More on the Deutschlandtakt

The Deutschlandtakt plans are out now. They cover investment through 2040, but even beforehand, there’s a plan for something like a national integrated timetable by 2030, with trains connecting the major cities every 30 minutes rather than hourly. But there are still oddities that are worth discussing, especially in the context of what Germans think trains are capable of and what is achieved elsewhere.

The key is the new investment plans. The longer-term plans aren’t too different from what I’ve called for. But somehow the speeds are lower. Specifically, Hamburg-Hanover is planned to be a combination of legacy rail (“ABS”) and newly-built high-speed rail (“NBS”), dubbed the Alpha-E project, with trains connecting the two cities in 63 minutes.

The point of an integrated takt timetable is that trains should connect major nodes (“knots”) in just less than an integer number of half-hours for hourly service, or quarter-hours for half-hourly service. Trains connect Zurich and Basel in 53 minutes and each of these two cities with Bern in 56 minutes, so that passengers can change trains on the hour and have short connections to onward destinations like Biel, St. Gallen, and Lausanne. To that effect, Switzerland spent a lot of money on tunnels toward Bern, to cut the trip time from somewhat more than an hour to just less than an hour. So the benefits of cutting trip times from 63 minutes to just less than an hour are considerable.

What’s more, it is not hard to do Hamburg-Hanover in less than an hour. Right now the railway is 181 km long, but the planned Alpha-E route is shorter – an alignment via the A 7 Autobahn would be around 145 km long. The Tokaido Shinkansen’s Hikari and Nozomi trains run nonstop between Nagoya and Kyoto, a distance of 134 km, in 34 minutes. Kodama trains make two additional stops, with long dwell times as there are timed overtakes there, and take 51 minutes. Shinkansen trains have better performance characteristics than ICE trains, but the difference in the 270-300 km/h range is around 25 seconds per stop, and the Tokaido Shinkansen is limited to 270 km/h whereas an Alpha-E NBS would do 300. So doing Hamburg-Hanover in less than 40 minutes is eminently possible.

Of course, major cities have slow approaches sometime… but Hamburg is not a bigger city than Kyoto or Nagoya. It’s about comparable in size to Kyoto, both city proper and metro area, and much smaller than Nagoya. Hanover is a lot smaller, comparable to cities served by Hikari but not Nozomi, like Shizuoka and Hamamatsu. Hamburg-Hanover has 12 km between Hamburg and Harburg where trains would be restricted to 140 km/h, and around 6 in Hanover where trains would be restricted to 130 km/h; in between they’d go full speed, which at the performance characteristics of the next-generation Velaro would be a little more than 35 minutes without schedule padding and maybe 38 minutes with. This fits well into a 45-minute slot in the takt, permitting both Hanover and Hamburg to act as knots.

Moreover, if for some reason a full NBS is not desirable – for example, if NIMBY lawsuits keep delaying the project – then it’s possible to built a partial NBS to fit into an hourly time slot, trains taking around 53 minutes. The cost per minute saved in this context is fairly consistent, as this is a flat area and the legacy line is of similar quality throughout the route; if for some reason the cost per minute saved is too high, e.g. if nuisance lawsuits raise construction costs above what they should be on such a route, which is around 15-20 million euros per kilometer, then going down only to 53 minutes is fine as it makes the hourly takt work well.

And yet, it’s not done. The biggest cities are not planned to have regular half-hourly knots, because there’s too much traffic there. But Hanover is in fact a perfect place for a knot, with trains going east to Berlin, west to the Rhine-Ruhr, north to Hamburg, and south to Frankfurt and the cities of Bavaria. Hamburg is at the northern margin of the country, with trains going mostly south to Hanover, but having some timed connection with trains continuing north to Kiel and eventually Copenhagen is not a bad idea.

