No Cafe Cars, Please
European and American intercity train planning takes it as a given that every train must have a car dedicated to cafeteria service. This is not the only way to run trains – the Shinkansen doesn’t have cafe cars. Cafe cars waste capacity that could instead be carrying paying passengers. This is the most important on lines with capacity limitations, like the Northeast Corridor, the West Coast Main Line, the LGV Sud-Est, and the ICE spine from the Rhine-Ruhr up to Frankfurt and Mannheim. Future high-speed train procurement should go the Shinkansen route and fill all cars with seats, to maximize passenger space.
How much space do cafe cars take?
Typically, one car in eight is a cafe. The standard European high-speed train is 200 meters long, and then two can couple to form a 400-meter train, with two cafes since the two 200-meter units are separate and passengers can’t walk between them. In France, the cars are shorter than 25 meters, but a TGV has two locomotives and eight coaches in between, so again one eighth of the train’s potential passenger space does not carry passengers but rather a support service. Occasionally, the formula is changed: the ICE4 in Germany is a single 12-car, 300-meter unit, so 1/12 of the train is a cafe, and in the other direction, the Acela has six coaches one of which is a cafe.
A 16-car Shinkansen carries 1,323 passengers; standard class has 5-abreast seating, but even with 4-abreast seating, it would be 1,098. The same length of a bilevel TGV is 1,016, and a single-level TGV is 754. The reasons include the Shinkansen’s EMU configuration compared with the TGV’s use of locomotives, the lack of a cafe car in Japan, somewhat greater efficiency measured in seat rows per car for a fixed train pitch, and a smaller share of the cars used for first class. An intermediate form is the Velaro, which is an EMU but has a cafe and three first-class cars in eight rather than the Shinkansen’s three in 16; the Eurostar version has 902 seats over 16 cars, and the domestic version 920.
The importance of the first- vs. second-class split is that removing the cafe from a European high-speed train means increasing seated capacity by more than just one seventh. The bistro car is an intermediate car rather than an end car with streamlining and a driver’s cab, and if it had seats they’d be second- and not first-class. A German Velaro with the bistro replaced by a second-class car would have around 1,050 seats in 16 cars, almost even with a 4-abreast Shinkansen even with four end cars rather than two and with twice as many first-class cars.
How valuable are cafes to passengers?
The tradeoff is that passengers prefer having a food option on the train. But this preference is not absolute. It’s hard to find a real-world example. The only comparison I am aware of is on Amtrak between the Regional (which has a cafe) and the Keystone (which doesn’t), and Regional fares are higher on the shared New York-Philadelphia segment but those are priced to conserve scarce capacity for profitable New York-Washington passengers, and at any rate the shared segment is about 1:25, and perhaps this matters more on longer trips.
Thankfully, the Gröna Tåget project in Sweden studied passenger preferences in more detail in order to decide how Sweden’s train of the future should look. It recommends using more modern seats to improve comfort, making the seats thinner as airlines do in order to achieve the same legroom even with reduced pitch, and a number of other changes. The question of cafes in the study is presented as unclear, on PDF-p. 32:
| Food and Refreshments | Willingness to Pay |
| Coffee machine (relative to no service at all) | 3-6% |
| Free coffee and tea in each car | 6% |
| Food and drink trolley | 11% |
| Cafeteria | 14% |
| Restaurant with hot food | 17% |
Put another way, the extra passenger willingness to pay for a cafeteria compared with nothing, 14%, is approximately equal to the increase in capacity on a Velaro coming from getting rid of the bistro and replacing it with a second-class car. The extra over a Shinkansen-style trolley is 3%. Of course, demand curves slope down, so the gain in revenue from increasing passenger capacity by 14% is less than 14%, but fares are usually held down to a maximum regulatory level and where lines are near capacity the increase in revenue is linear.
Station food
Instead of a bistro car, railroads should provide passengers with food options at train stations. In Japan this is the ekiben, but analogs exist at major train stations in Europe and the United States. Penn Station has a lot of decent food options, and even if I have to shell out $10 for a pastrami sandwich, I don’t think it’s more expensive than a Tokyo ekiben, and at any rate Amtrak already shorts me $90 to travel to Boston. The same is true if I travel out of Paris or Berlin.
