like this without visible signs of stagnation. Nonetheless, what Japan has, and to some extent the other rich Asian countries, remains a valuable lesson, which good public transport advocates and managers must learn to adopt to the European case.

Urban rail and regional rail: network design

Tokyo and Seoul both have stronger S-Bahn networks than any European city. This is not just an artifact of size. Paris and London are both pretty big, even if they’re still only about a third as big as Tokyo. In Tokyo, the infrastructure for urban and regional rail is just far better-integrated, and has been almost from the start. Among the 13 Tokyo subway lines, only three run as pure metro lines, separate from all other traffic: Ginza, Marunouchi, Oedo. The other 10 are essentially S-Bahn tunnels providing through-service between different preexisting commuter lines: the Asakusa Line connects the Keisei and Keikyu systems, the Hibiya Line connects the Tobu Skytree Line with Central Tokyo and used to through-run to the Tokyu Toyoko Line, etc.

This paradigm of cross-regional traffic is so strong that on lines that do not have convenient commuter lines to connect to, there are suburban tails built just to extend them farther out. The Tozai Line hooks into a reverse-branch of the Chuo Line to the west, but to the east has little opportunity for through-service, and therefore most trains continue onto an extension called the Toyo Rapid Railway.

On the JR East network, there are a few subway connections to, but for the most part the network has its own lines to Central Tokyo. This is an early invention of mainline rail through-running, alongside the Berlin S-Bahn; the Yamanote ring was completed in 1925. Further investment in through-service since then has given more lines dedicated tracks through Central Tokyo, for capacity more than anything else.

The issue is not just that there are many through-running lines. Tokyo has 15-16 through-running trunks, depending on how one counts, and Paris, a metro area about one third the size, will soon have 4.5. It’s not such a big difference. Rather, Tokyo’s through-running lines function well as a metro within the city in ways the Berlin S-Bahn, the Paris RER, the Madrid Cercanías, and any future London Crossrail lines simply don’t.

What’s more, future investment plans in Europe do not really attempt to turn the commuter rail network into a useful metro within the city. Berlin has a strong potential northwest-southeast S-Bahn route forming a Soviet triangle with the two existing radial trunks, but it’s not being built, despite proposals by online and offline advocates; instead, current S21 plans call for duplicating north-south infrastructure. In Paris, the RER C doesn’t really work well with the other lines, the RER E extension plans are a mess, and most of the region’s effort for suburban rail expansion is spent on greenfield driverless metro and not on anything with connections to legacy mainlines. In London, the subsurface Underground lines are historically a proto-S-Bahn, with some mainline through-service in the 19th century, but they are not really used this way today even though there is a good proposal by railfans.

While Europe generally does the longer-distance version of regional rail better than Japan, the vast majority of ridership is S-Bahn-type, and there, Japan absolutely crushes. What’s more, Korea has learned from Japan’s example, so that the Seoul Subway Line 1 is an S-Bahn and many other lines are very long-range; Seoul’s per capita rail ridership is much lower than Paris’s, but is increasing fast, as South Korea is a newly-industrialized country still building its infrastructure at low cost to converge to Western incomes.

Rolling stock

Tokyo outdoes the closest things to its peers in the West in S-Bahn network design. Japan is equally superior when it comes to the rolling stock technology itself. It has all of the following features:

  • Low cost. Finding information about rolling stock costs in Japan is surprisingly hard, but Wikipedia claims the 10000 Series cost 1.2 billion yen per 10-car, 200-meter train, which is around $60,000/meter, compared with a European range that clusters around $100,000.
  • Low weight – see table here. European trains are heavier, courtesy of different buff strength regulations that are not really needed for safety, as Japanese trains have lower death tolls per p-km than European ones thanks to accident avoidance.
  • All-MU configuration – Japan has a handful of locomotives for passenger service for the few remaining night trains, and runs all other trains with EMUs and sometimes DMUs. Parts of Europe, like Britain, have made this transition as well, but Zurich still runs locomotives on the S-Bahn.

The one gap is that Europe is superior in the long-range regional rail segment with a top speed of 160-200 km/h. But Japanese trains are better at the more urban end up to 100 km/h thanks to their low cost and weight and at the high-speed end of 300+ km/h thanks to low cost and weight (again) and better performance.

Shinkansen equipment is also more technically advanced than European high-speed trains in a number of ways, in addition to its lower mass and cost. The N700-I has a power-to-weight ratio of 26.74 kW/t, whereas European trains are mostly in the low 20s. Japanese train noses are more aerodynamic due to stringent noise regulations and city-center stations, and the trains are also better-pressurized to avoid ear popping in tunnels. As a result, the Shinkansen network builds single-bore double-track tunnels hardly bigger than each individual bore in a twin-bore European rail tunnel, helping reduce cost relative to Japan’s heavily mountainous geography. The EU should permit such trains on its own tracks to improve service quality.

The Shinkansen

The Shinkansen works better than European high-speed rail networks in a few ways, in addition to rolling stock. Some of it is pure geographic luck: Japanese cities mostly lie on a single line, making it easy to have a single trunk serve all of them. However, a few positive decisions improve service beyond what pure geography dictates, and should be studied carefully in Britain, Germany, and Italy.

  • Trains run through city centers with intermediate stops rather than around them. This slightly slows down the trains, because of the stop penalty at a city, and sometimes a slightly slowdown for an express train. This is especially important in Britain, which is proposing an excessively branched system for High Speed 2, severely reducing frequency on key connections like London-Birmingham and London-Manchester.
  • Trains run on dedicated tracks, apart from the Mini-Shinkansen. This was enforced by a different track gauge, but a sufficiently strong national network should run on dedicated tracks even with the same gauge. This is of especial importance in Germany, which should be building out its network to the point of having little to no track-sharing between high-speed and legacy trains, which would enable high-speed trains to run more punctually.
  • Train stations are through-stations (except Tokyo, which is almost set up to allow through-service and errs in not having any). If the legacy station is a terminal, like Aomori, or is too difficult to serve as a through-station, like Osaka, then the train will serve a near-downtown station a few km away, like Shin-Aomori 4 km from Aomori and Shin-Osaka 4 km from Osaka. Germany does this too at Kassel and has long-term plans to convert key intermediate terminals into through-stations, but France and Italy both neglect this option even when it is available, as in Tours and Turin.
  • Rolling stock is designed for high capacity, including fast egress. There is no cafe car – all cars have seats. There are two wide door pairs per car, rather than just one as on the TGV. There is full level boarding with high platforms. Express trains dwell even at major stations for only about a minute, compared with 5 minutes on the TGV and even slower egress at the Paris terminals. Trains turn at the terminals in 12 minutes, reducing operating expenses.
  • Pricing is simple and consistent, without the customer-hostile yield management practices of France, Spain, and much of the rest of Europe.

Reliability

Japan is renowned for its train punctuality. As far as I can tell, this comes from the same place as Switzerland: system design is centered around eliminating bottlenecks. Thus it’s normal in both Japan and Switzerland to leave some key commuter lines single-track if the frequency they run allows timed meets; both countries also employ timed overtakes between local and express trains on double track.

Where I think Japan does better than Switzerland is the use of track segregation to reduce delays. Captive trains are easier to control than highly-branched national rail networks. In Switzerland, there is no room for such captive trains – the entire country has fewer people than Tokyo, and the city of Zurich has fewer people than many individual Tokyo wards. But a big country could in effect turn long lines into mostly separated systems to improve punctuality. This goes against how the S-Bahn concept works in the German-speaking world, but the Tokyo and Seoul lines are in effect metros at a larger scale, even more so than the RER A and B or the Berlin S-Bahn. France, Germany, Spain, Italy, and Britain could all learn from this example.