For some reason, German rail activists, including presumably the ones who pushed the Deutschlandtakt from the bottom up while the ministry of transport was run by pro-car conservatives, are just too conservative about the capabilities of trains. I’ve seen one of the D-Takt groups, I forget which one, criticize plans to build an NBS between Hanover and Bielefeld, a segment on which the existing line is fairly slow, on the grounds that it could never fit into a knot system. It is not possible to do the roughly 100 km between Hanover and Bielefeld (actually closer to 95 km) in less than half an hour to fit a knot, they say – average speeds higher than 200 km/h are only found on very long nonstop stretches of high-speed rail, as in France, they insist. Shinkansen trains achieve such speeds over such segments every day, and even with the slightly lower performance characteristics of the next-generation Velaro, Hanover-Bielefeld in 24 technical minutes and 26 minutes with 7% pad (and the Shinkansen only has 4% pad) is feasible.

I genuinely don’t know why there is such conservatism among German rail planners and advocates. It could be that Europeans don’t like learning from Asia, just as Americans don’t like learning from Europe. There are examples of faster trains than in Germany within Europe, but maybe German advocates discount French and Spanish examples because of genuine problems with French and Spanish rail operations, leading them to also make excuses like “the trains run nonstop for 500 km and that’s why they’re fast” to avoid adopting the things where France and Spain are genuinely superior to Germany.

Nothing about the integrated timed transfer schedule idea impedes high speeds. On the contrary, in some cases, like Hanover-Hamburg but also the planned Frankfurt-Stuttgart line (already in place south of Mannheim), high speeds are necessary to make the desired knots. Moreover, where distances between cities are long compared with desired frequency, as on Berlin-Hanover, it’s possible to build 300 km/h lines and cut entire half hours or even full hours from trip times. Germany could innovate in this and build such a network for an amount of money well within the limits of the corona recovery package, which includes €50 billion for climate mitigation.

But either way, Germany is about to make mistakes of underinvestment because it’s not quite willing to see where the frontier of rail transport technology is. This is not the American amateur hour, it’s not the sort of situation where I can spend a few hours with maps and come up with better timetables myself, but even so, the plans here are far too timid for Germany’s medium- and long-term transportation needs.

The D-Takt is a step forward, don’t get me wrong. None of the investments I’m seeing is bad. But it’s a small, hesitant step forward rather than a firm, bold walk toward direction of intercity rail modernization. A country that expects intercity rail ridership to double, putting Germany’s per capita intercity rail ridership in the vicinity of Japan’s, should have something similar to the Shinkansen network, with a connected network of NBS links between the major cities averaging 200-250 km/h and not 120-160 km/h.


  1. michaelrjames

    I genuinely don’t know why there is such conservatism among German rail planners and advocates. It could be that Europeans don’t like learning from Asia, just as Americans don’t like learning from Europe.

    Remember my theory: the single biggest group of Americans w.r.t. ethnic origins is German. About 80m of them. One of them is the current president and while most Germans would disown him they do share a certain stubborn streak …

    Germany could innovate in this and build such a network for an amount of money well within the limits of the corona recovery package, which includes €50 billion for climate mitigation.

    I think a lot of the public have taken a message from this corona-crisis, that budgets and government spending has no need for the austerity mindedness they have made such a fuss over, for almost 4 decades now. But there are still plenty of politicians and even some economists who remain stubbornly austerian.

  2. df1982

    Alon, do you have a formula for turn-around times at termini for intercity trains? (I’m talking about when a service terminates and then a new one begins, not when a single service goes in and out of a Kopfbahnhof.) It seems to me that your Takt plans assume around 7 minutes, but I would think an intercity train would require substantially more than that once things like disembarkation, cleaning, decanting toilets, re-stocking the bistro car, boarding new passengers and timetable recovery are taken into account. And surely longer routes of 3-4 hours would require significantly more timetable recovery time in order to maintain punctuality than sub 1-hour routes. This would also affect how many platforms are needed at termini for these services

    But I’d be happy to hear about what best practices are in the real world.

    • Alon Levy

      The Tohoku Shinkansen turns in 12 minutes at Tokyo Station, with cleaning. But this is unusually low, and 15-20 minutes is more common.

      That said, 0:53 fits into a takt with 14-minute turnarounds, not 7-minute turnarounds. Hanover is a through-station and nothing terminates there, so a :00/:30 knot there means northbound trains get to Hamburg Hbf at :23/:53, turn in 14 minutes, and leave southbound at :07/:37. Likewise a 0:38 trip time fits into a takt with one city having a :00/:30 knot and the other a :15/:45 knot.

