Large Urban Terminals are Unnecessary
London and Paris are both famous for their elaborate rail terminals, each with many tracks to store trains connecting the capital with one direction of the country’s provinces. This leads rail advocates to romanticize this type of station; Jarrett Walker said in 2009 that these stations make the arriving passengers feel grander for, in theory, being able to walk directly from platform to station entrance without changing levels. Earlier this month, Ido Klein generalized from London and Paris to the rest of Europe, arguing that the setup in Tel Aviv, with only through-tracks on the Ayalon Railway and only six platform tracks at Savidor Center, is inadequate. Ido is incorrect on this; the Israel Railways network needs continued capital investment in the Ayalon Railway and not in alternatives such as more terminating tracks or a bypass.
What is the situation in Israel?
The Israel Railways network, owing to Israel’s small size and lack of connections to other countries, is in effect an interregional system, in which the most common intercity trips are 60 km (Tel Aviv-Jerusalem) or 90 km (Tel Aviv-Haifa). Due to the country’s narrow width, it is entirely centered on Tel Aviv, where the four stations of the Ayalon Railway – University, Savidor Center, HaShalom, HaHagana – are on pre-corona numbers the four busiest in the country; this leads to much political distaste, as everywhere else in the Israeli economy.
The Ayalon Railway has three tracks, with plans for a fourth in the design phase; capacity is said to be around 14 trains per hour per track, even with the latest ETCS signals, and I’ve been told of at least one Israel Railways manager who disbelieved that the Munich S-Bahn could push 30 trains per hour per track. The main station in construction is Savidor Center, which has six platform tracks, but in ridership it is second to HaShalom, which is better-located in the Tel Aviv central business district and has only three tracks. Practically all trains run through Tel Aviv – only the trains to Jerusalem terminate in Tel Aviv.
Do trains need to terminate in or bypass Tel Aviv?
No.
There are plans for a bypass railway east of the city, but they’re largely for political reasons, signifying that the state supports decentralizing economic geography away from Tel Aviv. The most important city pair not involving Tel Aviv, Haifa-Jerusalem, does not get any faster via the bypass, and will be direct via the Ayalon Railway as soon as electrification is completed up to Haifa. The city pair that could most gain from a Tel Aviv bypass, Jerusalem-Beer Sheva, has its most direct route passing through Bethlehem and Hebron, which besides being in the Territories (and in Area A) have terrain from hell.
Terminal tracks are likewise useless. Demand from Tel Aviv to points north and south is fairly symmetric on both intercity and commuter rail; there was some asymmetry before the new Tel Aviv-Jerusalem line opened, but at this point Jerusalem’s one station has elbowed its way to the number four position, narrowly ahead of Tel Aviv University, and together with Beer Sheva, Jerusalem forms a fine counterpart to Haifa.
But what about Europe?
Ido’s thread goes over the largest 10 metropolitan areas in the EU and UK, and looks at their intercity rail stations, which he defines as stations serving lines of at least 200 km in length. Among those, he finds that London and Paris only have terminals (their through-stations are for regional rail), and the same is true of Milan, Rome, and Athens. However, from that point, things fall apart.
First, Madrid, Barcelona, and Warsaw have intercity rail through-stations; Ido incorrectly says they are not. Madrid Atocha is a through-station; there’s little to no AVE through-service there or at Chamartín, but they are both through-stations, and the medium-speed Alvias, which are faster than anything in Israel, run through routinely. Barcelona-Sants has AVE through-service between Madrid and France; most trains terminate, but that’s because of asymmetric demand, not because it’s inherently better. Warsaw has a through-tunnel and many through-running intercity trains listed as serving Warszawa Centralna.
Second, Ido skips over Germany. He portrays Berlin as atypical for having a through-station at Hauptbahnhof, but the Berlin way is what Germany wishes were the norm for all cities. Smaller German cities either have through-stations, like Hanover or the cities of the Rhine-Ruhr, or act as pinch-points, with through-trains coming in and reversing direction to continue onward, including Leipzig, Frankfurt, and (until Stuttgart 21 opens) Stuttgart. The pinch-point operations are as efficient as they can be, but still occupy more platform and approach slots than through-trains would.
Third, the actual practice of Paris and London is that it’s assumed people only travel between the capital and a provincial city. France and the UK do not have good everywhere-to-everywhere trains. Paris has a bypass, the Interconnexion Est, but the service quality on it is terrible: the operating paradigm is that trains that bypass Paris make every intermediate stop, which takes 5-10 minutes per station, as the TGVs are not designed for fast boarding and alighting at intermediate points but for nonstop Paris-provincial city trips.
The way forward
Province-province trips are difficult to serve with high frequency. Therefore, the best practice for them is to run through the main city if possible. Israel can do it, using the Ayalon Railway, and once electrification provides through-service from Haifa to Jerusalem, this city pair can piggyback on the higher demand of Haifa-Tel Aviv and Tel Aviv-Jerusalem to serve passengers frequently. Israel is famously small; Haifa-Jerusalem is around 150 km and 1:36 with upcoming speedups planned for electrification and other investments – trains have to run at worst every 20-30 minutes to avoid throwing away ridership, and this can only be supported if they run through Tel Aviv.
London and Paris have many rail terminals because they were huge cities in the steam era and private railroads figured they should connect the capital with one section of the country. This inherited infrastructure is a liability to both of their respective national rail networks, especially that of France, where Paris is centrally located and could be the center of a Lille-Marseille spine. Israel’s newer network lacks this seam, and this should be celebrated and form the basis of further investment.
The most important investment is to ensure that the Ayalon Railway can run at decent capacity. Electric multiple units on regional and interregional rail systems with more complexity than that of Israel Railways do much better than 14 trains per hour on each track: Zurich does 16 and is (I believe) capable of 24, Munich does 30, Tokyo does 24 on some commuter lines with comparable length to Israel’s intercity lines. This is not a problem of the signaling system, which has been upgraded to ETCS Level 2 at the same time as the electrification project. Rather, it’s a problem of how the trains are timetabled, and possibly also of infrastructure on the commuter rail branches, some of which are still single-track.
If it’s possible to cancel the Eastern Railway plans, it should be canceled. There isn’t much that’s being served on the way, and the split in frequency between Haifa-Tel Aviv, Tel-Aviv Jerusalem, and Haifa-Jerusalem trains would seriously hurt ridership on the last of the three city pairs. The fourth track on Ayalon Railway is useful, but long-term plans to go up to six tracks should be shelved – a four-track electrified line could support a large multiple of current traffic.
Pedestrian Observations 
The eastern track’s main purpose is freight. Since Ayalon railway is so crowded with trains (in Israeli terms, as you explained), there was a need to create an alternative route to connect the port of Haifa to the rest of the country. Only after the planning of this train route then (and now also) transportation secretary Miri Regev branded it as a project that will support “the periphery” over Tel Aviv.
Anyway, the eastern railway is already in construction, so, it’s a done deal.
And regarding tracks 5 & 6 in Ayalon railway, they are planned to be used exclusively by a future HSR (or “Bullet Trains” as Regev and Netanyahu like to referer to them).
There is no point in high-speed trains skipping stops in Tel Aviv, so they’d be running at the same speed as all other trains and don’t need any track segregation.
Separately, there are no destinations in Israel worth building high-speed rail to from Tel Aviv. Save that for when there’s geopolitical demand for lines to Amman, Beirut, and Damascus, and even those are close enough and in bad enough terrain they should run at 200 and not 300. Maybe Cairo, but that would require the Egyptian government to give a fuck about the economic development of the country, which it currently does not.
I agree completely, I was just saying 🙂
Excellent post Alon. Minor correction re: ETCS Level 2 – the system is not operational yet, there are delays. The current estimate is 2024 from what I’ve heard.
14tphpd is the capacity *before* the upgrade. After the upgrade it should be 17tphpd (or 34tph total)
re: the eastern track, that’s mostly for freight if we’re being honest. Freight traffic is at the moment limited to off-peak hours on the coastal main line, slowing off-peak passenger trains. The eastern track will be the new main line for freight, so it’ll free some off-peak capacity for more passenger trains, and it will allow increasing freight traffic without harming passenger service.
Passenger service on the eastern route is mostly a political gimmick, that’s for sure, but the project is still useful in my opinion.
17 tph is still pretty bad.
Then again, Berlin only runs 18 tph on the S-Bahn through-trunks, so S21 is doubling the North-South Tunnel to provide capacity that would be viable on current track infrastructure with Munich signaling quality.
17tph is horrible, given how much passenger demand there is.
In fact they’re planning on building a brand new metro line parallel to the Ayalon route, at a cost of many billions of dollars. This line would be superfluous if they could run a reasonable frequency (like 24 tph) on each of the 4 tracks that will soon be there.