The heavy emphasis on punctuality in Japanese railroad culture has been blamed for a fatal rail accident. But even with that accident, Japanese rail safety far surpasses Europe’s, approaching 80 billion passenger-km per on-board passenger fatality where Europe appears to be in the low teens.

Is this everything?

Not quite. I will write a companion piece about what Asia can learn from Europe eventually. For one, East Asia appears to optimize its rail operating culture to huge cities, much like France and Britain, and thus its smaller cities have less per capita rail usage than similar-size Central European ones; on this list, compare Fukuoka, Busan, and Sapporo with Stockholm, Prague, Vienna, Munich, Stuttgart, Rome, Frankfurt, Barcelona, and Hamburg. Europe is also better when it comes to 160-200 km/h regional rail.

However, the bulk of intercity rail traffic even in Europe is on high-speed trains, an area in which Europe has more to learn from Japan than vice versa. Similarly, the bulk of individual boardings on trains are on metro and short-range S-Bahn trains even in the German-speaking world; there there is a lot of learning to be done in both directions, but at the end of the day, Tokyo has higher rail usage than Paris and London.

130 comments

  1. Alon Levy

    P.S. I talk about Japan and Korea and not China because, frankly, China’s passenger rail network is not that amazing. The construction costs are nothing to write home about (and neither are Japan’s), the HSR rolling stock costs are intermediate between Japan and Europe’s, the regional rail is truly awful, the HSR station siting is just bad (lol Hongqiao). The one positive aspect is that China manages to have very high passenger rail density and very high freight rail density on the same network, whereas North America only does freight and Europe and Japan only do passengers; that said, because legacy passenger rail in China is so weak, what it really has is a legacy freight rail network and an overperforming high-speed rail network running mostly on separate tracks, with none of the branch lines that characterize older industrialized countries.

    • Eric

      China’s metro lines that extend far out into the suburbs may effectively be a good example of the “through running regional lines which run as a metro in the city” which you are praising here. A good example of this would be Shanghai.

      • Alon Levy

        They’re very slow, though. The average speed is in the high 30s, whereas Tokyo’s rapid lines are in the 50s and the Tokaido Main Line hits 60 between Odawara and Tokyo.

        • SB

          Aren’t those rapid lines fast because there are local lines with shorter stop distance right next to them?
          The rapids are fast because they can skip stations stopped by the local trains. And where local lines end and rapid trains makes all stops, stop distances are wide because stations are in less densely populated outer regions of the Great Tokyo.
          would consider local lines (Keihin–Tohoku, Chuo-Sobu) S-bahns and rapids (Tokaido Main, Chuo-Rapid) Regional-Express.

          • Alon Levy

            Yeah, but it matters that Tokyo (and to some extent Seoul) made sure to widen its mainlines to four+ tracks to make this possible, whereas Shanghai and Beijing have not.

          • michaelrjames

            This must be partly due to most of Tokyo being on viaducts versus bored tunnels in Shanghai? I don’t know the history of Shanghai’s Metro and agree that it looks like an error, or poor compromise, because it was already a mega (or super-mega) city when they began. The problem of frequent stops on such long lines, must have always been obvious but they went and did it anyway. (Unless there are station bypasses that I am unaware of?)

          • Alon Levy

            Tokyo’s four-track mainlines are all inherited from legacy operators, not the metro, but Shanghai could have had room for this recently, using the legacy lines into Shanghai Station and Shanghai South and maybe also building a four-track viaduct in Pudong.

          • Eric

            I suppose Shanghai and Beijing can dig express lines later on (though Beijing’s current attempt at an express line, line 19, seems badly done)

      • yuuka

        The same city that has Line 16 and 15, which require you to take a metro line out to buttfuck nowhere before getting on those commuter lines?

        Arguably they aren’t as bad, the worst example is with Beijing’s Yanfang line. Chinese planners have this obsession with matching capacity to demand, they end up forcing a lot of gauge changes and transfers.

        • Eric

          I think you mean 5 and 16 (and 17). Yes that is stupid. But there are many more lines which do run through the city center.

          • yuuka

            The thing is that lines 1 and 5 share a terminal station, so wouldn’t it have made sense to extend line 1 instead, with lowered frequencies past that point?

            The most insulting is Beijing Line 1/Batong line, where they use the same train size, power and gauge, plus the platforms are literally next to each other. That should have been through service from day one… Line 5/Yizhuang comes a close second.

  2. Martin Kolk

    I would add transparent pricing, where every trip is charged per distance and there are no massive bulk discounts benefitting some customers over others. I think this is considered fair by the citizens in most Asian countries and aligns the operators’ interests with the interest of the customers in a much better way. It is my belief that in most western countries giant cross-subsidies benefitting some groups of travelers over others weakens the general public stake in investment in the system. Just as having an efficient network means better utilization of resources, aligning pricing to the actual cost of trips just leads to more efficiency in the system. It is fair to have large subsidies in a public transit network as it is an important public good (50% of the running costs, excluding investments, is not unreasonable), but then just add them as a fixed percentage subside on the “market cost” of every single ticket, where the original ticket prices are as close as possible to what the trip actually cost to provide. It also makes investment in new infrastructure easier to calculate, and less prone to grandstanding politicians. Public transit should not be treated as a charity for some groups in society, but as an optimization puzzle where utility for the greatest amount of people should be maximized.

    • michaelrjames

      aligning pricing to the actual cost of trips just leads to more efficiency in the system

      … from the sole point of view of “financial performance” … whatever that may mean. Not in terms of maximising passengers, or optimising the actual use of an expensive bit of government-funded infrastructure, nor getting people out of cars (and reducing the negative impact of their pollution and lifelong healthcare from accidents) or airplanes, or reducing the trade balance by reducing oil imports … or the positive effects of agglomeration by bringing cities closer together.

      A mass transit system should not be run by the government (ie. our proxies) to optimise revenue as its sole or major objective.

  3. Coridon Henshaw

    My understanding is that Europe’s fondness for dual-bore tunnels in new construction is driven by safety concerns related to derailments and fires. As the Mont Blanc (road) tunnel fire demonstrated, single-bore tunnels are deathtraps in the event of a major fire. As three major fires in the Channel Tunnel have demonstrated, fires in multi-bore tunnels are very dangerous but are survivable.

    Whether the Japanese single bore approach would provide better value for money depends on the constructing government’s views on preventable deaths and the statistical value of lives.

    • Nilo

      Road tunnel fires are much more common than rail tunnel fires. Notably no HSR trail carries a giant extra fuel source for the fire to feed on with it, and given the fact the Japanese have far fewer rail fatalities by any metric compared to Europeans it seems silly to say they value life less.

      The chunnel tunnel also isn’t comparable to most rail tunnels in that it’s incredibly long and underwater. Notably the Japanese do use a multi-tunnel approach in the Seikan tunnel for similar reasons.

      • Coridon Henshaw

        According to the OECD[1], the mean value of a statistical life used for health planning purposes in Japan is %30 of that used in Sweden and Norway and only 10% of that used in the UK. The OECD was not able to generate cross-country comparison figures for the mean VSL for transport design. However, if the pattern in the health VSL can be generalized, then that would indicate that the Japanese have a much greater tolerance for fatal risks than Europe and the UK.

        [1] PDF page 53: https://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/epoc/wpnep(2008)10/final&doclanguage=en

        This is a rather old source but it is the best I can find on the public Internet. More recent data might be available from the Economist Intelligence Unit if anyone here has access to it.

        • Alon Levy

          Judging by the history of rail accidents in the last 30 years, it’s the exact opposite – Europeans are more cavalier about human life than Japanese, and Americans far more so. After Amagasaki, there was a lot of English-language reporting criticizing Japanese culture for its indifference to human life, and after Wenzhou there was even more reporting about Chinese culture. No such reporting haunts Germany despite Eschede and multiple smaller fatal accidents, because the racists who spread the Asians-don’t-value-life memes stereotype Germany differently from how they stereotype East Asia.

