How Fast is Necessary?
Swiss intercity rail planning follows the maxim “run as fast as necessary, not as fast as possible.” This is usually uttered in opposition to high-speed rail in the sense of the French TGV network. But what does this slogan really mean? And how does it inform good planning?
The issue of timed connections
The origin of the Swiss planning system is in the 1970s and 80s, as it was refining intercity rail, taking what West Germany was doing with its InterCity brand and going further. Two key elements were present from the start: timed connections, and regular clockface timetables (initially every two hours in West Germany). The clockface timetabling facilitated the timed connections, since it’s easier to figure out how to schedule a timed transfer at (say) Frankfurt if the same train moves happen at the same intervals.
With these elements in place, Swiss Federal Railways (SBB) set up a schedule in which trains would be timed to arrive in Zurich all at the same time, currently just before the hour, and depart all at the same time, currently on or just after the hour.
The issue is what to do at stations other than Zurich. Optimizing for timed connections at Zurich means compromising on the question of what to do at other cities. If trains arrive and depart Zurich roughly on the hour, then the terminal at the other end may have ugly arrival and departure times – for example, a 1:40 trip time, with Zurich arrival :56 and departure :04, would mean that the other end has an arrival time of :44 and a departure time of :16. If it’s an hourly train, it means it’s not possible to time connections there – an arriving bus or regional train would have a transfer time of perhaps 20 minutes, which in a country the size of Switzerland is a large share of the overall trip.
And then there is the looming issue of intercity connections. Zurich is located fairly centrally in Switzerland, but there are some key connections that don’t go anywhere near it, led by Bern-Basel via Olten. Bern’s central location makes it a great node for timed connections as well. However, the Zurich-Bern trips took 1:09, making it impossible to have timed everywhere-to-everywhere connections in both cities.
Speeding up trains to make connections
In the 1980s, the Zurich-Bern trips took 1:09, so connections could only be timed in Zurich, not Bern. At the time, there were plans for a French-style high-speed rail network connecting Zurich, Bern, and Geneva, but those plans were canceled due to high costs relative to Switzerland’s size.
Instead of running as fast as possible, enough to connect Zurich and Geneva in perhaps 1:30, trains would run as fast as necessary, just enough to make connections. The centerpiece of this plan, dubbed Rail 2000, was to speed up Olten-Bern by just enough to shorten Zurich-Bern and Basel-Bern to 0:56 and 0:55 respectively. This way, trains could arrive in all three cities just before the hour and depart just after, facilitating more timed connections.
This system was a resounding success. Swiss rail ridership has been sharply rising in the last 20 years, from an already fairly high level; by all metrics I am aware of, such as modal split, ridership per capita, and passenger-km per capita, it is Europe’s strongest rail network. More speedups are planned, all aiming to add more points where trains can be timed to connect, called knots, till the network looks like this:
The shape of Switzerland
The notion of running trains as fast as necessary is in one sense a global principle. But its surface manifestation of a system designed as an alternative to high-speed rail is a product of Swiss geography; in Japan, the current speed of the Shinkansen is also as fast as necessary. Switzerland’s current knot diagram has to be understood in the context of the following geographical features:
- Switzerland is small enough that the strongest trunk corridors, like Zurich-Bern, can support just a train every half hour.
- Switzerland is also physically small enough that the Zurich-Bern-Basel triangle has legs of about 110 km, which an upgraded rail system can connect in an hour minus transfer time, and which it is very hard to speed up to 30 minutes minus transfer time.
- Switzerland is two-dimensional: there is no central trunk through which all service must pass – the diagram above depicts Zurich-Olten as a key link, but Luzern, though not yet a knot, connects separately to both Zurich and Bern.
- Swiss train stations have a surplus of tracks, while still functioning as central stations, without the separation into different stations for different directions that Paris and London have.
Switzerland is not unique in having these features. The Netherlands is the same: it’s a small two-dimensional country with demand for many connections to be timed. The Netherlands built a 300 km/h high-speed rail line between Amsterdam, Rotterdam, and the border with Belgium toward Antwerp, Brussels, and Paris, but this line is not important to its intercity rail planning, which instead lives on Swiss-style knots.
However, not even all small countries are like Switzerland and the Netherlands. Israel is one- rather than two-dimensional: all trains pass through Tel Aviv and the Ayalon Railway, and there is no real need for timed connections, as the most important city pair not involving Tel Aviv, Haifa-Jerusalem, passes through Tel Aviv, with through-service to begin as soon as electrification is completed. Taiwan is so linear that a single high-speed rail line connects all of its cities, with higher ridership per capita than in any other country I have data for except maybe Japan; even before high-speed rail, the shape of its mainline network was a single line on the west coast connecting the cities and another on the east coast connecting smaller settlements.
And then in large countries, running as fast as necessary requires choosing a speed. Germany is two-dimensional like Switzerland, but has a string of city pairs 90-100 km apart – Hanover-Bielefeld-Dortmund, or Mannheim-Stuttgart, or the Fulda-Frankfurt-Würzburg triangle, or Würzburg-Nuremberg – for which running a fast as necessary entails a choice between connections in 60, 45, or 30 minutes. Longer-distance city pairs, like Nuremberg-Munich, Berlin-Leipzig, and the Hamburg-Hanover-Bremen triangle, are even less constrained.
Capacity
The Swiss network is based on the idea that near major stations like Zurich, trains should arrive and depart on pulses every 30 minutes. This way, the Olten-Bern line runs eight intercity trains per hour, but they are not equally spaced: they are timetabled in two platoons over a period of seven minutes each, with the off time used for regional trains not participating in the knot system.