Munich works (probably, I’ve got no idea how reliable it is in practice – for Paris I found a 2017 document indicating that the then nominal 30 tph peak frequency was almost never achieved in practice) not just because of modern(ish) signalling (as well as the trains being somewhat more optimised for shorter dwell times than some outer suburban or especially regional rolling stock, though that doesn’t matter for the specific comparison with the Berlin S-Bahn), but also because the three most frequented stations in the central tunnel have separate platforms for boarding and alighting, which also helps to keep dwell times down.
Sorry, got a little confused there. That document about the RER A was from 2012 actually: http://web.archive.org/web/20170811183736im_/https://www.iledefrance-mobilites.fr/wp-content/uploads/2017/04/Deliberation_no2012-0163_relative_au_schema_directeur_du_RER_A.pdf#page=38
I think most passenger services through the eastern railway will be oriented towards Tel Aviv anyway, I saw a mention of one ring service using the central part of it, the 431 railway and I’m guessing the sharon railway, and I’m guessing there’ll be another service in the northern part of it that takes either the sharon or yarkon railway into the central tel aviv section.
According to the Wikipedia track diagram, the Eastern Railway will not include a connection from Haifa to Jerusalem (only the existing connection allowing trains from Ben Gurion to go to Lod) so is there even a plan to run trains direct from Haifa to Jerusalem? Aside from the freight issue, is there benefit to the line for RegionBahn/RegionExpress type service for the intermediate cities (Elad to Haifa, Rosh HaAyin to Airport City, etc.)? If so, wouldn’t there be benefit to running those RB/RE trains to Jerusalem in addition to the through service Jer-TelAviv-Haifa (so those eastern communities can reach Jerusalem directly)?
Wouldn’t Beersheba to Jerusalem via Lod be a lot faster than a connection in Tel Aviv via Ashkelon, or even a new track segment connecting from before Ramla to Modi’in at the narrow point? You make a good point about Tel Aviv-Haifa balancing with TA-Jer/Beersheba, but is there enough demand for a direct Jer-Beer train?
The plan is a direct connection between from the beer sheva railway to the rishon letzion-jerusalem railway that’s under construction right now. As far haifa to jerusalem goes the plan as far as I’ve seen is to extend the tel aviv- jerusalem line along the coastal railway as electrification continues up the line, not going through the eastern railway
The British province to province trains aren’t as good as the province to London trains. But the cross country trains are actually pretty decent.
Bristol to Manchester by cross country train is probably still how you’d do it after HS2 is built and the journey averages ~90km/h which isn’t bad.
Also if the interconnection track in Paris was used more the service from England, Northern France, Belgium and the Netherlands to provincial France would be improved by around an hour even with the stops at the airport and Disneyland.
Plus while a lot of those places are far away between them its like 80m people so you should get pretty solid ridership. Additionally you avoid having to travel through Paris on the RER or Metro – which for anyone living in the provinces is definitely a bonus.
Also Parish Hbf or London Hbf would have a mad number of passengers. And you’d need waiting rooms etc for all those people so it would need a lot of space.
What waiting rooms? I would show up to domestic TGV trips a few minutes before departure.
I have personally crossed paris between Gare du Nord and Gare de Lyon in 30 minutes or less by Metro however Seat61 recommends a minimum of 60 minutes and ideally 90 minutes – https://www.seat61.com/changing-stations-in-paris.htm
Now maybe you don’t need traditional waiting rooms but you do need somewhere for people to wait assuming all goes well and they make the transfer quickly.
A worst-case scenario Paris Hbf, absorbing 100% of pre-pandemic traffic from the four long-distance termini (Nord, Est, Montparnasse, and Lyon), would have some 550 million passengers per annum, or about 1.5 million per day.
Before the pandemic, the JR East portion of Shinjuku Station (incidentally a through station) had 790,000 daily passenger boardings, translating to around 1.6 million total passengers per day. It’s a horrendously crowded complex, but does its job without taking up too much space – Shinjuku’s platform level measured between extremities covers around 7 ha, compared to Gare du Nord’s 11 ha plot.
One could argue that long-distance traffic needs more spacious facilities than commuter traffic. In that case, we might look at Tokyo Station. Of its 678,000 pre-pandemic daily boardings – or 1.35 million daily passengers – 27% were HSR users, and a few more were using limited expresses. Tokyo Station is 18 ha if we include the forecourt above the underground platforms, and 14 ha for the surface/elevated portion alone.
Wait, why would Paris Hbf have 550 million passengers a year when the TGV (which is most long-distance traffic) only has 120? The ridership you’re adding up from the four termini is mostly commuter trips, which are mostly on the RER at this point and not even at the terminals.
I’ve looked at the Chiltern mainline in some depth and the biggest stops are Gerards Cross, Beaconsfield, High Wycombe, Princes Risborough, Haddenham and Thame Parkway, Bicester Village/North, Oxford Parkway, Oxford, Banbury, Leamington Spa and Warwick Parkway.
So an RER service that would run all stations to High Wycombe wouldn’t really cut it – as it’s really the medium distance passengers – plus people driving in from surrounding areas – who dominate.
@Alon Levy
Which is why I said ‘worst case scenario’ – the absolute maximum overall traffic such a station could have, and how it’s still very much possible to accommodate that upper bound number in one location. Also, it enables a simpler comparison with the Japanese stations, which also have mostly commuter trips.
A big station with lots of commuter traffic is easier because most people are doing the journey regularly so will know when to arrive so they don’t arrive too early. And even if they do arrive early likely they can get an earlier than expected train so they don’t in general hang around.
If you are travelling long distance that is a lot less appealing – especially as you are more likely to have non flexible tickets that you have to offer to compete on price with the airlines. And people want seat reservations etc anyway for the most part.
@Matthew Hutton
Which is why I mentioned Tokyo Station, where the vast majority of capacity on Shinkansen and express trains is reserved seating. The point is, with a decent layout and sound operations, the scale of a hypothetical Paris Hbf would not be unrealistic.
Greatest Britishest FL000 Airways Franchises Pty Ltd may indeed wish for you to “want this”, but the reality is that actual human beings travelling between “long distance” endpoints as London and, say, exotic peripheral colonial far-off mirage-beset Leeds, might find it more “appealing” to show up at the station and just, you know, jump on the next train headed that way, the one that runs every 30 minutes.
A handful of self-important wankers may claim they want their seat reservations. And complimentary chair massages, and Platinum Executive Class loyalty perks. Wankers always need something.
Meanwhile most human beings just want to get to their destination, and don’t necessarily feel the need to carry water for the franchise model of Greatest Britishest FL000 Airways Franchises Pty Ltd
To be fair with Tokyo my understanding of the rules is that if you miss your Shinkansen you can sit in the unreserved carriage on another train on the same day. Plus it’s a terminal, so if people turn up 15 minutes early they can get straight on the train, or at worst wait a short time for the train to be cleaned.
And we can make all the rude comments we like about cheaper fixed train tickets, but they do sell and people like them.
@Matthew Hutton
“cheaper fixed train tickets”
This is bizarre to me–in Japan the flexible tickets are the cheaper ones, a seat reservation is an additional surcharge.
And this makes sense, because planes and trains have reversed incentives around capacity. With planes, you have infrequent low-capacity trips, because each departure is hugely expensive. With trains, each departure is very cheap (especially relative to the infrastructure costs). And trains need high frequency to provide sufficient convenience anyway. The combination of those factors means for a well-run railroad, capacity is abundant and cheap.
An airline has to charge you a lot for flexibility: if you could take two flights, that’s at least one seat they can’t sell (at least until the last minute), because they have to hold it open for you. For a railroad the opposite is true: it’s the *reserved* seat that they have to hold open, since otherwise it would be available as part of the overall flexible capacity.
Sure, people like cheap train tickets, but they sell because they’re cheaper, not because people like being locked into a schedule. A railroad that’s charging extra for flexibility is mishandling its incentives by applying an inappropriate pricing model.
I seem to remember the speech was “Your ticket is a contract for carriage, not a seat” back in ancient times and any coach ticket was good on any train. More people show up than there are seats, someone is going to be unhappy.
The Europeans with their fixed train cheap tickets do better than the Japanese at getting a good market share for rail over medium distances. Rail has a 57% share between Edinburgh and London for example – https://www.cityam.com/hop-on-rail-overtakes-aviation-as-preferred-mode-of-transport-on-london-edinburgh-route/. I don’t think the Shinkansen holds that market share over that distance/time even though the journey quality is better.