          The problem with VSL calculations is that beyond insurance, they need to be translated into policy and planning somehow. The US VSL is pretty high in your link, but in practice the auto insurance requirements are so low that the VSL for traffic is more like $50,000 than $17,465,000; American transportation planning likewise does not treat 40,000 car accident fatalities per year as a 2005$700 billion a year deadweight loss, around 5% of the US’s 2005 GDP.

          • Onux

            To be fair, Coridon wasn’t referring to safety outcomes, but comparing VSL to suggest why Japan might favor/allow construction that is objectively “less safe” from a design perspective. This of course is not the same as less safe from an operational perspective; as you’ve shown the Japanese rail network is far safer overall with fewer accidents. Yet if an accident occurred in a dual track tunnel and involved an oncoming train the result would be worse than in a single track.

            This paradigm of building more infrastructure at less cost via looser safety standards but then operating it safer is worthy of study itself, a sort of “organization before concrete” for safety not planning. Any ideas why/how the Japanese manage a network that is so densely trafficked, punctual and yet still safe? Better training? Better signals/early adoption of PTC? More regulatory oversight? Less regulation, more responsibility on operators?

      • po8crg

        One additional reason for the Channel Tunnel to have unusually strict fire safety rules is that the trains do carry giant fuel sources for a fire to feed on, in the form of all the cars and trucks being carried by the car-carrying trains.

        I believe that all of the major fires in the Channel Tunnel have been on the road-vehicle-carrying trains, not on the passenger trains.

        • Herbert

          Switzerland and Austria have several train lines that carry motorcars through tunnels. I’m not sure they won’t eventually offer similar things for the Gotthard Base Tunnel or the Brenner Base Tunnel…

          • po8crg

            Cars, sure. But lorries/trucks (which were involved in all three of the Channel Tunnel fires – 1996, 2006 and 2008) are something that only that tunnel carries; on land, the container is generally just transferred onto the train and then onto a different lorry at the other end, rather than loading the lorry on a train.

            I’m not sure why/whether cars are safer, but HGVs do seem to explain why the Channel Tunnel has a notably worse safety record than other tunnels.

          • Herbert

            Ehm… Google “rollende Landstraße“. As the Brenner Base Tunnel is largely being sold on its ability to relieve narrow alpine valleys off the truck traffic and as Germany is dragging its feet on building proper railway feeder lines, I am virtually certain they’ll have trucks on trains through that tunnel. Why wouldn’t they?

    • Marco Chitti

      The problem with tunnel safety regulations is that can drive costs up very quickly for minimum overall benefits.
      The existing high speed network in Italy was built with single bore tunnels for the 150 or so km it has. Some tunnels are 18 km, single bore, with parallel smaller escape tunnels. The new 25 km tunnel on the Genoa-Milan line is twin bore and much costlier for that reason.

      Besides, 2004 EU regulations are also imposing huge improvements in existing tunnels to upgrade to reach excessive safety standard (1event/10billion passages), that are hardly required for road tunnels, that are way more dangerous. These rules are good in theory but bad in practice. Italy, with almost 2100 km or rail tunnels, more than any other European country, is paying several billion to upgrade it, money that would be better spent elsewhere to improve the network.
      Copying Japan, that has a similar very mountainous topography, would be better for Italy than copying Germany.

  4. Rover030

    The HS2 network is really weird. I guess it comes from the same idea as existing mainlines that bypass cities, so that each major city has its own trains to London which is theoretically the fastest because there are no major cities to pass through.

    But considering that the narrative on HS2 has changed from speed to capacity, and that they are building many tunnels anyway, they might as well go straight through Birmingham and Manchester. Then those cities get full service to points further north as well, and it would obviously be much more convenient to be able to take a train every 3-5 minutes than having to wait for the 20 minute frequency train (or 30-120 in the case of Birmingham) that goes to your specific destination branch.

    From what I’ve read they’re assuming long turning times as well. 7 platforms for 9tph at Curzon Street and 11 platforms for 15-18 tph suggest so as well. That should make the benefits of through service even stronger, because the trains would be utilised more.

    • fjod

      Current speculation seems to be that Manchester will be getting a through station that allows east-west running, in conjunction with a higher-speed line from Liverpool to Leeds. Not that that’s of huge help, however – as traffic going north-south would benefit more from a through station.

        • fjod

          that defeats the object of a through station for HS2 traffic going northwards through Manchester

          • adirondacker12800

            I got as far as seeing they are building two lines on the low demand part and one line on the high demand part and decided that it’s something in twice boiled water or maybe it’s the Marmite. Trains can go east/west on a north/south line or vice versa and stations can have north/south platforms and also have east/west platforms. I got as far as seeing they want to build one line on the high demand part and two lines on the low demand parts.

      • adirondacker12800

        …and Liverpool, Manchester and Leeds have more people than all of Scotland…

  5. Eric

    In this post you highly praise regional rail systems that through-run with many stops in the urban core. Isn’t that exactly what BART does?

    • Alon Levy

      Sort of. It’s not really laid out in a way that’s useful for urban rail; the contrast is with the Washington Metro, or better yet a counterfactual Washington Metro in which the Red Line would’ve run on the Brunswick Line tracks, the Yellow on the VRE tracks, the Orange on the disused now-trail going west of Vienna, etc.

    • Onux

      The problem with looking at BART as a success is that it is built like a metro system (with attendant cost) but has the regional rail typology. The counter-factual in the Bay Area is “BART” running on the current corridors for Caltrain, Capitol Corridor, San Joaquins, ACE etc. (tracks which all existed when BART was built) and the center city tunnel only running from Bayshore to W. Oakland (most of the current SF tunnel, except turning east to the Caltrain tracks instead of west to Daly City). Instead BART took some legacy corridors and expensively rebuilt them so they can only run BART, or built duplicate, parallel corridors at great cost.

      The other issue is that BART has good urban stop spacing for regional rail, but not for urban rail (compare as noted to DC Metro). Since it tries to be the area’s subway, there is no good actual subway network for SF or the East Bay. If Muni Metro had a Van Ness/Mission subway it would be different (BART would be the express tracks between it and the Market subway) but as it is the lack of stops at 20th and 30th Streets in the Mission (or stops actually on Mission St in the Excelsior) means Muni has to run rapid AND express busses that exactly parallel the BART line through SF to handle demand. Oakland has similar gaps up to over two miles through dense neighborhoods like Pill Hill, Temescal and San Antonio.

      • Lee Ratner

        In an ideal world, the Key system would have been turned into a light rail system for the Berkeley-Oakland area rather than being ripped up and replaced wholesale by BART. San Francisco should have also built a muni light rail subway under Geary rather than replace it with buses. There should also have been more North-South light rail lines for Muni, so you don’t need to go downtown to get to another neighborhood.

  6. df1982

    HS2’s London-centric orientation is very obvious. The high overall frequencies mean you still get acceptable frequencies for the major cities to London, but Birmingham-Manchester, for instance, ends up being poorly served.

    Alon, for rolling stock costs, is the price advantage of Shinkansen trains offset by the fact that, if I recall correctly, they have a much shorter life expectancy than European produced trains (ca. 20 years vs 30-40 years)? Or are you already factoring this in?

    • Alon Levy

      I’m not factoring the short shelf life, no. But two important things to note:

      1. Mid-life refurbishment is not free.
      2. The trains end up more modern, so fleet renewal (for energy efficiency, performance, etc.) happens faster.
      3. Spending 60% of the money on a train that lasts around half as long is a good deal when interest rates are positive (which, granted, isn’t really the case in either Japan or Europe…).