      • df1982

        Trying to get a better understanding of the principles behind Taktfahrpläne here, but do Intercity termini also function as Knotenpunkte? I would guess so, taking regional connections into account. Having a 14-minute spread for a pulse seems like a lot, I thought the idea was to have everything clustered within 5 or so minutes of each other. Would there be a benefit to holding a terminating train over for the following pulse? So it arrives at say :58 and then leaves at :32. Obviously it means more down time, but not an unreasonable amount if you’re talking about longer trips (in the 3-4h range), and would also prevent cascading effects from train delays.

        • Sascha Claus

          There would be a definitive benefit—for the train manufacturers! On a short line in the 3–4 h range, you would need 12.5 % more trains to have one of them sitting idle at each end and doing nothing. What a way to get the train manufacturing economy going!
          To prepare for delays, it should suffice to have one or maybe two spare trains at important stations where they can serve multiple routes. If a train is too late, it can be cut short and take the place of the reserve train or reverse and continue in the other way on its own route.
          If this happens too often, just try to run the trains on time or adapt the timetable to reality instead of having two(!) trains per route sitting idle at the terminals. (On an half-hourly interval with 34min turnaround time at each end, there are at all times two trains costing millions of euros / dollars / pounds / whatevers doing nothing!)

          • df1982

            I don’t think 12.5% down-time is too bad, considering there is a bare minimum of turn-around time at terminals anyway (14 minutes in Alon’s plan). Having stand-by trains dotted around the country to deal with routine delays is also costly, and disruptive (particularly if you have multiple train types). Surely it’s better to minimise the cascade effects of delays, precisely by having decent recovery time at termini.

            And of course you want to maximise reliability as much as possible. But if you want to minimise dwell time at termini, then that means increasing journey times (what you call “adapting the timetable to reality”) which also means increased rolling stock costs, not to mention longer journey times for all passengers, even if the train isn’t subject to delays. And if you’re running a Taktfahrplan then a few minutes extra padding can totally blow out the timetable. In this case I would argue that longer dwell time at the terminus is the lesser of evils.

      • ckrueger99

        If there are transfers at Hanover, then those trains have to arrive/leave at :58/:02, right?

    • Richard Mlynarik

      I’m talking about when a service terminates and then a new one begins, not when a single service goes in and out of a Kopfbahnhof.

      If an urban terminal station has enough platforms that trains can sit around and be serviced, there are too many platforms! That’s a poor use of real estate.

      Either the train should run into the Kopfbahnhof and reverse out as “the same” train (typically 6 to 10 minutes — see eg timetables for Frankfurt am M HB or Stuttgart HB, or for Zürich or Luzern for that matter); or there should be Tokyo levels of in-train freshening (litter collection, etc) by staff on the approach to the terminal; or it should terminate and depart “not in service” to some other location where toilets are emptied, etc, to free up valuable platform space for another train service.

      (Note that it’s not just JR companies at Tokkyo that do “in-train freshening” — I’ve encountered this in German and Switzerland, too, if not to the same level.)

      Trains are expensive capital equipment and should be in revenue service, not depreciating while crews are paid not to be in motion. Station real estate is valuable and should be used by trains in revenue service, not repurposed as a maintenance yard.

      • df1982

        Reversing out into a yard for routine turn-arounds is not ideal either, as it can take up slots in the station throat that could otherwise be used for revenue service. Having dwell-time of 6-10 minutes for long-distance services doesn’t give you much recovery time if there are delays. And in Germany at least, the major termini have ample platforms (20+ in places like Frankfurt and Leipzig), so this is not really a limiting factor.

  3. oevans82

    You compare German operations assumptions unfavorably with those of Japan here, but that is in the context of a post on scheduling, not ops. Have you done much analysis of scheduling practices in Japan?