Even better, if the station is well-designed and placed in a central area of the city, then passengers can get from the street to the platform very quickly. At Gare de l’Est, it takes maybe two minutes, including time taken to print the ticket. This means that there is an even broader array of possible food options by buying on the street, as I would when traveling out of Paris. In that case, prices and quality approach what one gets on an ordinary street corner, without the premium charged to travelers when they are a captive market. The options are then far better than what any bistro car could produce, without taking any capacity away from the train at all.
Pedestrian Observations 
There is also the half-car option which is increasingly popular in the UK – done by eliminating the seating from the cafe and just selling food and drink in bags for people to take back to their seats.
yes, I found the example of the West Coast Main Line odd here, as it doesn’t have cafe cars. In fact I don’t think any British trains currently running do.
This is probably the best compromise (or maybe even vending machines dispensing ready-to-eat boxes), but it’d only be viable east of Chicago/New Orleans.
Westbound, I daresay a kitchen is pretty much compulsory, even if cafe seating won’t be missed by the fact that one can eat at their seat/room.
JR Central still has food carts, even though they seem to be unprofitable (the ekiben are cited as the primary reason for this): https://asia.nikkei.com/Business/Transportation/Bullet-train-mystery-Snack-carts-trundle-on-despite-slow-sales2
Historically all food cars were unprofitable. Railroads in 1880 ran them because a good meal was seen as a worthwhile marketing expense. Things have changed in 140 years, so it is worth evaluating how many customers actually choose a train because of meal service. If Alon’s numbers are right (I see no reason to doubt them), then get rid of them.
“worthwhile marketing expense”
This hasn’t changed. If anything the demanding end of the market–businessmen travelling on business class–is even more demanding. On trips like Paris/Brussels/Amsterdam (future Munich etc) to London, I couldn’t conceive of HSR without such a service. Often catching early (and returning late) they need to be able to eat on the train, instead of being under time stress when they depart/arrive and being forced to grab some junk (at elevated station prices). Of course this is true for all travellers and when you factor in the alternatives the relatively modest extra cost (ie. the differential between on-train versus at-station) is worth it. This is why even Ouigo has retained it. Also the restaurant/bar car offers a welcome respite from your seat (even a business-class seat) for a longish train journey. Either the service would lose these travellers, to air for business or to bus for economy, or they would force a price drop because it has been a very long established practice. And as even Alon knows (but seems to ignore here) the economics of most travel revolves around the huge margins on business travellers (which includes a lot of tourist travel too).
Alon wrote:
What this shows is that the differences are so complicated and diverse that it would be difficult to ascribe any particular outcome (such as econometric performance) to any one thing. And all this angst over the difference of 80 seats between the TGV and Shinkansen! I mean how often are the trains running at capacity to worry about this relatively minor difference? Japan is a very singular culture so I would be reluctant to make any extrapolations to European expectations. Perhaps the bento-boxed lunch is more than good enough to compensate for no cafe, but the kind of crap take-away one buys at rail stations in most of the world is not. Not to mention that the main Shinkansen corridor services about 80m people one of the densest such situations in the world (anything remotely equivalent like Beijing, Shanghai or Pearl River Delta is not in a confined narrow and corridor). Five-abreast seating, really? No thanks. (Even the Velaro e-320 used on Eurostar which can have 3+2 (economy I presume) because they are 300mm wider than TGVs designed for all the French network. Not to mention that TGV duplexes carry 100 pax more than the Velaro!) Trains have some big advantages over flying (or buses) yet this seems determined to jettison all advantages for the most miserable ‘gains’.
As happens from time to time, Alon puts on his econocrat hat and it fuzzes his thought processes. Zombie economics takes over. They need an intervention!
Eki-ben are great, but that is mostly due to two things; the rail+property model that JR Group heavily participates in, so that the stations serve heavy foot traffic just visiting the shops (and so dedicated stores for food are a lot less big of an ask), and the general high quality of even packaged Japanese food from a 7-Eleven. (No resemblance to the American 7-Eleven in terms of quality or stock.)
In comparison, getting some warmed up frozen pastry from Starbucks, Dunkin Donuts, or Pret A Manger and trying to eat that later on your journey when it got cold again would be quite depressing. And you can barely get something that level at most Amshacks.