      • Herbert

        Do you think DB’s extending of the life of its ICE 1 & ICE2 fleets is due to the lack of readily available trains for the German market, in response to or anticipation of rising demand (German press was surprised both by the success of VDE8 and the lowering of the tax rate on train tickets and the ICE1 fleet was refurbished to include ECTS), a penny wise pound foolish decision or something else?

      • df1982

        There are environmental benefits to having longer-lasting rolling stock (i.e. less scrap metal), although these would be offset by lower fuel efficiency and reliability. Do you have any info how much mid-life refurbishment tends to cost?

  7. Diego Beghin

    Does it ever make sense to have such close stop spacing as in Paris? Even if you could through-run regional lines there, you wouldn’t because it would be so slow. On the other hand, it is quite nice to have such a dense network in Paris proper. They just need to expand the RER a bit to pick up the slack for longer-distance travel.

    • Alon Levy

      It’s useful for connections – Tokyo is the #2 city in the world in the number of missed connections on the subway. But the RER still misses a bunch of Métro connections, especially the RER B-D tunnel. Paris is programming new missed connections in e.g. the M14 extensions.

      The Seoul method of just building a four-track mainline with local and express trains rather than kludging it is useful here; Seoul has missed connections, but not many – just 8, of which only one is a total miss (1-6), one is a seam with a nearby connection (1-U), and the rest are reverse-branches whose main lines do connect (2-5, 2-B, 2-GJ*2, 6-GC, 5-GJ).

      • Herbert

        When the metro started construction, Paris had one of the longest streetcar networks in the world. Streetcar have short distances between stations to provide good coverage. Historically that’s the main reason the distances between stations on the Berlin, Munich, Nuremberg and Hamburg subways are comparatively short. Of course there may be other reasons at play in Paris, but in Germany no city has a shape that it would make much sense to average two kilometers between subway stops…

        • Alon Levy

          New York, and I believe there are 43 misses, counting the local and the express separately if both are missed but counting neither if there’s a local transfer (as at Columbus Circle or 51st/Lex).

          • adirondacker12800

            Yeah well, if the express stops a lot because railfans think it would be kewl for it to stop there, it ceases to be so express-y.

          • michaelrjames

            if the express stops a lot because railfans think it would be kewl for it to stop there, it ceases to be so express-y.

            Exactly. And why the Paris-RER (especially the new lines RER-A, D, E, compared to the old -C & -B) have relatively few stops.
            In fact I am going to guess that just like with Paris-RER* being built almost 80 years after the Metro, many of NYC’s “missed” connections are due to ad hoc construction of (private) lines that eventually turned into an imperfect network.
            (*The only entirely newly-built RER stations inside Paris are Chatelet, Nation, St-Michel-ND, CDG-Etoile and Auber.)

          • Nilo

            No. Actually the IRT and the BMT had exactly two missed connections. One is Julius Livonia way out at the end in Brooklyn between the 3/L the other is Whitehall and Bowling Green, between what become higher in Manhattan the BMT broadway line, and the IRT Lexington Line.

            The IND the city operated system intentionally missed connections in its war to drive the IRT and BMT out of business. Alon’s cataloguing here is excellent. https://pedestrianobservations.com/2018/03/13/missed-connections/

            One must remember that the system, though privately operated via concession, was largely publicly designed (some of the Brooklyn Els predate that design). The City of New York decided as a unit to have these missed connections to try to facilitate a more rapid takeover of the IRT and BMT.

          • adirondacker12800

            Whitehall/Bowling Green: Meh, they are more or less a block away from each other between Union Square and what is now all called Atlantic Avenue, meh. And there is Borough Hall too. Meh. South Ferry/Whitehall Street was more annoying. And that wasn’t all that bad.

    • michaelrjames

      That close spacing almost certainly played a role in the creation of the RER. The Metro cannot be extended much into the suburbs because the journey times would be too slow. Where they have been extended, eg. my old line M13 about 5km beyond the Peripherique into the southern suburbs, the stations are much further apart, and of course the train gets a chance to reach its top speed too. This is the same as to why the M14 long extension south, and the M15, M16 orbitals have more widely spaced stations.

      Alon always wants more intra-muros stations, and connections, for the RER lines but my opinion is that they are pretty good. Not only are there 33 stations but they include all the major nodes (Metro + Mainline, except Gare Montparnasse) which means, at most, one might have to take one Metro trip to connect. One could argue that Luxembourg & Port Royal are excess to RER-B, and RER-C has too many (9) along the river.

      Incidentally, the close spacing is what makes the Shanghai metro so slow. For such a giant city it seems a bit of a design error but a deliberate one in that they have the goal to make everyone no more than 600m from a station–which is claimed will be achieved when the 25 lines are complete by 2024. The lines I used on my sole visit were only 2-track and AFAIK that is the case for the system, so no faster express trains possible?
      ………………….

      Paris, a metro area about one third the size, will soon have 4.5

      Surely you meant “it currently has 4.5 and will soon have 5”, because the half-line of RER-E will become a complete line in the next year or so?

        • michaelrjames

          Hah, ok it took me a while but you are referring to Alon’s pet hate of the Nord-Chatelet tunnel shared by B + D.
          Nah. Don’t buy it. But clearly, until they finish its western extension RER-E is only half a line. Then Paris will have 5 🙂

          • Nilo

            Alon’s just talking about how many city center tunnels there are really. More fundamentally it means B/D can’t run 30 TPH each like RER A does.

          • michaelrjames

            More fundamentally it means B/D can’t run 30 TPH each like RER A does.

            Well, yes, but RER-A has about 300m pax pa while B + D combined has more (310m as of 2006, cannot find more recent). So, not a pressing need. At some point I presume they will update B and D with the newest duplex trains (used by A) which will increase capacity. The airport express will also remove stress and trains from RER-B. Also, GPX will probably reduce the stress on such lines converging on the centre of Paris, so maybe never needed … Of course RER-E will take a lot of stress off RER-A when the western extension opens because it is taking over the Passy branch of A.

          • Alon Levy

            The combined RER B+D tunnel is consistently more crowded at rush hour than the RER A tunnel (citation: I took both pretty recently). Egress capacity – keeping dwells at Gare du Nord and Les Halles to 60 seconds at the peak – is as big a challenge as on-train space (and the trains have commuter-style seating rather than metro-style seating).

          • Herbert

            When did Paris last have a realistic chance to build something like Berlin Hauptbahnhof?

          • michaelrjames

            Alon beat me to it, though he possibly meant it in a negative sense while I believe it was an amazing bit of planning. With its recent cosmetic upgrade (well, above ground, La Canopée) I think it is even better. Of course, despite the massive station fed by two new massive tunnels and three RER lines and 5 Metro lines, it still has been close to being overwhelmed with transit pax. That’s a sign of success rather than failure, IMO.

            Meanwhile Paris is doing a massive rejuvenation of Gare du Nord and Gare de l’Est, which incidentally has close to double pax flow of Berlin Hauptbahnhof. Also building a giant new station, Gare Pleyel which will have TGV, RER, Metro, tramway interchange. Maybe when Pleyel takes some pressure off Nord and Chatelet they might get around to building Alon’s missing tunnel!

          • Nilo

            I guess the question is Michael, apart from the CBD concerns, which get into your aesthetic objection to high rises, why would it not have been better to rebuild Les Halles with three RER tunnels and a mainline through connection to the TGV Network, than what they did?

            Sometimes we can’t anticipate how much demand there will be, but it sure seems like a through running TGV station would have been welcomed even back then.

          • michaelrjames

            Of course in a perfect world with genius prevision, they would have done that.
            But try to remember the enormous financial outlay that RER-A and RER-B projects were at the time. As Alon says, they remain today among the most expensive transit tunnels + deep stations ever built. And look at Chatelet-Forum-les-Halles and what an audacious project in the heart of ancient Paris it was. In the 1970s.