    I recall there is not much emphasis on takt- like clockface schedules (and instead rely on printed, posted, and widely distributed printed, exact, all-day timetables – which generally adopt a format that is concise, compact, and simultaneously much easier to use than anything I’ve seen in the US).

    But regardless of a lack of clockface schedules, timed meets (knots/pulses) are definitely a thing where frequency is low enough to justify it.

    • Andrew Gilbert

      In general Japan doesn’t need to bother with takts because intercity frequency is ridiculously high along the major routes like the Tokaido Shinkansen, so if you’re coming from a smaller city and transferring in Tokyo there’s no need to plan around anything other than the infrequent line.

      I’m not as convinced as Alon of the value of easily memorized clock-face schedules. Maybe for regional rail it’s more valuable, but for intercity service you’re going to usually plan around an exact train whose schedule you write down. Even for regional rail, you can just check Google Maps for the ETA of the next few trains and plan around that.

      • Rover030

        It’s not just about the memorization of clockface schedules, it’s about the knots it almost automatically creates, which makes transferring between infrequent lines much easier. Most countries don’t have the high frequencies of Japan on their IC networks. Countries that do, like the Netherlands, need the fixed half-hourly timetable to plan their infrastructure around it, because the schedule determines whether a junction can be at grade or not for instance.

        One thing I also like about clockface scheduling is that it ensures that you have a good level of service all day. In the Netherlands for instance almost every service runs half-hourly frequency throughout the day, and on busy lines, these combine to give higher frequencies, for instance between Amsterdam and Eindhoven (1:20 trip time) where there is a 10 minute frequency.

        In countries like England or France, there are often gaps in the timetable, not to mention the US.

      • fjod

        Adding on to Rover030’s point above, there are real benefits to memorisation, which play out the more often you take the train. Repeatedly checking Google Maps and noting down times (bearing in mind four semi-random digits is not an intuitive thing to remember) is just far more effort than remembering a couple of numbers which will get you where you want whenever you want – and this is key for the leisure, irregular-shift and infrequent riders which American regional railroads do so badly at attracting. It isn’t an academic debate; I have the numbers 18/49 and 05/35 etched in my head from taking the train growing up, because I know those were the two minutes every hour that would get me to and from the nearest big city.

        Rover030: The UK has slowly clockfaced its intercity trains over the past couple of decades (regional and suburban trains have been clockface for many years). Here’s an example from LNER, and the other timetables I checked (CrossCountry, West Coast) are the same. The regularity is much less strict than I’d be used to in the Netherlands, but it’s definitely there.

  4. MucklaSSR (@mucklasuntermf1)

    IMO German rail advocates generally idealize Switzerland but tend to dislike French TGV top-down planning. They put much focus on rehabilitating upgrading branch lines falling in neglect during the Kohl-Bundesbahn years but when a NBS is on the drawing board they tend to side with environmental groups. HSR (in the rare cases it’s framed with positive connotations) is often viewed as purely a measure to remove delay and distribution causing long distance traffic to enable more “sustainable” regional traffic.

    Among them is a certain sub-group that dislikes any “new built” infrastructure past 1945 (e.g. fanatically hatred of ESTWs or ECTS)

    • Nilo

      Of course this is a problem of reusing to acknowledge scale variance. Switzerland doesn’t really need HSR because it has fewer people than NYC and is 80% the size of NY state. Neither of these conditions is true in Germany which has far more people and far more land area.

      • Car(e)-Free LA

        This is only true if you view Switzerland in isolation instead of part of Western Europe as a whole. One could just as easily say Bavaria doesn’t need HSR because of it’s size and population while ignoring the trips in/out or across it. At some point, 350 km/h service across Switzerland will be needed, at least to shorten trips from Germany to Italy, not to mention cutting travel times from Zurich to Paris, Frankfurt, Munich, Milan, and so on.

  5. adirondacker12800

    Germany is somewhat smaller than the states in the Census Bureau’s definition of the Northeast. Though if northern Maine was lopped off it wouldn’t matter much. Switzerland being smaller than New York State means they get the urge to go to Italy, France and Germany. Like Northeasterners get the urge to go to the Southeast or Midwest. Or Montreal or Toronto.