@Matthew Hutton
“Rail has a 57% share between Edinburgh and London for example – I don’t think the Shinkansen holds that market share over that distance/time even”
That’s an odd claim to make. London-Edinburgh is around 600 km and 4.5 hours. Distance-wise, it’s similar to Tokyo-Okayama (676 km), where the Shinkansen has a 69% share against air. The Shinkansen maintains a 60% share as far as Tokyo-Hiroshima (821 km in 4 hours). There’s no indication that ticket pricing has seriously dented its popularity, not to mention that discounted deals do exist on the Shinkansen.
Time-wise, the Shinkansen certainly does poorer over the 4 hour mark – its share on Tokyo-Yamaguchi (4.5 hours) is only around 30% – but the faster trains are covering nearly 1,000 km by that point, right about where planes begin to claw back ground delays.
German ICE ridership on the few city pairs for which I have data overperforms my Shinkansen-trained model by 20%, even with meh speeds; TGV ridership looks like it overperforms by a lot more, with the caveat that my data quality there is basically alchemy.
Certainly when I have looked at trips from my house to Southern Germany, France and Italy by rail the price has been broadly competitive with flying if not a little cheaper. This is even with Eurostar’s high fares and includes travel to/from the airport.
Whereas I am sure you can fly Fukouka-Tokyo for €200/person/direction less than the standard train price – depending of course on the true availability of discounted fares and how big the discounts are.
Maybe if they’re forced to take another train or a train on another day. If they’re lucky, they get to decide if they want to stand or reserve a seat or take another train.
How are you going to force them to take another train? Throw them off at the next station?
Japanese domestic air travel is also very convenient compared to in other countries.
Security is quick and ID checks are not required so it’s normal to show up 20-30 minutes before departure rather than 45-90 minutes. Transit to/from airports is punctual so people pad less at the origin end. Airlines themselves are relatively punctual so people pad less at the destination end. Flights between major cities are frequent so people get closer to their optimal departure time. And Fukuoka airport is 6 minute subway ride from the city center, which means air travel to/from Fukuoka is even more unusually competitive with train travel.
All in all, Japanese domestic air travel is probably over an hour faster for short haul trips, compared to domestic air travel elsewhere. While Shinkansen travel is also way more convenient than European rail travel, it certainly isn’t an hour or more faster for similar time spent on the train.
@Alon Levy
I’m hesitant to read too deeply into the ridership model’s predictions. As discussed in and under your April 20 post, even within Japan, there are many city-pairs for which it either gives a missed prediction (on both rail travel and total travel), or cannot explain disparities between city-pairs it predicts ought to be similar.
Even if the model’s predictions hold, it’s difficult to pinpoint ticket policy as the cause of underperformance, as Matthew Hutton posited. Sassy correctly points out that air travel in Japan is much more competitive on many routes, and there are many more possible variables that might affect modal choice or the overall frequency of trips.
Steam trains and airplanes need waiting rooms/lounges because {there is only one per day; there are a handful, spaced several hours apart; there are many but the passenger’s ticket (bought in advance) is only valid for a specific one}, and because connections are either uncoordinated, or else they are coordinated to be reliably made — which implies the inclusion of sufficient padding to absorb the typical delay, which is on the order of an hour.
Ok so what happens when people follow the advice of Mr Seat 61 or frankly the national rail change in London timings and turn up at London Hbf 30 minutes early for their train?
“Waiting rooms” are simple Anglosphere insanity, based on some combination of nostalgia for choo-choos of the 1930s, trains-as-flight-level-zero-airline brain death, and a super masochistic desire to be utterly reamed by engineering and construction consortia. Anglosphere obsession with airline-style landside/airside controlled gated “secured” access to train platforms (see also: “a super masochistic desire to be utterly reamed by …”) then doubles the insanity. Gotta have separate waiting rooms and holding pens and VIP lounges and queueing space and cafés and book stores and duty free booze on both the “secure” airside and for those waiting and queueing to be pre-cleared to approach the trains!
In the real world, stations that any non-negligable number of people would ever want to use are located in “cities” and “towns”, and the places to while away the five or tens minutes on might allow before ducking into the station, directly to the platform from the street, are called “streets” and “commercial districts” and “retail” and “cafés” and “restaurants” and “bus stops” and “taxi stands” and “museums”, and and and and …
That stuff comes for free! The Magic of the Free Market provides! You don’t need to make your Intermodal Statement Gateway Nexus four times as big as a functional train station, constructed at nose-bleed rent-seeker public-private contractor rates, in order to make space for cafés in a city. You can have other people do this for you. For free!
The only thing a train station needs to provide are fast, direct, clear, and clearly-signed routes to and from the platforms, where the trains arrive and depart on-time and predictably.
Grand Central Waiting Rooms and World’s Most XTREME Oyster Bar and Majestic Sense of Arrival and Statement Gateway Portals That Reflect Civic Pride are all bullshit that do nothing but get in the way of getting on and off the trains. Sanity before Concrete.
Just stop it!
“I simply can’t wait to spend some time visiting the ticketed concourse area!” said nobody, ever.
They are imagining something from a movie not the actual Grand Central Terminal. The usual place to wait for your train to depart, in the older parts of Grand Central, is on the train. Without using any stairs or elevators.
You cant wait on the train unless its a terminal station
Then they aren’t imagining Grand Central Terminal
Matthew Hutton: “Ok so what happens when people follow the advice of Mr Seat 61 or frankly the national rail change in London timings and turn up at London Hbf 30 minutes early for their train?”
Why on Earth would anyone follow, never mind give, such advice? The nearest I can imagine is that an attempt was made to remind people to pad their schedules for the bus/tube/etc. travel to the train, but this ended up gold-plate-bloated into ridiculousness.
Anyway, on any notably high-traffic corridor, with any luck trains run every 30 minutes, so these people can take the “previous” train. If not, any the intercity trains only run every hour (or even two hours), they are perfectly able to do exactly the same thing that would have happened if they only showed up 3-5 minutes before the train arrived, just longer. Yes, it’s somewhat less comfortable, but that’s exactly why I asked why anyone would follow (or give) this advice.
Richard Mlynarik: “You don’t need to make your Intermodal Statement Gateway Nexus four times as big as a functional train station, constructed at nose-bleed rent-seeker public-private contractor rates, in order to make space for cafés in a city. You can have other people do this for you. For free!”
What sort of naive optimism is this? Of course other people cannot do this for you. The reason they cannot do this for you is because you forbade them from doing so by zoning.
@Basil, outside of the United States and maybe Canada, zoning isn’t going to be much of an issue. Even in the United States, most places with train stations tend to have them near the commercially zoned areas of the municipality they are in. That being said, there are going to be lots of people who don’t want to leave the station while waiting for their train but will want to eat or drink in the station or get something for the trip.
@basil, you follow that advice for the same reason a flight takes a minimum of 5 hours – because you want an easy stress free journey.
What’s stressful about showing up to an international flight where I’m checking a bag 1.5 hours before departure? I get more time to relax at home this way.
The 5 hours is 1 hour for getting to the airport, 1 hour minimum for security etc, 1 hour slack, 1 hour for the flight, 1/2 hour to get bags and 1/2 hour to get to your destination
The only way making every flight take a minimum of 5 hours is an easy stress free journey, is if you have lounge access and not too many other people have lounge access.
The real easy stress free airplane journeys are in Japan, where you casually show up to the airport 20ish minutes before a domestic flight.
@Richard Mlynarik: While waiting rooms that dominate stations are indeed insane I think it’s wrong to particularly associate them with the Anglosphere. By far the worst cases are in China, with long security queues to get into the station, queues for ticket checks at the top of the escalator when your train finally does board, a ticketing system that mostly only permits booking specific trains, and a general assumption that people arrive at the station 30+ minutes before departure. Arrangements for intercity trains in Spain are similar, though at least they don’t impose them on shorter-distance trains. Amtrak *tries* to be like this at Penn Station but there’s still no security and frequent passengers can and do arrive a few minutes before departure and wait on the platform. And in the UK waiting on the platform is completely standard and “waiting rooms” are almost unheard-of (though of course a station concourse is often a good place for a bunch of shops to serve people passing through).
Many of the Cross-country services are a menace, they propagate timetable risk everywhere they go. Not to mention they push expensive multi-power trains onto the networks with their expenses and lower reliability.
They should exiled from the mainlines until electrification at the very least. Exeter-Bristol-Birmingham, to Manchester via Shrewsbury and Birmingham to Derby-Nottingham via Burton on Trent are fine.
Provincial England definitely suffers from “lets keep doing things the way they are because we don’t know better and wasn’t it better in the 19th century” ideas about how trains go. Electrification is treated as something that makes trains go fast and nothing else (see Great Western’s lack of infill stations).
Note that in Israel the lines come pre-infilled – the stop spacing is very short in the built-up areas of Tel Aviv and Haifa. But the dwells are long (or used to be until recently) and the timetabling really isn’t good.