            Also remember that D wasn’t planned, not even in crayon fantasies, at the time because a second north-south line was originally pencilled in to be west of the centre (I presume thru St Lazare and Montparnasse) to form the H plan.

            Then, no one could have predicted such ginormous passenger loads. Despite all this, and the grumbles about crowding etc, A, B & D actually do the job, and way above whatever was planned.

            So, I don’t disagree with Alon, but find the insistence on perfect planning 42 years ago a bit … capricious, is all.
            Surely there must be plenty of similar examples around the world, ie. with imperfect sharing of tunnels by different lines? I’d also say that the 8km tunnel to complete RER-E (and relieve A & M1) makes sense re prioritising over an doubling of the Nord-Chatelet tunnel.

          • michaelrjames

            I didn’t address the TGV thru-running question.
            As I’ve said before, I am not sure it serves much of a purpose in a mega-city like Paris which is the origin/endpoint of most TGV & long-distant journeys. It’s not like so many secondary cities where thru-running makes sense. And Inteconnexion-Est serves that purpose for north & east pax headed south &bwest, while being much less expensive than building a tunnel under Paris (reusing lots of existing rail ROW) plus serving CDG and Euro-Disney.
            Perhaps I’m wrong but I reckon the number of pax arriving by Eurostar or Thalys at Gare du Nord and wanting to immediately hop on another TGV heading south at Gare de Lyon or Montparnasse, is trivial.

            Also, I just received my copy of Meunier’s On the Fast Track, a history of the origins of the TGV and see that its progress to reality was a rather lucky convergence of politics, industrial strategy and timing. (Doubtless I am guilty of having a bit of a rosy picture of wise planning by the French.) It only just escaped closure when the sceptic Giscard d’Estaing took over from the enthusiast Pompidou. Given the concurrent mega-project of the RER-A & -B, it is all a bit of a miracle. (So, good that Pompidou plans to demolish half of Paris to build hi-rise but kudos for his enthusiasm for progressive and ambitious infrastructure, part of his ambitions for industrial strategy.)

            My only mild objection to Alon–and other transit geeks–is they tend to ignore the hostile obstacles that have to be overcome in the real world to get any project off the drawing board. Especially in the contentious western democratic world unlike, say, Japan, Korea or Singapore (or China)! Simultaneously demanding perfection is just a bit too much.

          • Alon Levy

            Wait, in what way is France more contentious than South Korea, where mass protests forced the political system to impeach and remove a corrupt president? Or Taiwan, where the two-party system disagrees not just on tax rates and social welfare but also the character of the regime (independent democracy vs. part of the PRC)? I guess Japan doesn’t have much partisan competition, but neither did Italy from the war right up until mani pulite.

          • michaelrjames

            I was comparing 40 to 50 years ago when both Korea and Taiwan were nothing like the democracies they are today. Japan and Singapore need no explanation.
            Even so, today all these manage to make broadly bipartisan decisions on the big stuff, don’t you think?
            In fact I thought the same broadly applies to France–which explains their success in long-term infrastructure–compared to the brainless and toxic automatic partisanship we see in the Anglosphere.

          • Alon Levy

            No, I don’t think this at all. In South Korea, conservative media keeps looking for reasons to attack Moon’s government and likes demagoguing against Chinese-Koreans. In Taiwan, the KMT seems happier with an enemy state than with the DPP.

            And in France, every party except LREM immediately sided with anti-Semitic rioters because the rioters were against Macron. LR eventually relented and had the shame to lie that the rioters were Islamists rather than white supremacists, but that took 2 months.

          • michaelrjames

            You’re talking about something else. Like I said/implied, messy democracies bring vicious partisanship to the foreground (which are always there but suppressed under their former strongman systems) but still manage to support the same broad national interest policies.
            Taiwan (which I didn’t mention in my first post on this) has the special case of the KMT. I have zero idea of why, of all groups or parties, they support reunification. The fact that they seemingly are now willing to sell out most of their own citizens is at one with their duplicitous atrocious history. Once, I would have said if the PRC took back Taiwan and lined all the old KMT guys against a wall … I would not have objected. Bunch of thugs, gangsters and anti-patriots (20s thru 40s China). Though I suppose there is none of the originals still alive, when it comes to that particular bit of history I am a Maoist. I find it impossible to believe the PRC party elders would ever forgive them, which I perfectly understand.

          • adirondacker12800

            There are people who have been telling each other tall tales for decades and people who can agree on reality. The ones agreeing on reality seem partisan but they aren’t. Reality is. it’s not partisan. And reality is about to get very very real in a few ways if it already hasn’t and we just haven’t noticed yet.

      • Nilo

        To highlight Alon’s point about South Korea here is an example. When Lee Myung-bak was mayor of Seoul he shifted money from subway construction towards completely rebuilding the bus system, and faced monetary difficulties in doing so as the national assembly was controlled by parties opposed to him. This is to say though South Korea invests in mass transit, what it should invest in and what gets priorities are certainly not agreed upon by all.

        • michaelrjames

          Those examples are not convincingly counter to my point, which I think you actually agree with but are being obstreperous.
          I’m talking about the gutting of any in-house planning & execution capabilities inside government and the erratic support & withdrawal of support that we have seen across the Anglosphere over the past 4-5 decades. Maggie Thatcher not only closed down BR’s research into tilt trains but got them to sell off the IP cheaply to the Italians. She passed specific legislation to prevent any government money going towards the Channel Tunnel or what became HS1. When Tony Abbott became Australia’s PM in 2013 he blatantly declared not a dollar of federal money would go to rail projects (he touched Maggie’s skirt when he was a Rhodie in Oxford and this was assuredly his adoption of her policy, verbatim) not to mention trying (but mostly failing!) to shut down anything but coal-fired electricity generation. His inheritors seriously consider forcibly buying a 50-year-old coal power station (Liddell) because its private owner says it is going to close it earlier than planned due to its lousy economics.
          … And Trump. Rejuvenate coal. Fantasy trillion dollar infrastructure ‘plan’. Petty withholding of project funds from blue-tinged districts/cities/states.

          Of course any polity is going to have factionalism, sometimes vicious–though in Asia often hidden from public view (less so in the modern era)–but it hasn’t resulted in utterly erratic reversal of major policy or paralysis in policy like we see in the Anglosphere. And even less so in their glory, fast-growth phases where it was strongman rule & one-party rule.

          Incidentally, it kinda demonstrates my point that even though Giscard d’Estaing apparently would have closed down the TGV plans, its champions had managed to lock it in just before Pompidou’s death. Ironic that d’Estaing subsequently takes credit for ‘his’ far-sighted TGV program, and so much other modernisation that occurred under his (single) term. He does deserve some credit but much was long gestating. When he lost re-election to his–and the conservative establishment’s–worst enemy, Mitterrand, apocalypse didn’t happen as predicted. (Even his nationalisation/hasty re-privatising of banks doesn’t look such a bad move in retrospect; not sure what Piketty says, except that giant banks are one of the scourges of the world.) The nuclear program was only picking up steam … and continued. The reason France & Paris and its cities and towns have a coherent transport system is due to a robust civil service staffed with people who know what they are doing, and the general acceptance, if sometimes reluctantly, by incoming administrations of this status, and thus much more continuity of policy across time and changing of the guard. Almost Japanese! Well, dirigisme is shared amongst all those Tigers we have been discussing …

  8. Luke

    I’m curious about what sort of rail service category the upcoming GTX lines in Seoul fall into. They’re fast–the expected average speed on the lines is 100km/h–stops are fairly widely spaced–should average 5km or so, but that average is skewed by multiple stops within Seoul mixed in with longer-distance stops outside of it–and they’re functionally intercity lines with intracity stops in Seoul, even if they’re all within the same larger metro area. Would these count as S-Bahn style lines?