Even in the Netherlands, this isn’t quite tenable. Amsterdam-Eindhoven trains come at regular 10-minute interval, each third train requiring a connection at Utrecht and the other two trains in three running directly. In Israel, Tel Aviv-Haifa trains run at 30-minute intervals midday, but for two four-hour peak periods each day this is boosted to four trains per hour – and this is even before electrification has been completed.
The situation in Germany deserves especial mention. Germany is investing in it intercity rail network in a way that sometimes tries to be Switzerland except bigger (such as the Deutschlandtakt and other measures supported by technical advocates) and sometimes tries to build high-speed rail lines and through-stations like Stuttgart 21. The technical advocates dislike Stuttgart 21 and argue that it’s un-Swiss to timetable intercity trains throughout the hour rather than in two pulses with inactivity between them. But Germany has large enough cities that it can’t afford to keep intercity lines out for so much time. Already, with mediocre speeds, the Deutschlandtakt plan for 2030 prescribes 4.5 trains an hour between Frankfurt and Mannheim and between Frankfurt and Cologne. As more high-speed lines come online, demand will grow; Deutsche Bahn projects to double ridership between 2019 and the 2030s, which will force the busiest links to operate a train every five to 10 minutes.
Is Swiss planning useful outside Switzerland and the Netherlands?
Unambiguously, yes. However, it would look different.
The best place to see how different it should be is, naturally, the Northeastern United States, my area of research. None of the features that made Rail 2000 work is present there: the region is large and has huge cities, is one- rather than two-dimensional, and has capacity-constrained stations forcing round-the-hour use of every approach track.
What this means is that there is little optimization from running slower than as fast as possible on the Northeast Corridor trunk line. However, running a fast as necessary remains a solid planning maxim on all the branches that connect to it, with their own timed connections to one another and to local buses in secondary cities like Worcester and Springfield. Most trains between Boston and Washington should run as express as practical based on station track speeds, and the local trains may plausibly only run every half hour, making them ideal for a system of timed half-hourly connections.
Pedestrian Observations 
Most trains between Boston and Washington should run as express as practical based on station track speeds, and the local trains may plausibly only run every half hour, making them ideal for a system of timed half-hourly connections.
Timed half hourly connections to what? If I’m at a station that only has Kodama service and I want to go to station with Nozomi or Hikari service the train I AM ON goes there too. I don’t have to change trains.
If you’re in Meriden CT and going to Boston, I’m assuming that the fastest way is take a New Haven local or even Hartford to NYC service and then transfer to a Northbound NEC service.
Once you get frequency high enough, I’m going to assume the only guaranteed stops are Philadelphia and NYC. Everything else would be a matter of having a clear track ahead or not.
Right now the easiest way to do it is to get on a train, in Meriden, that is going to Washington D.C.
Even if it gets downgraded to commuter service only you still run into the problem of connecting to what? The Acela to D.C. that only stops in New York, Philadelphia and Baltimore or the Regional that makes a lot of stops before getting to D.C? Or the train that goes to Cleveland via Pittsburgh? That would be quite handy if you want to go to Pittsburgh… What train is it meeting?
None of your proposed routes gets the Meridian passenger to Boston.
The people who used it to get from Boston to D.C. will want to come back eventually, Use one of those.
“If I’m at a station that only has Kodama service and I want to go to station with Nozomi or Hikari service the train I AM ON goes there too. I don’t have to change trains.”
You don’t have to, but local-express transfers are often considerably faster. Kakegawa is a station that only has Kodama service – the fastest way from there to Shin-Osaka is changing to a Hikari at Hamamatsu, which despite being a long transfer still saves around 25 minutes of travel time (a 20% reduction) compared to riding the Kodama all the way.
The Kodama is doing a lot of loitering? North America doesn’t have that problem because the only place, in some wildly futuristic scenario, there are 12 high speed trains an hour, in each direction, is between New York and Philadelphia. Almost all of the route has four tracks now and could have six easily.
The idea is that the Kodama is timed to meet sundry regional lines, like Shore Line East, or New Haven-Hartford regionals, or New London-Mohegan Sun-Norwich, or Wilmington-Delmarva.
There’s never going to be a Delmarva train, there aren’t enough people. There is no traffic, they all own cars and they can drive to the parking garage in Wilmington. Raise the toll on the Delaware Turnpike another buck they could have free Super Shuttle style service.
There are 11 million people in Southern New England. They get Japanese levels of urges to use high speed trains it’s four trains an hour. The demand in Meriden, is for New York, Philadelphia, Baltimore and D.C. not Princeton Junction. It can meet the New England Nozomi.
Scads and scads of capacity to use the tunnel between Manhattan and Farmindale for turning the Springfield-New Haven local into the Wilmington slow train…. Farmingdale is as far away from Penn Station New York as White Plains Airport.
It gets tangled in how much money they piss away. Wilmington is almost exactly halfway between New York and D.C. They can spend enormous amounts of money straightening curves in downtown or they can have a few trains use the existing commuter tracks and have the express trains blast along the swamps near I-495 and the freight bypass. 30th Street is good enough because everything is going to stop in Philadelphia. If there is enormous demand for City Hall they can run separate trains that terminate at Suburban. Or they can run through Suburban to Jefferson and Temple to automagically become the West Trenton Hikari.
…. the trains that are not stopping there don’t have to pass through the platforms. They can spend enormous amount of money tearing down Downtown Elizabeth so trains that aren’t stopping there are passing through the side platforms. Or there can be in a short tunnel under downtown. You can “fix” the Metchuen curves with another short tunnel. Trains that are NOT stopping in Metropark don’t have to pass through the platforms. A slightly longer tunnel, in 2090, solves problems with the bridge at New Brunswick only having four tracks.