You’ve said before Israel has some cultural cringe, who exactly is it towards? Just the US?
The entire developed world; note that Ido even brings up Warsaw and Athens. When it comes to public transport, Israel is trying to be generic Europe, or generic Western Europe, which makes it really convenient to study best practices because the engineering standards are all generic European – ETCS Level 2, 25 kV 50 Hz, 60 cm sleeper spacing, etc. On private business matters, Israel usually tries to be the United States with some European modifications (e.g. worker protections), but Israelis are very well-traveled and know that the US is not good on public transport.
Would Switzerland be the model that you think they should follow? It seems the right scale.
No, because Switzerland is two-dimensional and Israel is one-dimensional. The concept of a timed connection doesn’t really fit Tel Aviv (or New York, for an American example that Adirondacker keeps pointing to).
There is a difference between the Toonerville Trolley meeting every train and there being so many possibilities it’s impossible to time anything.
To be fair Chiltern doesn’t suffer massive delays at Banbury and Leamington spa from the cross country trains.
The main place Cross-country causes trouble in the East Midlands and Yorkshire where there are very few passing loops and relatively busy mainlines. Oxford-Banbury branch has very few services (but when they go on the Great Western). But if Chiltern’s got the upgrades in services it deserves (i.e. electrification) that would change quite radically.
Until we can get central European levels of infrastructure, UK railways should go on an interlining diet.
I would not call Bristol to Manchester 90km/h decent, I would call it slow, flat terain should allow more. Looking on openrailwaymap tracks seems good, so maybie doing electrification would improve it.
To be fair 90km/h is competitive with driving, and it’s a lot faster than a French regional train. The TGV from chambery to Turin for example only averages 75km/h with many fewer stops.
Is it still competitive with driving if you need half an hour to make your way to the departure station and another half hour from the arrival station to your destination?
Pre-Covid it was increasing but low.
So this is a very fair point.
How desperate you must be to compare Bristol-Manchester to Chambery-Turin. Do I get to see the beautiful Alps on a CrossCountry train?
in theory, being able to walk directly from platform to station entrance without changing levels.
Because yokels from the hinterlands are confused by three dimensional space?
Union Station in Washington D.C. is a glory of City Beautiful planning. It’s a realllllllllllly long trek from the back of the train. It’s a long trek from the front of the train too. You finally get to the front of the station and it’s a realllly long trek to the street. I’ll take the three and half-ish levels of Penn Station Newark or the four of Penn Station New York.
Yeah, Washington Union Station blows, but the European terminals are nicer and many have no-elevation-change walks to the exit. But you’re right that it’s not that important, and Berlin Hbf, which has two level changes between each mainline platform and the exit, is very nice too.
Union Station is spectacular. I’m not there to be impressed. I get the full experience on the long trek to the taxi stand. Or from it.
Seems to be that a nice station is nice, but an efficient station is better. It can be both, but the latter takes precedence.
I honestly think that when Israel was expanding its rail network, they should have imitated Japan rather than Europe. Something like 45% of the population lives in or around Tel Aviv according to Wikipedia and the Tel Aviv Metropolitan area seems to be only 585 square miles with a population density of around 6,000 per square mile. Not as dense as the Kanto area but getting there. Having relatively short frequent electric train lines connect the different municipalities of the area would make sense. Same with the rest of the county. Israel should have also just electrified the entire network from the start rather than building a diesel network and then electrifying it. The county is small enough where they could do entirely electric effectively.
Adding to my above comment, does Israel have enough train stations itself? The number of existing passenger stations seems to around 70 while similarly sized countries in Europe have hundreds of stations. Taiwan has two hundred forty stations. South Korea, another country where half the population lives in and around the economic capital, has over 700 stations in the capital region. Would Israel be better serve by having more stations, especially in the Tel Aviv area like similarly sized/dense countries?
Taiwan has 2.5 times Israel’s population…
I was referring to geographical size more than population. Yes, Taiwan is still bigger by about 5,000 square miles but most of the population lives on the west coast of the Island like most of the Israel population lives on the Coastal Plain.
From what I heard, the French approach with many terminal stations was a feature imposed by King Louis Phillipe, who divided the country in separate domains, each with their distinct company and head station in Paris. The idea was to ensure no company could grow too powerful and enshrine the central role of Paris. I doubt user convenience was a priority. This “étoile” (star) architecture was reproduced at the local levels, of course.
Britain ended up looking exactly the same without any state involvement and with a lot more capital available in Birmingham and Manchester than in any provincial French city.
It would seem more like unless there is some state involvement to encourage otherwise, railways naturally ended up building several terminals in major cities.
Overwhelmingly immense commuter demand encourages through running to handle it. Struggling demand encourages through running to attract and retain ridership through convenience. In absence of either, it would seem better for a real estate-railway company to force people to change trains at a terminal. Even the increased dominance of the central city from that decision, benefits real estate-railway companies.
And even the big projects to through run trains happened mostly with state involvement.
Many of the termini were established before 1850, certainly before the ~1870 “conceptual maturity” and the big construction boom. I understand they were a function of:
– potential connecting modes being so terrible (even horse-drawn urban trams not being common yet!) that the railways were willing to make tradeoffs to decrease “last-mile” transportation, which was easier with a radial approach RoW than a tangential one;
– railways were built piecewise, point to point, with great uncertainty about whether there would be enough capital (political and/or financial) to build some kind of grand network, and the obvious way to connect two cities (or to connect a new, independently destination-level city to a spindly assemblage of lines that does not yet merit describing as a “network”) is with termini;
– the engines available at the time wouldn’t have been able to execute through-running anyway;
– even at small towns most definitely with through-running, if the train did stop, then the station dwell would easily be 10+ minutes until well after WW2. The time necessary for the porters to handle luggage and mail, perhaps taking on water or changing engines, made it common practice for passengers to alight and buy food/drink/tobacco/etc. on firm ground before continuing. (I’ve seen safety films from ~1960 where these are “background” establishing events for the mishap. Allegedly it’s still common practice in the ex-USSR.) Given this, and the laughable frequencies (“24-hourly takt”, the existence of station hotels isomorphic to airport facilities, etc.), termini just weren’t operationally constraining.
Maybe the available capital didn’t lend itself readily to compel different companies to connect their stations, in addition to the reasons mentioned by Basil Marte.
The same happened in Germany, with Berlin and Leipzig having separate stations fror many directions, until the state railways took over and consolidated.
Here’s a pretty interesting historical study on Gare du Nord:
https://www.theses.fr/2015PESC1204
Page 92: 1.2.2 Paris-Nord II : la co-évolution de la gare et du quartier:
“On notera deux types d’éléments qui ont structuré cette trame urbaine et conditionné le maintien de la gare en forme terminus : les conséquences des projets urbains qui ont précédé la première gare du Nord d’une part, et les travaux liés à l’haussmannisation de Paris d’autre part. ”
“Dans les premières études de la fin des années 1830, la ville était envisagée comme une « plaque-tournante » que des lignes ferroviaires pouvaient traverser du nord au sud. Sous le Second Empire, il n’était plus question de mettre en œuvre cette transversalité ferroviaire dans Paris246. Les travaux haussmanniens ont relié les gares par de boulevards routiers (boulevard de Strasbourg, boulevard Haussmann) et la liaison ferrée entre les gares a été réalisée en contournant le tissu urbain (petite ceinture). Dans ce nouvel ordre urbain, les grandes gares étaient pensées et conçues comme les nouvelles portes de la ville247.”
(Rescued from spamfilter. My spamfilter swallowed a bunch of real comments in the last few weeks, not sure why.)
A simpler explanation is that London and Paris were already large developed cities when the railways were built, so urban right-of-way was expensive even in the 1830s.
On the other hand, railway was a very lucrative business in 19th century and early 20th, so it was easier to build through-running tracks and consolidate different terminals 100 years ago than now.
People didn’t travel long distance as much then. In the 1930s there were two decent trains a day from London to Norwich for example. Now there’s one every 30 minutes.
So the number of people who wanted to go from Norwich to Oxford or whatever you might have done with a through train would be very small.
On the other hand, even lucrative businesses had to skirt around the rich and powerful. There was an act of parliament that forbade mainline railway from entering Central London bordered by approximately what’s now the Circle line.
I suspect also that old terminals were favored as a way to limit the noise, smell, and grime of the trains back then to as small an area as possible.
The main problem with Israel Railways is that they have a tendency to put their stations inside of a big parking lot in the middle of nowhere.
They even did that in Tel Aviv, it’s just that the Tel Aviv “middle of nowhere” had ended up growing a highway and CBD around it. There are still big parking lots there.