  9. orulz

    Alon, you have mentioned on this blog often that Seoul (and South Korea in general) have enviable construction costs.

    In my very limited travel experience in Seoul (1 week as a side trip while I was living in Japan) I was struck by how these cities have very wide arterial roads. Six lane roads are everywhere; eight to twelve lane roads (surface arterials with intersections, now; not freeways) are not unheard of either. Much of the metro network runs beneath these massive roads. Could this be a contributing factor to why construction is so cheap?

    • Alon Levy

      I don’t even know if Korea is building tunnels cut-and-cover nowadays. I don’t think Japan is (and I know the Oedo Line was deep-bored), but when you go on level -5, you pretty much have to bore anyway. Singapore likes boring tunnels near or underneath very wide roads and the construction costs there are very high, since the station pits are not under the street but often involve extensive demolitions of older buildings.

  10. Patrick Jensen

    The BOB project (https://www.jordaanelectronics.com/pdfonline/ip_camera/benchmarking_paddengers_in%20railways.pdf) did some benchmarking with regards to punctuality. The factors they noted echo Alon’s observations:

    “The continued belief in Japan in the value of retaining vertical integration between the infrastructure provider, capacity allocation, traffic control and operations makes the planning and control process easier compared to the situation in those European countries where infrastructure and operations have been separated. The decline in punctuality experienced in certain countries as a result of such separation have therefore been avoided in Japan. However, other factors are of equal or greater importance, since even before the European policy was implemented Japan’s punctuality was considerably better.

    • The rolling stock is far more reliable than in Europe; failures in services are rare. This is not achieved by reliance on modern techniques such as computer chips or systems, but more by simple tried and tested robust systems, preventative replacement of parts, double or triple back-ups systems, and better checks and controls
    • The infrastructure is likewise simple and robust, with preventative maintenance of vulnerable aspects such as signals and points. Maintenance is mainly carried out in nontraffic hours and is finished punctually
    • Staff is exceptionally service (if not market-) oriented, rarely absent for any reason and considers punctuality as of major importance. Drivers consider not running to time a personal failure
    • The network and timetable are designed to make optimum use of the capacity, and because of the high level of reliability timetable constructions involving passing loops, overtaking and intensive traffic on single-track sections are common. Such timetabling would be unacceptably vulnerable to delays in the European context.
    • Train services are designed with relatively little through running between lines. Staff and rolling stock remain on the same services so that delays on one line are not transferred to another. This means that passengers have to change trains more frequently but this is regarded as acceptable given the high level of reliability of such changes

    When delays do occur management pays attention to the lessons that can be learnt for the future by examining the causes in detail.”

    The Japanese way of thinking is that punctuality is a prerequisite for achieving a high degree of utilization for staff and capital, whereas in Europe the common wisdom is that punctuality and a high degree of utilization are mutually exclusive after a point. Thus Japanese operators will typically have running time supplements of 5–2% whereas European operators will pad schedules with running time supplements on the order of about 10%.

  11. Andrew Gilbert

    I found it interesting that you cited the simple and consistent pricing on the Shinkansen as a plus, even though the “customer-hostile yield management practices of France, Spain, and much of the rest of Europe” result in significantly lower ticket prices. It’s my understanding that the TGV and Ave prices are significantly lower than the Shinkansen for day-of travel, let alone far out when the European prices are even more favorable. As an example, tickets Madrid-Barcelona tomorrow are ~€70 each way, vs. the absolute cheapest deal on the Shinkansen Tokyo-Osaka (Puratto Kodama) being ¥10,700, about 20% more for a journey that’s 100km shorter. The standard price is around 14,000.

      • Brendan Dawe

        Given that they’re building a privately-financed, new-build, maglev, third mainline between Tokyo & Osaka I suppose this checks out

    • Herbert

      Plus yield management absent all other factors leads to better equipment utilization, lowering the peaks and smoothing the valleys…

      • michaelrjames

        leads to better equipment utilization, lowering the peaks and smoothing the valleys…

        Is there any good data on that. I mean I know that is one interpretation but I’ve never been convinced that it isn’t simply a means to exploit those who have to travel at a certain time. It’s not so much that fares are cheaper off-peak or whatever, rather that they are much more expensive at the most convenient times. But then my initial experience of such toxic practices was in the UK …. I never found it anything as intrusive in France (perhaps it post-dates my time there; my last visit I used a Eurail pass so would not have been subject to its whims or tyranny).
        A particularly egregious example in the UK which I am sure I have given several times on this blog, is the Heathrow Express which manages to have ten different fares! Approximately zero of the journeys are “discretionary” so it is nothing more than a highly evolved method to rip off travellers to the max (a veritable plucking of the goose while avoiding injury from fightback). In fact IIRC when first opened I think they had poor patronage due to excessive fares and had to reduce them (or fudge around with highly conditional fares) … which perfectly demonstrates both the case against them and my hatred of them. An airport express should have a single fare 24/7 (except a heavy discount for airport workers).

        • adirondacker12800

          American Airlines computerized all of their bookings in 1964. The accountants have been mauling those numbers, with computers, since. If there was something askew with their assumptions somebody would have noticed by now.
          No one forces you to take the express train. If you don’t like the prices, you have alternatives.

          • michaelrjames

            The only alternatives to the Airport Express are awful (the Piccadilly line that is crowded, very slow and no provision for luggage) or of course a very expensive (and still slow) taxi. A train from a major airport like Heathrow should be for everyone and I don’t mind if the price is above what the regular Metro would cost, but still should be reasonable and 24/7 (since you don’t have any real choice about arrival/departure times). IIRC the non-special price of Heathrow Express is close to US$30.00. You can travel between London and Paris on Eurostar for not much more.

            Re airlines: there it is more understandable because you do have more choice as to airline or time of flight or class of seat or route etc.

          • adirondacker12800

            And one of your alternatives is a different airport. It’s too bad you preferred choice isn’t cheap. Don’t go that way.

          • michaelrjames

            That is much less of a choice. As I’ve pointed out here before, in general one cannot even choose between the different airports in London, depending on your origin. I would always choose Gatwick instead of Heathrow but generally can’t.
            But as it happens, I do make the choice to avoid Heathrow (and if poss. UK entirely). Comparing like-with-like it is about $100 cheaper to fly into Paris-CDG than Heathrow which has the highest landing charges in Europe. So you can fly into Paris, arriving at about 6.30 am and spend the day in Paris then take the Eurostar to London at night (if you must go to UK) and it will not only be cheaper, or no more expensive, but will be much more enjoyable and hassle-free. If you are a boring busy-bee bizoid, you can probably just take Eurostar directly from CDG though I haven’t looked at that. Voila.

          • Herbert

            Heathrow is already the busiest two runway airport in the world by far and there’s not much more room for additional slots so they have every reason to charge high landing fees… If only the airlines got the hint and started flying bigger planes into Heathrow…

          • michaelrjames

            The real reason is that the airport is close to a monopoly and is privately owned (not even by a British company anymore) so they simply charge the most they believe they can get away with. The so-called free-market doesn’t exist (if you are flying in from US or Asia there is close to zero choice).
            Oh, and despite this, do you think these captains of the free market will pay for the third runway? It is mooted to cost $30bn, so no, the government has admitted there will have to be public funds involved. So the public have historically, and for the future third runway will, fund all this expensive infrastructure but then stand back and allow private interests to maximise their profits (not even British so who knows if they pay any tax).

            Of course the third runway is dead in the water.

          • adirondacker12800

            Don’t live someplace with lousy destinations. Too bad.