In Spain, the regional and interregional run slower conventional commuter trains (max about 75mph that I’ve clocked) and ultra-velocidad trains (200mph+) along separate but roughly parallel track systems. Small towns aren’t serviced by ultra-velocidad trains to avoid delay by stop-start time at stations. But, more numerous spur conventional train lines into mountains, such as from Barcelona Sants station to Ripoll-Nuria stations in Catalan Pyrenees and the even longer climb from Zaragoza-Goya/Delicias Stations to Jaca/Canfranc stations in Aragon Pyrenees do provide reasonable speed commuter service to connect with an ultra-velocidad train for farther destinations of Paris and Madrid. Barcelona Sants combines all trains, metro train, and busses in central thruway hub, which is convenient. Zaragoza unfortunately divided off tracks for ultra velocidad junction from more city central station at Goya Station, which has excellent surface tram connection to the central pedestrian plaza, to what appears to be landscape unfinished modern building contractor over-reach and slightly remote from city center and car-culture dominated Delicias Station. But, the two Zaragoza stations have no stops between them on a conventional Renfe train line, and trains do run frequently.
Spain’s larger train routing problem is in Basque Provinces and Cantabria where ultravelocidad train connection from Madrid to Bilbao-San Sebastian-Bordeaux is not complete, and even direct regional routing from Vitoria Gasteiz isn’t complete.
“egs of about 110 km, which an upgraded rail system can connect in an hour minus transfer time, and which it is very hard to speed up to 30 minutes minus transfer time”
Is it though? Everyone assumes Bos-Wash (735 km) can be done in 3 hr if built to a high standard, which would imply 110 km in 27 min. Isn’t that just about right for takt?
Conversely, couldn’t you do 45 min legs, with knots happening every 15 min instead of every 30 at those main stations?
720 km in 3:10 is the most aggressive I’ve ever talked about and at this point I’m saying 3:20. The cost of going from 3:30 to 3:20 is likely justified and the cost of going to 3:10 may be too, but that’s because it’s a corridor whose fourth largest metro area is almost as large as all of Switzerland combined and whose largest is almost as large as #2-4 combined.
You can do 45-minute legs, yes, but 15-minute knots wouldn’t really work. First of all, a decrease in travel time from 56 minutes to 41 isn’t going to double ridership on the trunk. And second, the point of the half-hourly Takt is that the connecting services all get to the main stations at the same time, and those services (to Biel, St. Gallen, Luzern, Schaffhausen, etc.) are definitely not coming every 15 minutes. I can see a world in which 40-minute legs and 20-minute Takts may be viable, but that’s with a lot of future growth and a lot of rail-first policy that would make SVP very sad.
I am all here for a 20-minute Takt and rail-first policy but sadly don’t see how that would be possible with Ölbert Rösti in charge of the Departement of Transport now for the next few years.
45 minutes is Wilmington and New London-ish. Not particularly useful.
…and get a grip. It’s not the Tokaido and all of Southern New England is going to generate demand for 4 trains an hour. So is Washington-Baltimore. The high demand part is between Philadelphia and New York and there is enough space for six tracks. Sometime in the distant future.
Yes, knots (or nodes) are possible at half the takt-time, so every 15min for a 30min interval. So, if you find a way to cut Zürich – Bern travel time to 45min (not 47 or 51…) while keeping the node at Olten, a small town nobody would know if it weren’t one of the most important railway interchanges of (Northern) Switzerland …
You can have knots at half the Takt-time, but you’re not allowed to have triangles – all cycles must have length equal to the full Takt time.
You can still have a triangle if the sum of the sides is divisible by takt time. So if you sped up both Zürich-Bern and Zürich-Basel to 45min you could have:
(:15/:45) Zürich – 45m – Bern (:00/:30)
(:15/:45) Zürich – 45m – Basel (:00/:30)
(:00/:30) Bern – 60m – Basel (:00/:30)
“two platoons over a period of even minutes each”
I’ve always wondered how this works in practice. I understand the concept of pulses an trains arriving/departing “every 30 min”. But it is impossible for multiple trains from the same track to arrive simultaneously, so they do not all actually arrive at :30. Four trains over 10 min implies that even with a moderate dwell, train #1 from Bern is leaving Olten before train #C from Zurich heading to Biel arrives, so you miss that connection. For all four trains from each direction to benefit from the knot, it would seem dwell times for train #1 are 14+min (#D arrives at 0:14, plus a few min time for those passengers to get to #1) which is absurd if the total trip time is :28 for takt.
Is the idea that the knot only helps connections to other knots, i.e #1 from Bern lets you connect to #A to Zurich or #W to Biel, #2 to #B & #Y, etc. If so, what is the benefit of the knot, if each train only really connects to one or two other members of the platoon wouldn’t it be better to spread the departures out evenly (the :00 trains all meet, the :08 trains all meet, the :15…) so someone arriving at the station at 0:12 doesn’t have a long wait because they missed the platoon?
It seems like there’s a sweet spot for Knoten, which is 30min frequencies. Trains running at hourly frequencies or worse would obviously benefit from time interchanges, but tweaking journey times to make an hourly Takt is much more challenging.
And once you get significantly higher frequencies, the benefits of time interchanges begin to decrease. If the timed interchange results in a transfer time of 8min, then this isn’t really much of an improvement over untimed transfers with trains running at 10-15min frequencies.
So to echo Onux’s question: would Switzerland be better off if they ran the 8tph between Zürich and Olten at regular 7.5min intervals rather than trying to jam all the trains into a half-hourly Takt, which ends up having a 16min spread of trains anyway?