The trains to Jerusalem and Beer Sheva also somehow always end up slower, less frequent, and less convenient than the corresponding buses.
The same problem that many post-World War II rail systems in the United States and I’m guessing other wealthy countries have. For whatever reason, the designers and planners just have to include big parking lots or garages, which at least allow other buildings near the train stations more than parking lots. It seems to be a combination of political pressure because politicians don’t want to piss off voters and just bowing to the fact that lots of people are going to drive to the station before going to a more dense less car friendly area.
It could be worse. You could be English local councils that invest in speculative parking projects and defend park-and-ride sites from being turned into apartments.
https://www.theguardian.com/society/2023/jun/06/woking-council-property-investment-spree-surrey-write-off
Private citizen NIMBYism and their representatives on the city councils do that over here. BART does seem to be turning at least some parking into housing though. Then COVID caused ridership just to drop and is threatening BART’s service level and funding.
The Northeastern US suburbs do that – Stamford NIMBYed TOD over the parking garage at the train station.
Parking is where the narrative Anglo-saxon neoliberalist mythos goes to die.
The Chiltern line, one of the biggest recent rail success stories has quite a few parkway stations, namely Oxford, Haddenham and Thame and Warwick.
And Banbury and High Wycombe have pretty big multi storey car parks.
Yeah Chiltern is a well run railway by British standards. But remember its a
1. London commuter railway.
2. London-Birmingham railway
3. London- to Oxford railway.
4. Oxfordshire to Oxford railway
This is a system with huge latent demand, which means even if you screw up you can be successful. I should say its not Chiltern’s fault, its Oxfordshire being a bunch of NIMBY dipshits and the Department for Transport sucking balls as usual. (N/b former Oxfordshire resident here)
Seriously can anyone say that Oxford parkway being surrounded by carparks in the most expensive urban area in Europe with less than 300,000 people is an example of “success”. We should judge by standards of the other London 2 track mainlines (C2C etc). Or by the railway it should emulate which Kintetsu’s Osaka-Nagoya line system which is a 2 track system mixing commuter services for 2 megacities plus express intercity service.
It’s utterly bizarre to me that anyone who’s had to deal with the pointless nightmarish hassle of getting from the Gare de Lyon to the Gare du Nord would have any sympathy left for psychological-majesty arguments about the importance of statement-architecture terminals.
If you think Paris is bad enough, London is even worse.
London is worse as there are more commuter terminals, on the other hand 3 of the 4 main long distance terminals (St Pancras, Kings Cross and Euston) are all next to each other.
And Paddington is only a few stops down the circle line from them.
One of the many failures of HS2 is not to create an easier walkway between the 3 terminal stations. But Camden is protecting key assets like….a community vegetable garden, the British Library car park and some 3 story shithouses.
Camden is Kensington with more hypocrisy. Ditto Islington.
Don’t overlook the Amy Winehouse statue …
Statues are fine as uses of space because they don’t take up much. You can put one in a station throughway. See Shibuya’s Hachiko* statue.
*of course Hachiko’s part of the Wartime loyalty-death cult of Japan so its actually quite creepy, but London has statues to Henry Frere of Zulu war fame (its one thing to have Imperial conqueror statues, but incompetent ones is even worse)
Our largest military base is named after Rommel and is located in Teutoburg Forest. I can see a world in which the Brexit press can be sold on a European army if it’s part of an effort to modernize its values and stop honoring the literal Axis.
We were talking about stations….
The Brexit press is dying, its influence has always been exaggerated, its downstream of readers attitudes.
Also remember the Rommel cult thing is a joint Anglo-American-German effort in the Postwar era to justify building the Bundeswehr and to big up Anglo-Saxon ground forces.
For British participation in a European army you need some combo of
1. Breakup of the UK 2. America leaves NATO 3. The Entire British military class going on a massive PR campaign to demand it. 4. Something that really scares people. 5. His Majesty’s Armed Services losing a Falkland’s scale war.
I’d prefer 1. obviously but actually 3. would be doable. But British armed services’ top officer class is very 1. Very well behaved on staying out of politics. 2. Thinks writing a letter to the Times is how you influence the national conversation. 3. Extremely poisoned by the British ideological system which tells you that the only reason Britain isn’t Great is willpower, loves magic asterisk technological solutions and is intellectually incapable of telling the body politic “sorry we can’t do that on 2.5% of GDP please prioritise”.
The Anglo-American Rommel cult was being developed at the request of Adenauer and CDU, who backed Clean Wehrmacht for domestic political reasons. It was never about any kind of realistic geopolitics. For the same internal reasons, CDU ran against recognition of the Oder-Neisse line, even as the US begged Adenauer to calm down and do what NATO needed him to do to mollify the Poles and possibly get Gomułka to be more independent of the USSR. Meanwhile, Germany has marched on – Brandt recognized Oder-Neisse and survived, the neo-Nazi faction of Ernst Nolte lost the Historikerstreit, and the average German thinks the country’s conduct in the war is cause for shame rather than pride. But the Bundeswehr sticks to the Clean Wehrmacht myth and is so full of neo-Nazis that entire units have had to be disbanded.
#4 is the Russian invasion of Ukraine… but all the europhobes took it as an excuse to dump on the EU further, over behavior by France and Germany that was exactly the same as what the US has done or was even slightly more supportive of Ukraine (Germany took until January to free the Leopards, the US took until May to free the Falcons). Poland’s throwing fits at Germany for not permitting Leopard 2 deliveries and then only sending 14 Leopard 2s while Germany sends 18 has been disgusting.
To be fair as far as I am aware Britain did more on Ukraine than France and Germany.
Britain signaled doing more on Ukraine than France and Germany, yes. But total aid given by Germany is higher, and Britain didn’t do the one thing the Ukrainians wanted the most from it, which was to freeze the property of Russian oligarchs.
The issue is that Scholz enjoys lying on this to the public. It makes the SPD elites feel better when they can correct mere mortals who say “Scholz hasn’t done X” after Scholz publicly refused to do X by saying “well, actually, here’s my insider information that he did do X.” SPD elites love ‘splaining more than they love winning elections; when they win elections they have to deliver on promises, whereas when they are junior partners to CDU they can always tell Jusos “we didn’t win so we need to be more moderate, now shut up, homos.”
But why would anyone want to transfer between long distance trains in London (except Eurostar)? All the long distance trains go to the north and west. If you want to take a train between any of Leeds, Birmingham, Manchester, Bristol, Oxford, Nottingham, Scotland, Wales, there are more direct routes that bypass London. I guess there are a few possible journeys like Cambridge to Bristol that are best done through London (although they should really restore the Oxford-Cambridge line so this is no longer necessary), but they are a small minority of trips and planning shouldn’t be done around them.
This is in contrast to Paris, where Gare de Lyon/Gare du Nord is a frequently needed transfer because Paris has a more central location in the country and the continent.
I was thinking smashing the existing buildings, building a commercialised walkway-plaza from Euston to St-Pancras with at most a moving walkway and pay for it by letting developers build 10 large luxury hotels etc.
The primary beneficiaries of this idea wouldn’t be long-distance travellers, its for all the smaller trips within the London commuter zone arc of Watford-Northampton to Stansted-Cambridge plus the Tube. Plus having low-middle density building next to the largest concentration of rail connections in the country is a real waste of resources.
N/b my model here is Umeda in Osaka which has a bunch of competing rail operators, yet its become the central hub for the Keihanshin area above previously bigger rival districts like Namba or Nihonbashi. And the other Japanese megastations which Alon thinks have “missed connections” but are connected in an underground pathway complex.
If you imagine a giant clock with London at its center, Cambridge is at about 12:30 and Northampton at maybe 10:30. The angle between them is about 60 degrees. For almost any trip in that wedge, it will be quicker to travel directly, rather than through London at the center of the clock. Since there isn’t rail infrastructure for such trips nor (generally) demand which justifies building new lines, it will have to be by bus.
*by car.
Even at ~US$60/km PPP the Oxford-Cambridge line is going ahead – and thats 3x what the Belgians or Swiss spend on full high speed rail adjusted for PPP.
So connections between the main lines are possible – especially if we could build them at less than a Belgian or Swiss high speed line costs.
Checking Google Transit, travel by bus between Cambridge and Northampton in fact takes over 3 hours, while taking trains and changing in London takes under 2.5, so you’re simply incorrect about it being quicker to travel directly by bus. The same goes for many other city pairs; the radial routes out of London have enough demand to justify investment in speed & frequency far greater than circumferential routes, so it’s very common for changing in London to be the fastest and most frequent option. A good link between Euston and St Pancras would also be helpful for WCML travellers continuing to Brighton and other southern points (Gatwick etc) via Thameslink.