        • adirondacker12800

          But these days they don’t have to hike there. If the Swiss don’t want fast service to Frankfort and Milan the people in Frankfort and Milan can fly between the two.

          • michaelrjames

            I recall a plan to transport up to 24,000 cars a day thru the Gotthard Base Tunnel on special shuttle trains a la the Chunnel. But a quick search doesn’t yield up any new info on whether that plan went ahead? The shuttle trains were to be in addition to the freight shuttles and freight trains, and of course passenger trains.
            What’s going on Herbert?

          • michaelrjames

            That’s the old Gotthard line mostly for tourists due to its very scenic route. The interesting thing about that accident (as poorly reported as it was) is that it involved a car-shuttle that side-swiped a passenger train. Wiki:

            3 July 2020, a shuttle train hit the side of the passenger train, which was carrying 30 passengers. A few minor injuries were reported. The wreck occurred in the small town of Oberwald in south Switzerland, near the Italian border. The Matterhorn Gotthard Bahn train line, which passes through the Alps, was halted on account of the crash

            I have no idea if those shuttles are anything like the Eurotunnel shuttles which are duplex, ie. hold cars on two levels. This means they are top heavy compared to most other trains and thus possibly more prone to leaning more on bends, and susceptible to loading effects.

  6. Tatil Sever

    How does somebody traveling between Frankfurt and Stuttgart decide on whether to drive or take the train, if their destination and starting point requires some additional local and regional transit? I’m assuming driving time is probably 2~2.5 hours and door-to-door trip time of 2.5 hours vs 3 hours by transit. Does it make a huge difference if train takes half an hour less for a person who already owns a car if there is parking at the destination? My hunch is some people will still chose to drive to avoid all these transfers even if they need to deal with some traffic, especially if they need to carry luggage. Some will take the train anyways to avoid having to pay attention to the road (or if they don’t own a car). Wouldn’t price be a bigger differentiator at that point, especially if more than one person is traveling together?

    • Herbert

      Cagers by and large have a skewed perspective on both time and money they spend on their hobby…

  7. Gustav Richard Fan

    > if NIMBY lawsuits keep delaying the project
    The area between Hamburg and Hannover is one of the anti-rail NIMBY hot spots in Germany. They have strong enough political connections and got a “compromise” preventing any further developments. There is a german Wikipedia article (“Y-Trasse Hamburg/Bremen–Hannover”) about the decades long struggles to get anything done.

  8. Gag Halfrunt

    Over at RailUK Forums they think that the strategy is too ambitious and unlikely to be realised in full. In particular, the second post says:

    As just one example, the Taktfahrplan assumes that all trains will operated by rolling stock capable of reaching the maximum speed the route permits, wheras in fact Deutsch Bahn have already decided the mainstay of their fleet for the next few decades will be the 250km/h “budget” ICE4 and not the 330km/h ICE3.

      • Tonami Playman

        JR Tokai’s initial plan was to retire the 700 series in March 2020, I guess they extended it to May due to Covid 19. This Japanese pdf shows that they added 2 more Nozomi trains to the schedule for a total of 17 tph. I don’t know if it’s top speed that improved or acceleration.

        This article shows that the N700 has 1º tilt

        Click to access 51-60web.pdf

        Eventually, the Series N700 was fitted with an air-springdriven system, tilting the body by about 1º in a 2500-m curve. The system helps reduce centrifugal force felt by passengers and permits trains to run through 2500-m curves
        at 270 km/h.

  9. Herbert

    The Alpha E is an unmitigated disaster of over empowered NIMBYs precisely because it isn’t a NBS. The originally planned idea was a “Y” connecting Hamburg to Hanover and Hanover to Bremen. Largely to relieve the busy freight lines but also to increase speeds. Then they started something called “Dialogforum Schiene Nord” where everybody got their say except railway experts who know what the capacity of infrastructure is. Those tracks simply can’t take any more trains, but the NIMBYs won’t listen to reason… the channel that published this video goes into considerably more detail (In German)

Leave a Reply

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

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

Google photo

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

Twitter picture

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

Facebook photo

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

Connecting to %s

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