On trunk routes like this, generally the past has built a two-track line. If you can reliably run trains on time, you can treat them as two parallel single-track lines and schedule a train of the two per platoon per direction to use the “wrong” track for the first one or two stations.
“ schedule a train of the two per platoon per direction to use the “wrong” track for the first one or two stations.”
Yes, but the importance of the takt only applies at the knot station at the end, not the first few stations. With enough approach and platform tracks you could do this in reverse and use both tracks past the midpoint after passing the other platoon, and get four arrivals in about 3-4 min, but Alon clearly said Swiss Railways have four trains in 11 min.
Easy: All trains arrive before :25/:55, wait and leave after :00/:30. That’s all! 😀
Or you have one track from each direction, which isn’t unheard of in German-speaking countries.
You didn’t define “moderate” 😀 , but take an example from now from DB Navigator: Zürich -> Olten has:
12:04 IC to Geneve (taking 34min)
12:06 IR to Bern (needing 49min)
12:08 IR to Basel (change at Aarau to S-Bahn, 45min)
12:30 IC to Lausanne (31min)
12:38 RE to Aarau (change to RE, 46min)
12:53 IR to Bern (35min)
The arrival is likely the reverse, with the IC overtaking the IR.
This shows: trains leave (and probably arrive) every 2min, don’t ask me about that last IR there, and …
… yes, in extreme cases, like abovementioned three trains per track, you get extreme dwells. But if you start with a 1-hour-interval, that seems quite good when compared to a 55min transfer time between trains.
Better for whom? Better for the people travelling all the way from Zürich to the faraway metropolis of Olten, but certainly not for anybody who has to change from an :08 train to a :00 train.
And if one arrives at :12 at the station (Why would anybody do this?), what use would they have for an :16 train, if that goes not where they want to go because the :00 or the :08 train goes there?
Apropos of this, 20 years(!!!) ago I hacked up some code that scraped the once-ok SBB.ch web site for train arrival and departure times and generated diagrams showing station platform occupancy.
http://www.pobox.com/users/mly/SBB/
It seems I last run this in 2012, so http://www.pobox.com/users/mly/SBB/20120410-anab-Olten.pdf is only suggestive of present day operations, but gives a pretty nice clear picture anyway. (Unlike, say, in USA USA USA, there have been improvements in service every single year since then.)
There are a whole bunch of other April 2012 station platform occupancy diagrams in that directory, and if that’s the sort of thing you enjoy, you may enjoy perusing them, God bless you. Takts are immediately pleasingly visually apparent. (I personally for decades have been fond of the clever setup at Arth-Goldau with alternating Luzern/Zürich direct/connecting service.)
In the meantime, the SBB.ch site has gone to hell, while the supposed data portal “öv-info.ch” https://www.xn--v-info-vxa.ch/de is an example of just about every single type of “modern web design” fuckery and is almost totally unusable and randomly laughably incomplete. Just a total fucking disaster, but, hey, you have a button on every page to choose “dark” or “light” page backgrounds, so somebody got paid for something. Death to everybody involved! Anyway, if I had the motivation or energy or the non-suicidal inclinations to do this over again I’d massage the (seemingly OK!) GTFS data from https://opentransportdata.swiss/en/cookbook/gtfs/ but I don’t, not remotely, so selected Swiss node’s April 2014 platform occupancies are what you get, frozen in time. Enjoy!
Switzerland wasn’t the origin of the Integraler Taktfahrplan, “merely” the perfecter.
The Netherlands was the innovator, as I’m sure Max Wyss will chime in to note.
And it’s not just Takt that makes an Integraler Taktfahrplan, it’s Takt und Symmetrie.
BTW FYI Marco Chitti has a nice didactic exposition on the hellsite “The Val Pusteria/Pustertal valley rail line is a great example of the importance of two fundamental concepts of service planning in sparsely populated areas:
– Regularity (🇩🇪Takt or 🇮🇹 Cadenzamento)
– Symmetry, which is necessary to ensure coordinated schedules”.
(Of course he has a nice exposition! He always does.)
Yeah, I keep procrastinating plans to talk about that specific thread of his (also on Substack if you want a version from a different hellsite).
“in Japan, the current speed of the Shinkansen is also as fast as necessary”
This is mostly the case, though sections like Morioka-Aomori or Fukuoka-Kagoshima are only as fast as legally permitted (the 260 km/h specified in the 1973 Basic Plan) when long-distance through traffic arguably necessitates higher speeds.
“Taiwan is so linear that a single high-speed rail line connects all of its cities, with higher ridership per capita than in any other country I have data for except maybe Japan”
If we’re comparing HSR ridership per capita based on 2019 figures:
– Taiwan High Speed Rail: 67,411,248 trips / 23.6 million population = ~285,640 rides per 100,000
– Shinkansen: 370,451,000 trips / 126.6 million population = ~292,620 rides per 100,000
> Even in the Netherlands, this isn’t quite tenable. Amsterdam-Eindhoven trains come at regular 10-minute interval, each third train requiring a connection at Utrecht and the other two trains in three running directly.
These 10 minute frequencies do cause some timetable challenges elsewhere. Previously there was a 15 minute frequency, so some transfer times increased by 5 minutes or weren’t possible anymore. Eindhoven – Sittard has a 10-20 spread instead of a 15-15 spread. The Netherlands is a bit in between “takt/timed transfer service” and “metro service” in this sense. But the ridership gain from introducing metro-style service between the large city pairs is much larger than the loss from the smaller city pairs that got worse transfers or inconsistent intervals.