Zürich combines a surface terminal (pretty much dictated by history and geography) with a steadily growing underground station with a mix of terminal and through tracks. While a few lines are all-surface or all-underground, others hop levels (e.g. the full hour train leaves underground, the half hour train on the surface).
I am not familiar with Israel, so will stick to talking about Europe for the time being. In countries with large cities and big distances, terminus stations work best. Beyond a certain size, terminus stations are actually more space efficient and easier to construct than through stations. Terminus stations might be fatter (require more platforms for trains to be turned around) but can be shorter as they only need one throat, and they lend themselves to being fairly easily constructed at edge-of-city-centre locations. Through stations have double throats so the overall footprint is considerably longer if slightly thinner, and new ones tend to need to be underground with giant caverns constructed because there is usually no above-ground site available. These underground stations cost a fortune and still end up being dark and dingy environments at platform level.
Large cities create enough point-to-point demand without a single train needing to serve multiple markets. A Munich – Cologne service need not serve Frankfurt; Frankfurt can send its own trains to both Munich and Cologne. Going through and calling at a major city en route on an inter-city train is disruptive. Through city alignments are usually slower than bypass alignments and you need a hefty dwell to get everyone off and on this train. A Hamburg – Munich via Berlin service needs that 10-minute dwell at Berlin Hbf and even then the boarding experience is still stressful, as regular users of that service will be able to testify. For cities that generate enough demand to fill whole trains, passengers should be able to board and alight at terminus stations in a leisurely manner, and the train should be sitting in the platform before the majority of passengers arrive. Intermediate passenger churn really degrades the inter-city experience and is a major source of performance risk. Intermediate calls should only really be for markets that produce 10 boarders and alighters each door that can be accommodated within a 2-minute dwell stress-free.
Berlin Hbf doesn’t work all that well. There are loads of inter-city trains carrying fresh air going to and from Ostbahnhof and Gesundbrunnen using precious city centre track capacity, and they actually travel further to be serviced and turned around at Rummelsburg depot. The Hbf – Rummersburg run easily costs an extra unit for an hourly service. The 8-platform Hbf low level has to handle both intercities and regionals, and is actually quite restrictive in terms of what you can do with timetabling. What Berlin would have been better off doing was keeping the old Potsdammerplatz, Anhalter, Lerhter and Hamburger stations; building a simpler two-track north-south tunnel for regional services and actually getting the S21 done at the same time.
Stuttgart 21 is more of a land-release real estate project with questionable transport benefits. Faster journey times and additional capacity could have been more cheaply delivered simply by building a by-pass high-speed alignments for through trains (a half-hourly Munich – Frankfurt service doesn’t need Stuttgart punters to wash its face), with a (reduced footprint) Stuttgart terminus station handling terminating trains only (trains terminating and going back the same direction are a lot less disruptive than trains reversing and going out on a different line).
It provides more frequency. That’s why TGV operations are so jank: the frequency is terrible, because each major city pair has its own dedicated trains. It works for the thick markets, like Paris-Marseille, but the thinner markets have too low frequency, so e.g. between Lyon and Toulon or Marseille, the car dominates and TGV ridership disappoints. Over the last 15 years, German intercity rail ridership has gone from lower than TGV ridership in absolute numbers to about on a par per capita, because the TGV tapped out all the thick markets already and can’t grow the rest due to its anti-Swiss timetabling practices.
The Shinkansen, KTX, and THSR do it right as well: trains stop at major intermediate cities – at no point has any Shinkansen ever skipped Nagoya or Kyoto.
In the case of Frankfurt, the issue is that the distances are short enough that a complex setup with trains that run only every two hours means that passengers would wait longer than the train trip. Even hourly frequency is low enough on a city pair like Cologne-Mannheim or Essen-Frankfurt that it would visibly discourage ridership. Those 1:00 and 1:30 connections are a half-hourly-or-GTFO situation, and being too creative with operating patterns is counterproductive.
In the case of Berlin, there’s a specific problem with the Hbf throat in that it’s not possibly to separate regionals and intercities perfectly cleanly. But at the end of the day, it means that people can actually connect between trains there, which they can’t in Paris or London and would not be able to do if it went back to steam operations with separate stations for separate directions. There is no need for a 10-minute dwell; Shin-Osaka has one-minute dwells and the ICEs are not that much worse at this than the Shinkansen, it’s just an expectation that can be changed over time (as it was in Japan!).
Apparently the Kodama do a lot of loitering. Shinkansen are a different loading gauge and track gauge from the other networks so they are restricted to their own tracks. It shouldn’t be a problem in most of North America because everything not-California and not-Texas will be “Amtrak”. And anyplace other than the Northeast Corridor there aren’t going to be a lot of different services. …. someday, far far in the future when there are six tracks between New York and Philadelphia the Springfield-Wilmington and the Boston-Harrisburg Kodama can sneak through NY-Philadelphia on the commuter express tracks. The Hikari using the West Trenton line will have to.
The first Nozomi introduced, in 1992, did skip Nagoya and Kyoto. The train would depart from Tokyo at 0600, Shin-Yokohama at 0616, arriving in Shin-Osaka at 0830.
Weirdly, the Nozomi going from Shin-Osaka at 0600 would arrive at Tokyo at 0830 without skipping Kyoto or Nagoya (but Shin-Yokohama was skipped).
It comes it goes, it mostly goes, Amtrak schedules a NY-DC train that only stops in Philadelphia. now and then. They skim off the peak hour surge.
Large city pairs don’t need through running to support frequency. TGV frequencies are shit because of SNCF conservatism – Frances LGVs and terminals have ample capacity. German frequencies are shit because of lack of or poor use of infrastructure. SNCF being ‘anti-Swiss’, if we call it that, isn’t because it doesn’t run through services, because it doesn’t understand marginal costing or the value of turn-up-and-go. The French would only run a train if it’s full – it doesn’t do all-day clockface timetabling, and as a result it has to resort to compulsory reservation to manage demand which puts off a lot of travellers. The Swiss run all-day clockface timetables and generally plan for (roughly) 75% peak and 25% off-peak occupancies on its inter-city services, and it justifies such service levels through higher prices of flexible tickets allowing you to walk up to any train without reservation. This is similar to the UK practice too.
Italian and UK examples show Paris – Lyon should be an all-day 2tph service with probably peak extras. Cologne – Frankfurt; Frankfurt – Munich; Cologne – Munich all justify point-to-point direct services in their own right if infrastructure capacity is there and journey time is attractive enough to hoover up all domestic air demand. London – Manchester is already a 3tph market. In theory you could run a London – Birmingham – Manchester service at 6tph, but London – Manchester journeys would be 15 minutes slower (alignment penalty and dwell), and Manchester passengers would have to suffer Birmingham churn. The journey time penalty and loss of comfort far outweigh the frequency gain. The Italians clearly apply this line of thinking to Milan – Florence – Rome too. I would also apply this model to Rhine/Ruhr – Frankfurt – Munich and Berlin – Leipzig – Munich axes too. Intra-Rhine-Ruhr journeys should be done on regional services at near-metro frequencies (Rhine-Ruhr Express will offer a 4tph frequency when complete) and Frankfurt – Mannheim should have metro frequencies from overlapping inter-city services.
Simple service patterns with clear service hierarchies produce memorable mental maps that facilitate interchanges. Cross-London and Cross-Paris journeys are easy – you just go into their metro system. The Dutch and the Swiss aren’t afraid to tell people to change trains (the Dutch through frequency and the Swiss through timed connections). Navigating German long-distance network is a mare – the plethora of low frequency services is impossible to memorise; interchanges are long and often unreliable.
In any case, in my mind Mannheim would be an intermediate stop for all Frankfurt bypassing trains as it’s low volume and low journey time penalty, so Cologne – Manheim would have a direct 2tph service out of Cologne – Munich services.
Generally in my book intermediate calls on inter-city services are only suitable for volumes up to 10 boarders and alighters per door, the sort of market volume that can be reliably handled within a 2-minute dwell. Manchester/Munich/Lyon sized markets – terminate; Preston/York/Nuremburg/Gottingen/Bologna sized markets – through.
1. Cross-Paris service is in fact terrible, and people who switch from Nord to Lyon constantly complain, like Diego who occasionally comments here. There are bypass trains, but they make all stops and TGV dwells are a horror story, partly due to train design, partly due to SNCF operations.
2. Re 10 boarders and alighters per door: I’ve observed much faster interchanges at Gare du Nord on the RER B, I want to say 25 per door in a minute but I don’t remember. We can also try to figure out Nagoya: total JR volumes between pairs of metro areas are Kanto-Keihanshin ~50 million, Kanto-Nagoya 28 million, Nagoya-Keihanshin 13 million, so close to half the volume of a Nozomi boards or alights at Nagoya, which at 70% occupancy is 16 per door. On an ICE with 72 seats of which 40 are occupied, if it’s built to have two door pairs per car (like the Shinkansen or like the intermediate cars on the Velaro CN), this means half the train is allowed to turn over by the 10 limit, or 80% is by the 16 limit.