  12. Thomas K Ohlsson

    I suppose it is not relevant to the readers of this blogg, but railway freight is almost nonexisting in Japan (and Taiwan). This of course simplifies passenger traffic – but I’m not shure it is ‘best practise’…

    • Herbert

      The only country that has high rail freight and high passenger usage without Stalinist legacy or low per capita income is Switzerland. Switzerland also happens to have one of the highest per capita spendings on rail. As the old German saying goes “viel hilft viel”…

    • Andrew in Ezo

      Rail freight has always been a marginal industry in Japan, even before the rise of road transport post WW2, mainly because of geography- being a mountainous island nation, all of the major industrial cities are located on the coast, so the majority of freight traffic has and is still moved by ship.

      • Nilo

        Is this still true? Naively it would make some sense since sea freight is pretty cheap?

        • Andrew in Ezo

          Almost all bulk freight (which is the majority of tonnage) moves by ship. High value freight moves mainly by truck, with ferries transporting those trucks on certain routes (e.g. Hokkaido to Honshu). Railways maintain some container freight traffic on the Tokaido and Sanyo corridors, as well the Kyushu-Hokkaido corridor that bypasses the congested Tokaido corridor via the Japan Sea side. A few dedicated fuel oil and cement trains remain, mainly to serve customers in inland locations such as Nagano, Gunma, and the interior region of Saitama Prefecture.

  13. Herbert

    Love or hate the yield management based prices of European railroads and global airlines, they’re the one proven method to ensure a high load factor without losing revenue or risking overcrowding. If you want trains at “unpopular times” to fill up, you have to entice people onto them. Under capitalism, that enticement is money…

      • orulz

        Standing is allowed on the Shinkansen. If you buy a non-reserved ticket for a Hikari from Tokyo to Osaka, you are free to board any Hikari train, you just might have to stand between Tokyo and Odawara.

        That is what makes JR’s fixed-price policies passenger-friendly: riding the Shinkansen takes just as much planning as riding the metro.

        Let’s not also forget that the Hokuriku line (when completed) will be *another* high speed relief line between Tokyo and Osaka. I suspect this will be useful when they begin to shut down the Tokaido Shinkansen segment-by-segment for major reconstruction.

        • Andrew Gilbert

          When are they planning on shutting the Shinkansen down segment-by-segment? I have never heard of a single train line in Japan ever being rebuilt like that, because they don’t let their infrastructure corrode into dust like we do in the US.

      • Herbert

        There’s simply more demand period. Japan has a surprising number of entries in the top twenty of the world’s busiest flight routes…

    • Nicolas Centa

      It is possible to do a lot without yield management, such as static discounts for hours that are considered off peak, because those tend not to change everyday. This is what electricity companies do in most respectable countries, and phone companies also did it in France when it mattered for capacity management. It is very easy to understand by the customer and therefore very effective, much more than constantly variable prices that only the most regular users can predict.

      For the Shinkansen there are discounts for slower trains (Puratto Kodama mentioned above), packs including round-trips and hotel nights, and also ski passes or anything else relevant for the season.

      And the Shinkansen fills with foreigners using the Japan Rail Pass outside peak hours, because otherwise it is crowded.

      • Herbert

        In my experience like a “nine o’ clock or later ticket” as is common for local public transit tend to confuse the non-expert more than “this is the price for a ticket for this connection, don’t worry how it was calculated” – after all, that works in aviation.

        Besides, customers who really need whichever benefit constant prices bring are free to purchase a BahnCard 50…

    • Henry Miller

      Is that what you want though?

      For short duration trips – a couple hours – you want to be show up and go for one price – don’t think about it, just get on the system and use it. If your plans change at the last minute it is no big deal, just get a different bus/train.

      For long duration trips your customers have already planned ahead – someone is watching the kids, your clubs have an alternate to take over your duties…. If they can save a buck by taking a different trip they will plan that in and take the cheap trip. They are already planning ahead (even if they have an emergency need they will be making calls on the way) Airlines only play in this area so of course they do yield management.

      For commuters transit systems get a choice. You can do yield management, but if you do you have become a commuter only line. Your yield management goal is to get people to change their work hours, and keep the prices as low as possible overall. Your yield management goal is to get people not going to work to get around some other way, thus the complexity of yield management is a feature: people won’t understand it and so they won’t muck up your system with random trips that might by chance increase the peak.

      Your other choice is to forget about it, charge every body the same price while still covering costs, ensure that there is always space for someone to get on when they want to get on, service is frequent all day, and you don’t worry about empty space off peak. In this plan you want to encourage people to ride for not just getting to work, but also for all their other trips. You should encourage trips to the dentist (twice a year for most people), shopping, concerts, and oh you can also get to work. Because of capital costs this plan can be cheaper – you already have the bus/train it doesn’t cost much more to run it vs keep it in the station. However as distances get longer it is harder to make it work.

  14. Nilo

    “Rather, Tokyo’s through-running lines function well as a metro within the city in ways the Berlin S-Bahn, the Paris RER, the Madrid Cercanías, and any future London Crossrail lines simply don’t.”

    My thought on this is that Munich is a good counter example no? The subway network intact looks stupid without the S-Bahn tunnel running through it.

    The use of through running onto metros is of course very clever, if effectively banned in the United States. São Paulo, which already has more ridership on its regional trains than Paris has on the RER (I think, it’s close either way), seems like a logical candidate. Line 9 in particular is currently a tangential legacy route that follows a river, it seems like it could be turned into two branches that feed into a tunnel that heads north towards downtown. (see attached graphic) Simultaneously Line 3 should really act as a local to Line 11, but it doesn’t— I think due to fare barriers between metro and regional trains at the time, and its not built with cross platform transfers to Line 11. São Paulo is also building/has built two monorail lines, which seem to be in the exact opposite direction of the advice given here, since of course they’re maximally incompatible with other types of rail services.

    • Herbert

      Munich is building a second tunnel for its S-Bahn at great cost because the existing infrastructure is maxed out. Munich does not have a Ringbahn…

      • Nilo

        You’re going to have to tell me what your criticism is more explicitly; I’ve missed it.

        • Herbert

          The Munich network has the big flaw of overwhelming the main Stammstrecke with capacity upgrades needing an expensive new tunnel. Maybe a Ringbahn would help some…

    • Eric

      Sao Paolo’s metro has the second most riders per km of any metro in the world (after Cairo). And it’s in a metropolitan area of ~20 million inhabitants. For these reasons, pretty much any line they build will see heavy usage. So it is not so bad if they build the “wrong” lines now, because they can always build the “right” lines later. They cannot be blamed for failure to reuse existing ROW, because all existing ROW appears to be used for high frequency services already.

      There is nothing really wrong with the monorails (which are cheaper than underground metro, and perhaps have less visual impact than elevated metro, while carrying a similar passenger load) as long as they do not end outside downtown with a forced transfer to other transit. (Which is the current situation with line 15, though that can be fixed.)

      The highest priority in Sao Paulo, though, should be through-running of suburban rail (lines 7,8,10,11 through Luz and Bras stations). Then, it will be worthwhile to extend line 12 towards Morumbi along the lines of your suggestion (it’s worth mentioning that around Morumbi there is a large business district, kind of an “edge city”). But this extension is not necessarily a higher priority than the various other metro or monorail expansions planned throughout the city.

      • adirondacker12800

        …. reuse existing ROW, because all existing ROW appears to be used for high frequency services already.
        Give it whirl, explaining that, the next time someone blurts out to use Metro North ROW for high frequency subway. Or running short half full trains to JFK airport displacing long standing room only trains to the rest of Long Island.
        …nothing really wrong with the monorails
        Monorail looks good in the pages of a 1960s Popular Mechanics magazine. People who aren’t fascinated by gadgetbahns look around and decide “we’ll have some of the NYC subway cars, the wider ones please” or “we’ll have some of those Northeastern commuter cars”. Because there will be multiple vendors offering moderately priced stuff forever. Or something that could be on the same tracks as a PCC trolley car.