The vision for the future is to further increase the frequency on the Schiphol/Amsterdam C – Utrecht – Arnhem/Den Bosch corridors. The end result consists of 8tph Utrecht-Schiphol, 8tph Utrecht-Amsterdam C, 8tph Utrecht-Den Bosch, 4tph Den Bosch-Breda, 4pth Den Bosch-Eindhoven, 4tph Utrecht-Arnhem non-stop, 4tph Utrecht-Arnhem semi-fast. The 4tph frequencies would be neat 15 minute intervals, the 8tph ones not necessarily.
With this, some of the transfers lost due to the 10 minute frequency can be restored again, so you get best of both worlds: both a takt/timed transfer service outside the Randstad, and a metro style service in between the major cities.
Hi Alon. I really appreciate reading your summaries and analysis of global rail systems. Could you recommend some books that are accessible to non-professionals that discuss the intricacies of Swiss rail planning, Istanbul metro construction principles, Japanese HSR principles, and the other topics that you study.
I’ve asked them this question too. The answer is there isn’t one other than this blog since Alon’s had to accumulate their knowledge via bits and pieces. Although the Transit Cost project’s Istanbul report by Elif Ensari has that particular metro-construction story.
Also many rail cultures are quite insular and the Anglosphere is among the worst which is a problem for global comparisons given English is the lingua franca. Alon’s superpower is sophisticated cross-country comparisons, whereas most people just unthinkingly live within their own railway culture bubble.
If Britain wants to fix many of its dysfunctions it needs to learn from other countries.
Or at least that is the easiest approach.
Certainly the next government is going to need to care about value for money if it wants to achieve results.
To learn from other countries properly would mean admitting Britain isn’t great. And if Britain isn’t great then….well. And improving things has consequences….if you construct your state on the basis that some people, communities and ways of life.
Successful passenger railways are based on urban service economies organised by competent governments administered by technically-professional state employees.
Everything in sentence goes against the ideologies of the British national project. Because that project is based on idea that urban England is worthless and if it refuses to die it must be punished until it submits to true Toryism/Socialism/Celtic supremacy.
Sorry I went on my pet rants, but the whole “Great British Railways”/William Review fiasco kind of gives the game away. I was literally turfed out of a public consultation when I dared ask a Civil Servant attached to the project why Great British Railways was only fully in charge of English railways. The Civil servant was pointedly a Scottish expat who spent 5 minutes talking about how great Scotland’s electrification campaign was going*.
The Williams review was such a joke, it detailed the various other successful 1st world models, SBB, DB, SNCF, JR etc (not especially well, whoever they hired was less sophisticated than this blog’s comment section)…. and resolved to go to the Overground model of pseudo-privatisation with Johnson swooping in to make sure that the word England was nowhere to be seen (he renamed all the England-only agencies “national” and spiked the census as Minister for the Union).
So many of these problems pre-date privatisation/Thatcherism, Britain’s railways system was and is highly insular because the industry can’t admit to its own failing (see Christian Wolmar or Gareth Dennis), the Civil Service is staffed by middle-aged know-nothing consultants and the major parties all want go back to their own golden age which weren’t golden at all. The only reason we aren’t Argentina is that London’s massive legacy rail network and agglomeration economies pay for everything.
*Yes Scotland is cheaper because its been chugging along consistently but that’s because Scotland’s budget is protected from Treasury discretionary cuts by the block grant system that is part of the Sewell Convention (Brown knew what he was doing).
Or put crudely can you learn/love others if you can’t bare to learn/love yourself?
Donald Trump ran successfully on Make America Great Again. And Americans are way more insecure about learning from others than the British.
You should definitely read “How Westminster works…and why it doesn’t”. Explains a lot about our political dysfunctions.
I think if we can pay technical experts up a grade or two and get some more people from a STEM background to work for the civil service we will improve a lot.
Especially if we stop this nonsense about moving civil servants around the whole time so they don’t learn any detail.
I fail to see why this couldn’t work if for example the train between Zurich and Bern in fact took 45 instead of 60 minutes. You can simply arrange the arrival and departure time of trains at Bern to be around :45 and :15 every hours, then they will reach Zurich at :00 and :30
It would take additional funding to upgrade the line to allow 45-minute travel between Zurich and Bern. The Swiss decided that the money was better spent elsewhere.
I mean it can be any arbitrary number, as long as it is a constant then timetable in local area can shift accordingly, no reason to say it must be about 60 or 90
Regarding Germany, one should mention that it’s a big country (although Americans might disagree 😉 ) and large enough for multiple layers of Integrated Clockface Timetables. Even if there is not enough demand for a train every 15mins, there is enough distance for an Integrated Clockface Timetable that serves the strings of city pairs 90-100 km apart and another, faster one that connects the bigger cities 500km apart and thereby skipping a takt. It might involve the former trains trundling with a leisurely 160km/h along a line upgraded to 220km/h for the latter ones.
Amtrak and NJTransit do 170 kph on the “local” tracks.
I am haunted by pesky pesky fifth grade arithmetic. At an average speed of 220 kph, in fifteen minutes the train has traveled 55 km. I’m not in the mood to calculate where it crashed into the train doing 160. Assuming the slow train leaves the station two minutes after the fast train, at 35km?
Specifically about Germany – Germany should aim to be a larger Netherlands as opposed to a larger Switzerland.
Germany is a big country – the aim should be to bring as many large city pairs to close to 3 hours as possible to kill domestic aviation. Sometimes to find a 30-minute journey time improvement you’ll have to find 5 minutes in 6 places and it’s not feasible to find 15 minutes in 2 places.
Germany being polycentric is neither here nor there – its cities are Milan/Rome sized. Milan Rome manages up to 4tph direct (i.e. not stopping at Florence) per hour – and this is the sort of service level Germany should strive for between its biggest cities – half hourly limited stop (current ‘Sprinter’ stopping patterns) point-to-point services with 400m trains should become the standard. As much of the inter-city network should operate on dedicated high-speed lines with uniform speed profiles – no mix of performance and (ideally) no or very few intermediate stations.