3. Yes, the frequency gain on very thick markets is not as important. HS2 markets are very thick, and I don’t really mind putting Birmingham on a branch (though I do mind the construction of an entirely new station at Curzon Street). Britain differs from France this way in that its cities are bigger and there just isn’t a long tail of Mâcons, Saint-Etiennes, Grenobles, and Bourgs to accommodate. I just mind the multi-billion pound project to add tracks to Euston where it’s a solved problem to turn faster in countries that don’t hire British consultants.
4. The Italians are building a through-station in Florence. The main reason they’re not doing it in Milan or Rome is that it would be too expensive for less gain – but Milan and Rome have a single intercity terminal each, not directional stations like Paris or London.
1. Improving the reliability of RER would be the answer to improving cross-Paris connectivity. The other problem is TGVs and Intercites are not half-hourly anywhere and there’s compulsory reservation. It means there isn’t always a train sitting in the platform ready to depart and you can’t just board any train.
2. Intercity stock with narrow end doors and separate vestibules are not comparable with metro stock. You plan a timetable for reasonable worse case, which means something like 120% occupancy (commonly seen all day on German networks in summer season).
3. Curzon Street is about the only place where land is available to put new platforms. It’s not perfect but there’s something called money. I’ve commented at length about respecting human behaviour and why 25-minute turnarounds are appropriate in a European context. You have a very sharp mathematical mind but sometimes your mathematical thinking only addresses part of the system. You seem to gloss over very important human factors and rail operations intricacies.
> Intercity stock with narrow end doors and separate vestibules are not comparable with metro stock. You plan a timetable for reasonable worse case, which means something like 120% occupancy (commonly seen all day on German networks in summer season).
You can also design intercity trains for more turnover faster, by providing level boarding and an extra door per car. The most aggressive Shinkansen timings are during their most overcapacity times, such as Golden Week or recovering from a major service disruption.
> I’ve commented at length about respecting human behaviour and why 25-minute turnarounds are appropriate in a European context. You have a very sharp mathematical mind but sometimes your mathematical thinking only addresses part of the system. You seem to gloss over very important human factors and rail operations intricacies.
What European human behavior necessitates long turnaround times? If it’s some deeply rooted in the European human behavioral context, why does it seem like Swiss turnaround times at busy terminals aren’t bad? And why does it seem like UK turnaround times are even worse than typical continental Europe?
Milan has Centrale and Porta Garibaldi, Turin has Porta Susa and Porta Nouva. So plenty of Italian cities have two stations.
It’s actually worse than Birmingham as you can’t walk between them in less than 5 minutes as you can from Moor Street to New Street.
“You can also design intercity trains for more turnover faster, by providing level boarding and an extra door per car. The most aggressive Shinkansen timings are during their most overcapacity times, such as Golden Week or recovering from a major service disruption.”
If a network is configured for point-to-point services then you would prioritise capacity over circulation, knowing circulation is only needed at beginning and end of runs, or intermediate calls are intended at low circulation locations only.
HS2 will have level boarding at new stations (would have required a derogation from EU interoperability rules if Britain didn’t Brexit). Mainland Europe will have a problem – a mixture of 550mm and 760mm platforms where member states can’t make their minds up on platform height. To me 300+km/h rolling stock should have ~1200mm flat floor height throughout and have usable accommodation throughout the length of the train – anything that involves power cars or rollercoaster floors represents a regrettable compromise. Much of Europe will be stuck with non level boarding for a very long time, and this is an important aspect of ‘European context’ that should be considered when specifying station dwell times.
London-Leeds-York would be 2h40 with a 5 minute dwell at Leeds vs 1h52 on the direct London-York service.
Same with Preston-London, which is 2h09 direct and 3h12 via Birmingham.
So the penalty is much more than 15 minutes.
Birmingham isn’t even on the WCML route between London and points north, which runs on the Trent Valley line.
“Birmingham isn’t even on the WCML route between London and points north, which runs on the Trent Valley line.”
Eh? Of course Birmingham is on the WCML. WCML refers to the whole system of through routes and branches / loops. Birmingham, Manchester, Liverpool are all WCML destinations (and considered to be on the WCML) as well as Lichfield, Preston and Carlisle which are on the ‘main’ line.
The question is would an HS2 configured to go through Birmingham and Manchester be a superior product than the currently planned HS2 with Birmingham and Manchester on spurs. As far as I’m concerned:
– A trunk + spur model delivers the fastest end-to-end journey time for each origin-destination pair. An alignment through Birmingham and Manchester would involve slower route sections.
– As a London – Scotland passenger, a 2tph direct service with limited churn is a more attractive offer to me than a more frequent but significantly slower journey with more churn.
– In the specific context of the HS2 corridor, a spine + spur + surface termini configuration is cheaper to build than one with underground through stations at Birmingham and Manchester (British ex-industrial cities still have that ex-industrial doughnut next to city centres which are easy to build platforms + sheds on)
– Operationally, operating a series of point-to-point services gives me more control over how many services I want to run over each segment.
Intermediate passenger churn has a disproportionate degradation impact on passenger experience.
When you have half the train getting off at an intermediate call people start standing up 10 minutes before arrival, the aisle gets filled quickly, people can’t get to their luggage and start panicking; if you are in the aisle seat and you window neighbour gets off, they stand up crouched under the overhead luggage rack, try to get their bag and coat down but can’t and there’s not enough space to actually put their bag and coat anyway and you are stuck in that limbo until the train actually stops.
Then as 25 people get on through each door the first person takes an age to settle into row 3 because they have to turf someone out of their booked seat, then the other 24 people are stuck in the vestibule and platform. The the second person spends 2 minutes trying to put their oversized suitcase into the luggage rack. The dwell gets breached (as is routinely the case on Germany ICE trains) and it’s not until 10 minutes after departure are all passengers settled into their seats.
If you are a through passenger you have no peace for anything up to 30 minutes. This is daily occurrance on German intercity trains.
Through arrangements impose significant journey time penalties especially over long distances. Tokaido Shinkansen has loads of 120-150km/h stretches through city centres, and a Tokyo – Fukuoka train has to go through about 10 of them (the trains do stop at these places). That’s a journey time penalty of at least an hour. If Tokyo – Fukuoka ran direct services at uninterrupted 300km/h it would be achieve sub 4 hours and make a significant dent in the air market, and a 2tph service would wash its face. Maybe one day the Chuo Shinkansen could be extended to serve that role.
“Through arrangements impose significant journey time penalties especially over long distances. Tokaido Shinkansen has loads of 120-150km/h stretches through city centres”
‘Loads’ and ‘120-150 km/h’ are exaggerations. The Tokaido Shinkansen ‘city centre slow zones’ can be counted on one hand:
1) Tokyo to near Musashi-Kosugi: an unavoidably slow urban approach
2) Atami: a sharp curve between tunnels with a 200 km/h limit
3) Shizuoka: around 240 km/h
4) Hamamatsu: around 250 km/h
5) Toyohashi: around 250 km/h
Many trains do transit these sections slower, but that’s more to do with dense scheduling for timed overtakes rather than a hard limit of the infrastructure.
You could count Nagoya and Kyoto as slow zones, but they are mandatory station stops justified by demand (given the traffic numbers Alon cited above), with dwells that take less than two minutes, and no turnover unpleasantness on the countless times I’ve ridden through.
If you were serious about the Shinkansen competing with the planes you’d charge the Tokyo-Hiroshima fare for an hourly or half hourly train that did Hakata-Kokura-Shin Yokohama-Shingawa-Tokyo
Your proposed direct service would probably shave 20 minutes off the current Nozomi schedule by speeding through the smaller Sanyo stops and eliminating dwells (though not deceleration) at the major stations.
That 20-minute saving would be almost entirely offset by longer waits from decreased frequency. The current service is 2 to 3 tph, and more during peak seasons, by virtue of traffic to and between all the intermediate stops. An optimistic 75% modal share of all Greater Tokyo-Fukuoka Prefecture trips (9.26 million out of 12.35 million per annum) would barely fill 1 tph each way.
Ironically, the markets forgone by a direct service (Fukuoka-Hiroshima/Keihanshin/Aichi, Hiroshima-Aichi/Tokyo, etc.) are also the ones more predisposed to rail travel given the shorter distances.
By my calculations currently Shin Yokohama-Kokura takes 4h19 on a Nozomi train. With non stop service at 285km/h from Shin Yokohama through to Nishi Akashi and 300km/h from Nishi Akashi to Kokura with 4% padding that would take 3h34 or 3h41 with 7% padding.