          • adirondacker12800

            That wouldn’t look all that much different if it was unfabulous thing with more conventional wheels. Perhaps like this

            which is in New York City. The NIMBYs and BANANAs came up with all sorts of objections but the Port Authority and NYSDOT built it anyway.

        • Nilo

          Isn’t the entire criticism of Metro North and the LIRR they aren’t high frequency? São Paulo Regional Railways carry 3x what the NYC commuter railways do, and run frequently rail all day, and I agree with your criticisms of the monorail. The monorail when I visited also had huge problems with frequency despite having no driver.

          • adirondacker12800

            They don’t run frequently all day but that doesn’t stop them from being at or near capacity for a few hours in the morning or afternoon. Which precludes running a subway on the tracks. All those people could attempt to drive into Manhattan but even if they made it there, there is no place to park.
            Someone did a very rough estimate of what would happen. It was something along the lines of everything north of 59th Street would have to be parking lot and then you would need a high capacity people mover on Lexington Ave and Seventh Avenue and likely something like Eighth…..

          • Tonami Playman

            And I also read complaints of bouncy ride due to shoddy construction of the monorail beams compared to other monorails. And some people already prefer to use Keikyu Airport Line instead of the Tokyo monorail due to better comfort.

        • SB

          “we’ll have some of the NYC subway cars, the wider ones please” or “we’ll have some of those Northeastern commuter cars”
          Maybe they should buy trains which doesn’t break down easily, have open-gangways, are lightweight EMU, etc.
          Emulating NYC or anywhere in US for matters regarding transit is not a good idea.

          • adirondacker12800

            And it would take five years and a lot of grief to do that. When NYC starts buying them they can buy those. It’s going reaalllllllllll well for Caltrain. In the mean time Denver is actually running trains. With guards at the grade crossings instead of signals but they are running trains.

          • Nilo

            I mean Caltrain’s problem isn’t rolling stock procurement its the fact the damn work windows are so small and the foundations so ridiculously large that its taking longer than expected to install the polls.

          • adirondacker12800

            And they have been designing a super duper extra special train, for years, that the FRA asked “Didn’t you read the regulations?” when they got the plans. In the mean time Denver decided to take some of that awful stuff from the Northeast and has been running trains with for years. And they will have multiple vendors with reasonably priced stuff to pick from forever.

          • Nilo

            The FLIRT’s they’re buying are approved, and couldn’t be running anyways because none of the system is currently electrified.

          • adirondacker12800

            I don’t know if it’s the twice boiled water or the Marmite in the U.K, and got as far as speculating there is something in the Hetch Hetchy water or perhaps organic free range alfalfa sprouts in California. Last I heard the FRA was asking “Did you read the regulations” and Denver has been running trains for years.
            https://caltrain-hsr.blogspot.com/2019/07/emergency-exit-fail.html

  15. yuuka

    On subway through-running – apart from the Ginza and Marunouchi lines (too old, too different) and the Oedo line (tiny loading gauge, they claim it saves cost), the other subway lines have through service to regional commuter rail lines. But I’d argue that it’s actually metro trains being sent up into the commuter rail system, more comparable to the DC Metro or all the Tube takeovers of NR lines in the 1950s, and on the Bakerloo/District where they still interrun even today.

    Firstly. for lines designated as “subway”, fire safety regulations require emergency escape doors at each end of the train. Secondly, to increase capacity, some companies (most notably JR) have curved-sided vehicles that are wider at waist level, so as to increase standing capacity since the seats can be moved out. These are banned in the “subway” (no idea why), which only allows the rectangular flat-sided stock.

    For some reason the above doesn’t apply to the Yokosuka, Keiyo and Rinkai line tunnels, which are classified as “railways”. But otherwise, if you want through service to the subway, you’ll need to buy new subway-compliant trains, like what Sotetsu is doing now to run through to the Mita and Namboku lines.

    Kansai area railways are different, they’ve gone to coupled sets with gangwayed cab ends, all subway compliant, and operate mixed-length services in a manner more like American commuter railways.

    On the topic of vehicle costs, the practice is, as I’ve been told, “half the lifespan, half the cost”, aka planned obsolescence, which keep headline costs low but require constant swapping of fleets. It does help keep the domestic railcar industry going though, which means what happened to Budd just wouldn’t happen in Japan. The size of the market also means they can have standard designs with small modifications for different clients, which makes it a lot easier to accommodate Keio and their weird 1372mm gauge since they just need to change out the bogies for the most part.

    What America could realistically learn from this would be to standardize on something much like the good old State of the Art Car, and Kawasaki/Bombardier/Alstom/whatever would just be selling clones of said standardized vehicle, forcing them to compete on things like cost, motor energy consumption, or lobbying politicians. Unless you’re Boston Blue Line, Chicago, or NYC A Division with unique loading gauge constraints and can’t pick from the catalog, but that would drive up your costs anyway.

    Shinkansen aerodynamics also come at a price, though, longer noses mean less seats in the head cars. The Akita Shinkansen went from 6-car E3 sets to 7-car E6 sets to maintain capacity, and the upcoming 7-car Yamagata Shinkansen E8 series have lesser seating than the 7-car E3 series they replace.

    In fact, if the stubby 0, 100, and 200 series trains are any indication, the long nose is more a workaround for legacy infrastructure first (the tunnels can be small, since the initial operating speed was just north of 200kph with not so much tunnel boom issues) and cost saving measure second.

    And as for turning trains in 12 minutes, well, they don’t have the luxury of building massive Chinese-style terminals, along with the packed timetable.

    • Andrew in Ezo

      The narrow bodied stock are required due the loading gauge of the single bore/single track tunnels. The JNR/JR tunnels as well as later builds have wider clearances or have wide twin track tunnels (shield construction) so have more leeway in rolling stock permitted. In case of evacuation, passengers can use the side doors in these wider tunnels.

  16. James Scantlebury

    The reason that the shinkansen lines at Tokyo station don’t do through running is down to a small quirk of Japanese geography – each half of the country has a different power grid – 50 or 60 Hz electrical mains frequency. JR East is 50Hz, and JR Central/West etc is 60Hz. The latest models of shinkansen are dual frequency – the Hokuriku shinkansen uses two frequencies depending on where it is…

    • Herbert

      Overcoming that should be easy. European trains need to overcome far bigger differences

    • Andrew in Ezo

      The cost/benefit ratio of having run through shinkansen at Tokyo Station is likely too high. JR Central and JR East are different companies with different operating philosophies (JR Central is running a high frequency [Nozomi 12 tph] service aimed at business travelers, while JR East is more varied, with a heavy tourist and “returning to hometown” mix of more casual travelers), and I reckon they deem the pax market from, say, Utsunomiya to Shizuoka or Nagoya does not warrant close coordination- easier to just have passengers change at Tokyo as the platforms are in the same general area of the station. The competition (i.e. air) doesn’t exist, and trains have 95%+ schedule reliability, so passengers don’t feel particularly inconvenienced.

      • yuuka

        Not only that, I’ve been told JR Central has a strict preference for 1323 seater, 16-car trains on their lines, to facilitate service recovery since all trains share the same seatmap and there’s no need to accommodate customers in the event of a stock change. Though that’s moot today since every train they operate is a N700.
        I believe that was an impetus towards getting the 500 series out to Sanyo services quickly.

  17. Pingback: Europe and Asia are not Liberal or Conservative America | Pedestrian Observations
  18. Pingback: More on the Deutschlandtakt | Pedestrian Observations

Leave a Reply to Henry Miller Cancel reply

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

WordPress.com Logo

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

Google photo

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

Twitter picture

You are commenting using your Twitter 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.