There will be lines with too many trains to concentrate all arrivals and departures around 00/30, and frankly German stations are too big and too crowded for super-short transfers. At big city nodes all services should be at least half hourly – when half hourly services meet each other they provide good enough connection times without needing to be precisely timed. For example, I envisage Munich – Nuremburg to have up to 10 tph (2 tph each to Leipzig, Berlin, Hanover, Hamburg, Cologne for example) – at that frequency the trains necessarily need to be spread around the hour. At large central stations the ideal interchange time for a granny with a suitcase is 20 minutes, not sub 10. Half hourly meeting half hourly spread around the clockface is more effective at achieving that than everything concentrated around 00/30.
Strict takt when applied too literally can be a blocker to frequency and speed, and in my view is only suitable for small and relatively rural countries like Switzerland.
Yes, agreed on all points, except that I think the strict Takt does have a general benefit: it counteracts the tendency among planners to slack and slow down the schedule whenever there’s a problem. Swiss city pairs that were connected in 53 minutes in 2004 are still connected in 53 minutes, but Berlin-Hamburg, originally 1:30 with a technical possibility to do it in 1:20, has been slowed to 1:40, while Paris-Marseille has been slowed from 3:00 to about 3:12.
Instead of a strict takt culture you just need an ‘every minute matters’ culture – and you need to decide which one as you can’t adopt both at the same time.
Additional time creeps in when you try to run a more intensive service on constrained infrastructure. If you run fast and slow services on the same line then fast services have to wait; if you have flat junctions then you need to wait for a gap to appear before you can cross. Germany is usually pretty good at grade separation of approach tracks, but is notorious for running services of different speeds on the same tracks. Berlin-Hamburg has a lot of regional stations on a two-track railway and I suspect freight too. When journey time creep happens the right solution is to build the infrastructure that supports the timetable, oh and run the right rolling stock as well.
I believe having lots of different speed trains hurts too. Having fast trains all at the same speed and slow trains all at the same speed probably helps.
Some of the slowdown is extra useful stops. For example the chiltern mainline is partly slower to accommodate more trains stopping at Haddenham and Thame parkway and High Wycombe where demand has increased significantly.
Whether also the originally planned times post-evergreen 2 were unrealistic is probably the case too. The Swiss probably did better at getting that right as their railways are much better run.
Berlin-Hamburg slowdowns are not extra stops, they’re timetable padding. Right now, the ICE pads 25%; the TGV on the LGVs pads around 11%. This matters.
I think passengers in general care about the speed being faster than driving/the plane and they also care about the trains being on time.
If the Germans need 25% padding to run an on time service on the ICE thats unfortunate but better than the trains repeatedly being late.
To be honest the UK long distance trains probably would be better off with more ability to deal with delays which I suspect is one reason they don’t appear to have sped up the east coast mainline timetable with their super fast accelerating class 800 trains.
“Berlin-Hamburg slowdowns are not extra stops, they’re timetable padding. Right now, the ICE pads 25%; the TGV on the LGVs pads around 11%. ”
Are you privy to working timetables by any chance? Or are those derived by comparing historical timetables?
In British parlance there are three types of time that one would consider ‘padding’
1. Engineering time – accounting for degraded infrastructure or likely temporary speed restrictions
2. Pathing time – constrained timetable when a fast train catches up with a slow train or has to wait at a junction for previous conflicts to pass
3. Performance time – additional time added to ‘mask’ delays
I’ve always been interested in finding out what type of padding has been inserted in European schedules. The first two types are not really padding – they are required. Only performance time is true padding, and I’m not a believer in performance time. Performance time just means on time trains hit subsequent timing points too early and hit yellow signals, which could create a ripple effect backwards (think phantom jams happening on motorways), or hang around outside stations which may block the passage of other trains. That said, anecdotal accounts of Swiss trains often arriving early suggests a degree of performance time in their scheduling (cumulative sectional running times being less than the system time and advertised time). Then again there’s another thing called public differential – the working timetable says xx:55 arrival but you print xx:57 on public timetables – that could also be a possibility.
Good timetabling (or infrastructure designed for the timetable) should also aim to minimise if not outright eliminate pathing time – again, having trains hitting yellows and reds is not good practice. Needless to say
Delays will still occur – I believe delay recovery should happen in dwells. Efficient runs + generous dwells.
Another account I’ve heard is that due to ICE fleets being mixed together, some sections of line are timed for the ‘lowest common denominator’ – i.e. 300km/h lines timed for ICE1. That said, the Hamburg line is only 230km/h which the ICE1 is capable of doing, so this can’t be the explanation in this particular case.
I’m basing this on reporting in Der Spiegel from a few years ago:
https://www.spiegel.de/reise/deutschland/ice-versus-tgv-warum-deutsche-schnellzuege-deutlich-langsamer-sind-a-1259209.html
Switzerland pads 7% relative to the technical time in practice, and the JR East Shinkansen pads 4%; one of the points made by some technical advocates in Germany, though sadly it’s still a minority view, is that if Germany completes building high-speed lines connecting all major cities rather than forcing every train to spend hours on a shared classical line, then it will separate slow and fast trains better and improve punctuality and lower the required buffer time.
(Edit: I should say – the reason you sometimes see Swiss trains arrive early is that the padding is added every few stations, rather than uniformly, on a one minute in 15 or two minutes in 30 basis.)
Meanwhile, here’s how actual competent professional successful adults go about things
To be fair 85.8% of British trains arrive within 3 minutes vs 90.2% of Swiss ones. So its not as bad as it looks.