Overall that is 38-45 minutes quicker than current service.
As per the formulae the 7 stops on route cost you 28 minutes. If we assume the 5 big stops have a dwell time of 2 minutes not 1 and that there are 5-10 minutes of other slowdowns which might or might not apply to a non stop train then we are in the right ballpark.
Your calculations assume full speed through Nagoya, Kyoto, Shin-Osaka, Okayama, and Hiroshima, which is impossible on the current station/approach tracks. As mentioned above, a non-stop train would still have to decelerate, pass those stations at low speed, and then accelerate again (ie. how the original 1992 Nozomi skipped Nagoya and Kyoto); thus, the time saved at each of these major stations would be the dwell time (60 to 120 seconds), plus the time loss of a stop/start versus a low-speed transit. That’s around 2-2.5 minutes per city, and so 10-12 minutes across all five.
Skipping Shin-Kobe and one other Sanyo stop (from the Fukuyama/Shin-Yamaguchi/Tokuyama rotation) would save more time as those stations can be transited near full line speed; not so much in terms of dwell time, which is sub-60 seconds at each. Supposing 4 minutes per stop, that’s another 8 minutes saved.
Overall that is 20 minutes quicker than current service, the figure I used. Even if it were a more substantial 25-30 minute travel time reduction, the counterpoint on frequency would still hold.
10 million annual trips is 3-4tph territory. Tokaido Shinkansen doesn’t have enough capacity to meet all viable objectives. 6tph Nozomis between Tokyo and Fukuoka isn’t very much. Rome – Milan has up to 4 ‘Nozomis’ per hour for much smaller cities (treat the Florence reversers as ‘Hikaris’). Incidentally Italian high-speed rail killed off Air Italia.
If the objective were to kill off domestic aviation then the line wouldn’t follow the existing alignment. You’d design the infrastructure and operation for sustaining uninterrupted 300+km/h running and without having to slow down for overtake moves. The largest cities would be served on loops or spurs with 230km/h turnouts (turnout speed at which a 3-minute headway can still be sustained between a diverging train and a subsequent straight-on train without the straight-on train having to slow down).
A hypothetical as-straight-as-possible HSL with spurs or loops for Nagoya, Kyoto-Osaka-Kobe (one loop), Okayama, Hiroshima would deliver a c. 3h30m journey time for Tokyo and Fukuoka and support 3-6tph point-to-point services for each origin-destination pair (Kyoto-Osaka-Kobe as one cluster) with 18tph coming out of Tokyo. Then the existing Tokaido Shinkansen can be exclusively used for Hikari and Kodoma stopping patterns.
In the even Chuo Shinkansen goes all the way to Fukuoka that would be the aviation killing service. Yes, in all likelihood it would just have through stations like the initial phase, but I suppose at 500km/h you can afford some stopping penalties in a way you couldn’t at 300km/h.
“10 million annual trips is 3-4tph territory”
9.26 million trips is 25,370 trips per day, or 12,685 per direction. An N700S Series Shinkansen seats 1,319, so an hourly service between 6am and 7pm (14 trains) would provide 18,466 seats. 1 tph would thus operate at a fairly reasonable 69% load factor.
“If the objective were to kill off domestic aviation then the line wouldn’t follow the existing alignment. You’d design the infrastructure and operation for sustaining uninterrupted 300+km/h running”
Of course you would, but Matthew Hutton’s proposal and calculations implied the existing alignment.
I don’t think you can get to 69% occupancy with a 1 tph operation; the TGV’s high load factor relies on a peaky, micro-targeted timetable, whereas the Takt-y ICE is around 50%.
And then, while the Shinkansen seats about 1,300, European trains never do – the ICEs seat about 900 and the TGV Duplexes seat 1,050.
“I don’t think you can get to 69% occupancy with a 1 tph operation”
Almost certainly, but if the question is ‘how many trains would 10 million annual riders fit into’, the maths works out.
“while the Shinkansen seats about 1,300, European trains never do”
Yes, but the Shinkansen capacity figure is most salient when the discussion is about structuring service in Japan (with the implicit constraint of currently available resources).
The question was never that. The question (implied) was ‘what is an attractive and viable proposition for serving a ~10m strong point-to-point market by HSR and kill aviation’. The answer is a 3-4 tph offer on an alignment optimised for Tokyo – Fukuoka while also taking in the major cities on a ‘spine+spur/loop’ model.
For journey times of more than an hour the train should really offer a bit of ‘frill’ – lower density seating, better catering offer, and lower average occupancy that are more towards the (Central) European norm.
Hey, the seat pitch on the Shinkansen, 1 m, is actually a bit longer than on the TGV, where I think it’s around 90 cm (or, rather, 180 cm for a pair of rows facing each other).
So the Shinkansen is better than the TGV on the schedule – especially outside the capital. It is slightly better on reliability and on seat pitch. In terms station location it is also better by having one in Tokyo and having better through stations everywhere else.
On the other hand the Shinkansen is much worse on price. Plus it doesn’t have a cafe car and the on board wine and possibly beer selection is presumably worse. Additionally it has fewer table seats in standard class.
Those are your propositions, elements of which I can agree with. My responses, however, have been directed towards Matthew Hutton’s proposition, where the words and numbers (eg. the current 285 km/h speed limit, the existing track length) imply something that could be introduced at the next annual timetable change.
Actually, Japan’s loading gauge helps with accommodating 3+2 seating comfortably. Don’t they do 3+3 as well? I think that could be done away with. And a larger shop area, though probably not a full dining car.
3+2 in Standard Class has long been the Shinkansen norm. 3+3 was trialled on the double-decker MAX sets and those have already been done way with.
“Those are your propositions, elements of which I can agree with. My responses, however, have been directed towards Matthew Hutton’s proposition, where the words and numbers (eg. the current 285 km/h speed limit, the existing track length) imply something that could be introduced at the next annual timetable change.”
Given this sub-thread stems from my original comment on journey time penalties imposed by through arrangements, forgive me for pointing out I defined the context. Matthew Hutton’s example represented a conservative illustration of journey time improvement potential, which supported my (implied) assertion that a trunk+spur set-up would offer significant journey time savings while still offering a high frequency. I’m pretty sure Matthew Hutton was commenting in the spirit of this said context.
The Japanese didn’t ask any of you and they are proposing to build mag-lev Chuo Shinkansen. If I want to go from Tokyo to Osaka I don’t care if the train goes through Nagoya or not.
In the theoretical world where an Australian HST system gets built, what would you do at Sydney? Terminus at the main station with northbound trains running back west before turning north somewhere around Parramatta, or running straight through north in a new harbour crossing?
Two points:
The UK has CrossCountry, an intercity operator centred around Birmingham New Street Station. It’s routes link Scotland and Manchester to the south coast, Wales and Cornwall. This network connects the country without going anywhere near London, the rail equipment of the M5 motorway. The train sets used are 100mph (160 kph) and 125mph (200 kph).
The Israeli Eastern railway will be a useful freight bypass, allowing freight traffic to the ports of Haifa and Ashdon to bypass Tel Aviv’s Ayalon rail corridor. It might not have much of a passenger potential as it skirts Greater Tel Aviv’s outer suburbs and the wrong side of the airport, but removing freight traffic from the busy core of the system will have a positive influence on the system as a whole.
I believe there might be a speed / headway issue with the Ayalon corridor. It may be that running speeds are set too high, as if the tracks are fast intercity tracks, rather than slow stopping tracks with multiple stations. The figure of 24 trains per hour per track is rather standard on many a system throughout Europe- S-Bahn, RER, the Elisabeth line etc. It may be a case of slowing the Core route from 160 kph to 80 kph or from 80kph to 50 or 40 in order to shorten the headway and increase the number of slots.
The wors of Israel Railways’ plans wasn’t mentioned in the article: The wasteful plan for unnecessary two high speed tunnels under the Ayalon corridor – designated as tracks 5 and 6. Israel Railways plan for 2040 has a maximum of 25 tph in the Ayalon. The forth surface track should allow for 48 tph. This, when the area will enjoy metro line 1. Israel Railways maximum planned speed of 250 kph puts it in line with AVANT which uses regular stations and lines between high speed sections.
The current plans for extending the Jerusalem line into the city with 4 tracks for 14 the are similarly perplexing, as two tracks could extend the line to Malha where it could use the existing station as a surface terminus at a fraction of the cost of the planned depo at Oranim Junction.
As to the London terminals: Railway construction within the core of the city was banned by law in the 19th century. This meant there was no possibility of a through line. The South Bank viaduct and the Nort London Line allowed orbital connections round the centre connecting the main lines, mainly for freight.