If we had 100% electrification like the Swiss we would probably have more reliable trains I would have thought.
https://dataportal.orr.gov.uk/statistics/performance/passenger-rail-performance/
“To be honest the UK long distance trains probably would be better off with more ability to deal with delays which I suspect is one reason they don’t appear to have sped up the east coast mainline timetable with their super fast accelerating class 800 trains.”
The ECML is too busy and too complex to simply run trains faster. There are fiendish flat junctions at Peterborough, Doncaster; there’s the Newark Flat Crossing; then there’s King’s platforming; York and Newcastle stations are both pretty restrictive. Thameslink paths are fixed. Even just speeding things up by a few minutes throws the whole thing up in the air. You need a systematic timetable recast and that’s going very slowly behind the scenes. The original Virgin Trains East Coast proposed timetable (that does use Class 80x timings) still hasn’t been implemented and the industry has been working on it for years. There are tonnes of non-compatible political aspirations and commercial requirements to satisfy on top of the all the infrastructure constraints.
“I’m basing this on reporting in Der Spiegel from a few years ago”
Not a bad piece of journalism from a general outlet. Still not entirely clear what type of slack is in the timetable. It looks like at least some of it is pathing time due to infrastructure constraints – it’s not true slack – if you removed those allowances trains would crash into each other.
Vagonweb suggests many Hamburg – Berlin – Munich services are worked by ICE-4s and ICE-Ts which are limited to 230 and 265km/h. Common timings between services with stopping patterns suggest everything is indeed timed to the slowest traction, so you get some ‘slack’ in ICE-3 workings. Again, this is really systematic buffer, just sloppy rolling stock procurement and deployment.
“…one of the points made by some technical advocates in Germany, though sadly it’s still a minority view, is that if Germany completes building high-speed lines connecting all major cities rather than forcing every train to spend hours on a shared classical line, then it will separate slow and fast trains better and improve punctuality and lower the required buffer time.”
I mean that’s such an obvious point it baffles me that this is a minority view in Germany. It’s as though after the very successful early high speed lines the Germans all decided ‘never again’. ‘Fear of being bypassed’ from the smaller places must also explain the lack of true HSLs in Germany.
No, it is true slack – trains here routinely recover from delays in line with that amount of slack.
And yeah, Germany has this problem of being able to build – construction costs here are not particularly high and are flat (whereas in most of the rest of non-Southern Europe, they’re rising) – but not wanting to. The costs per rider of the U-Bahn lines under discussion in Berlin are pretty low, showcasing recent past unwillingness to build.
For instance, the two-hourly xx:34 Hamburg Hbf – xx20: Berlin Hbf service overtakes regional services in at least two places:
Wittenberge – pass approx. xx:01, where the RE8 has an extended dwell between xx:57 and xx:03
Falkensee – pass approx. xx:39, where the RB14 has an extended dwell between xx:35 and xx:41
So the inter-city path is fixed by the structure of regional services. RE8 is a stopping service that uses 160km/h rolling stock. Holding Berlin arrival time fixed, you cannot actually run the express through Wittenberge any later because then the RE8 won’t have enough time to complete its journey before being run through by the next inter-city. Yes, on a particular day an ICE can pass Wittenberge late and still make it to Spandau on time, but that recovery isn’t a victimless one – you have already delayed an RE8 whose delay is unrecoverable, which will delay the next inter-city in the vicinity of Berlin.
What appears as a recoverable slack in the inter-city service is not optional time in the overall network schedule. In short, my educated guess is most of the observed slack is in fact pathing time not performance time.
Okay, but the trains were running 10 minutes faster in the 2000s.
Was the RE8 an hourly service, and was there 4tph regionals between Nauen and Berlin in the 2000s?
Alon, even when a intercity network could do with more than just 2tph isn’t a Takt still the best way to go? I understand that, at high frequencies, the issue of timed connections becomes less important. But I suspect that many of the intercity pairs mentioned in the article and the comments as deserving more than 2tph probably wouldn’t support it in early mornings and late evenings, such that in those periods the lack of timed knots would result in poor connections.
I’m reminded of the example of Zurich which I think you have mentioned elsewhere and probably applies elsewhere in Switzerland. Its public transport network lays down an all-day (easily rememberable) base Takt and then supplements it with ‘in-fillers’ in the rush hours as needed. Wouldn’t this work for intercity? For example, running a 30 minute takt which reaches knots at the appropriate time (say, :00 and :30) from 6am-midnight and then supplementing it with one that arrives at :15 and :45 through the busier periods or even for longer like 8am-7pm.
Naturally the trains that reach knots at even time would probably be busier than the gap fillers. But once you get to 5 or 6tph, for example, it may be that some people would prefer to catch a slightly earlier train to ensure that they have more time to make their connection or do something else at the station, and so that congestion problem wouldn’t be as serious.
So, re rush hour: Zurich actually has very little supplementary service at rush hour, especially on mainline rail. If it needs more capacity, it runs longer trains; the typical European regional train is a set of about 75 meters, which can be run in single, double, or triple traction depending on demand (I think Zurich tops at 2*75 but Max and Richard have forgotten more of this than I will ever know).
Intercity rail, likewise, has a relatively flat peak, unless it’s being used for high-speed commuting as around Tokyo. The TGV actually has a weekend peak rather than a traditional 9-to-5 peak; ICE trains mostly run a flat Takt all day and so do SBB intercities.
You’re completely right about the knot trains vs. the gap fillers. The reason I am enamored with the idea of having the Northeast Corridor knot trains be the locals is partly that – the expresses by themselves should be more popular because saving a couple minutes on New York-Washington brings in more ridership than adding the stop at Trenton, and then the knot would make the Trenton train more popular to secondary destinations. (And of course, timed connections at Trenton itself, from SEPTA trains ferrying commuters from northeast of Philadelphia to where they get connect to New York, must be to local and not express high-speed trains.)
I’ll bite. Propose a schedule. I can imagine that the :00 and :30, leaving Washington, only stop in Philadelphia and perhaps Baltimore. Skim off the crowds that are to and fro’ing from New York. Then the Hikari leaves at :05 and :35 and the Kodama leaves at :10 and :40. You then run out of demand. What train does it meet in Trenton? The :00 express to New York, the :10 local, the :30 express or the :40 local. How much of your imagined changing trains gets skimmed off by the Hikari or Kodama?
There can be six tracks between New York and Philadelphia. Pennsylvanians changing trains in Trenton to get to New York is a job that SEPTA should do and the northbound SEPTA Trenton local can automagically turn into an express to New York in Trenton. Not changing trains is worth a stop in Princeton Junction, New Brunswick, Metropark and Newark. Where they can change to the express trains to Wall Street or Grand Central.
The SBB has been confronted for the past couple of years with two problems, which are particularly acute in the french speaking part of Switzerland. First, necessary track maintenance and renewal was deferred roughly between 2010 and 2020, and second, the demographic increase and resultant demand.
The increased demand has pushed the SBB to offer 6 tph between Lausanne and Geneva (1 InterCity1, 1 InterRegio15, 2 InterRegio90, 2 RegioExpress) on the regular schedule, to which we add the occasional Geneva – Milan EuroCity, regional trains between Geneva – Coppet (4 tph)and Lausanne – Allaman (2 tph), the InterCity 5 coming from the Jura Foot line (1 tph), as well as freight traffic (1 tph on peak, 2 tph off peak). All this on a mostly double tracked main line (the regional trains between Geneva and Coppet have a third track for their own use).
This heavily used track, with quite varied traffic has resulted in a very fragile schedule. An InterRegio being a bit late can force a regional train to wait in the passing siding, resulting in further knock-on effects. Overall punctuality is therefore between 3 and 5 percentage points lower in western Switzerland than in other areas. This is a problem with the extremely short transfers mandated by a clock-face schedule. On some transfers, if your train is even 3 minutes late, you won’t make it.
The current schedule is still based off the Rail 2000 schedule introduced in 2004, just with many more trains squeezed in. Add to this the necessity of renewing much of the track over the next couple of years and you have a big problem.
The SBB has therefore decided to completely rework the schedule for 2025. In order to bring more robustness to the system, travel times are being increased (+4 minutes between Lausann and Geneva for example), some lines are being deinterlaced (the IC5 will no longer run to Geneva, only Lausanne), and Renens is being promoted to a secondary transfer hub in order to reduce the load on the Lausanne main station. One neat thing they did was to recalculate ideal travel times based on how the engineers actually drive their trains, and not just an idealized constant Vmax. See : https://news.sbb.ch/medien/artikel/122773/neugestaltung-des-sbb-fahrplans-in-der-westschweiz-ab-2025
This has resulted in much unhappiness in the press, but I am of the opnion that the current schedule desperately needs more robustness.
An amazing thing about the Swiss rail network’s dense traffic is just how much single track exists on so many routes, and how many trains run, with what feels like heroic reliability, despite this.
In Noosphere’s example of Lausanne—Biel/Bienne there’s a 2km section of single track around Ligerz/Gléresse through which they push up to 7 trains per hour per direction. In the 2023 graphical timeable it lies between the positions labelled “LIG” and “TWN”. There’s a somewhat out-of-date (2016) outline diagram of the layout of the route on page 20 of this document, where one can spot all manner of single-track, or double-track where three or four are desirable, sections in surprisingly gnarly parts of the national network.
There have been tons of projects, past and decades into the future, to double/triple/quadruple existing and projected pinch points, or to extend double track sections juuuuuuust enough to add timetable stability or to allow higher levels of traffic. If it were easy or cheap, it would have been done a century ago. (Here’s the web site for the 7 year duration project to build a double track tunnel bypassing Ligerz, with associated station upgrades and rebuilds.)
PS There are a ton of cab ride videos on the web (here’s one place to start to descend into that infinitely-deep time sink) and it gets to the point where one is surprised to go traverse a double-tracked passing section and not see an opposing direction train. (Here’s passing through Ligerz … without immediately seeing opposing traffic, as it happens.)
The Munich – Vienna – Budapest axis is an interesting ‘as fast as necessary’ vs ‘as fast as possible’ conundrum. Internally, Austria is much like Switzerland and the Western Railway is organised as an ‘as fast as necessary’ with symmetrical nodes. However, end state speed will remain relatively slow and Munich – Vienna has little prospect of coming much below 4 hours. On the other hand, if Frankfurt – Munich – Vienna – Budapest wants to be unambiguously competitive against air then Salzburg – Vienna really wants to be ‘as fast as possible’. That said, Munich – Vienna and Munich – Budapest have remarkably few flights. It’s only when you add Frankfurt – Vienna and Frankfurt – Budapest do you begin to have something that looks like HSR volumes. But then Frankfurt – Vienna will still be realistically the wrong side of 4 hours, so it’d be hard to get more than a 50% mode shift. Probably one for the back-burner …
I’ve been looking for such an explanation for a while so thank you . However one question : why can’t we have a faster connection between Zürich and geneva there’s really no good reason that it’s 3 hours I get the corridor etc. But we have the money and technology and it would help people commute between cities . It shouldn’t take more than an hour. It’s the same time as taking a car and it doesn’t make sense to me. I get the fast as necessary I think but why not improve