You forgot that Japan actually tried 3+3 aka 6 abreast seating
Not to mention, said foods are stinky and people on board trains (in Europe generally) don’t like the smell of said American fast food, as is mentioned in this travel video.
In 1880s, I would say it is less for marketing, but more for the fact that people might actually spend a long time (say, double digit hours) on the same train journey, and hence catering became an inseparable part of the experience in railway travel at the time.
On the one hand, the food at Penn Station or South Station is definitely better than what you can get on the train. On the other hand, being able to grab a drink or snack two hours in is welcome. I’ve seen vending machines on RegionalBahn trains – maybe that + ekiben is the future.
I think this also to do with the length of the trip, the shorter the trip, the less need for onboard food service. For example, on my most frequent trip, Wilmington-New York, I would just never visit the cafe car. If I were to go to Boston, roughly a six hour trip, I probably would.
Yep. For trips over 4 hours, the demand for onboard food becomes much much higher.
Is there anywhere an Acela calls that cannot handle a 14 car train? I get that Amtrak’s lack of 14 car NEC trains is a function of their preference to run a high cost, super-premium HSR service, but would seem to be a long ways off from platform length being a constraint
The Acela should be a 16-car train anyway.
They seem to think that the increased capacity of Avelia will be enough for now. At seat food service is one of the reasons they are able to charge high first class fares. Pesky pesky fourth grade math again. Half as many seats at twice the price kinda thing means they make as much money. And then are able to skim the profits from the business class passengers buying stuff in the cafe car.
It depends on fares… Acela fares are hideously high, partly due to scarcity pricing, partly due to insanely high operating costs. An Acela that averaged 200 km/h and charged $0.15/p-km would be filling 4 16-car tph all the way to Boston, and more on the NY-Philly segment. And a national HSR network, if you believe my Metcalfe’s law post, fills 12 tph all day into Penn Station (8 to Philly and DC, 4 to Harrisburg and Pittsburgh) or maybe even a bit more.
4 trains an hour to New England leaves plenty of capacity for 2 or 3 super express trains to Suffolk County, change in Farmingdale for local service on the Island. 4 is a bit low. Pittsburgh is in HSR range of Boston. Ohio and beyond would be going through Albany. Hartford, Springfield and Worcester have demand for beyond New York. Nice round numbers there are twice as many people on Long Island as there are in Connecticut. And there are more people in Hartford, Springfield and Worcester than there are in Providence. It’s all going to be the same loading gauge/platform height etc. there’s demand for two an hour to Wall Street and Brooklyn. You want this to come together by 2050 the Major Investment Study needs to start now.
New London is basically a glorified step box.
Between New York and Washington, I don’t think there would be problems. I don’t know about Route 128, but there should be no issues at Providence, Back Bay, and South Station.
The most common route I take on Amtrak is Pittsburgh to either New York or Philadelphia. I’ve also done Pittsburgh to Chicago, and then the next train on to Los Angeles. These routes are so long that a cafe car/dining car is necessary for me to be willing to take the trip by train rather than fly.
Your argument seems reasonable for shorter routes though.
I believe that for America, this is supposed to be in a world where trains run at an average speed higher than the highway speed limit.
I remember hearing that some carrier, probably Amtrak I believe was considering replacing meal service with vending machines. Would that be a reasonable compromise?
In 1964 when I was in college and interested in passenger car lengths for stations, I was told that in the Northeast corridor (which wasn’t called that then) the platforms could accommodate 17 cars of 85 foot length (besides the locomotive power).v
In this case, the “airline at flight level zero” approach gives the correct answer for the correct reasons (floorspace is expensive in both types of vehicle).
No, the basic economics works out a little different.
For one, there’s just more space available on a train, so optimizing it is certainly important, but not as big a deal; trains are comparatively more lenient with space than the race to the bottom that airlines have been racing with for seats and toilets, particularly with the higher charging they do for premium economy/exit row.
The other thing is that the labor component on airplanes is fixed. There is a fixed ratio of attendants to passengers for safety reasons, and so that (and sometimes short time windows where one is allowed to move about the plane) biases towards having all the flight attendants do one big meal service using tiny galley carts; if they’re there you may as well use the manpower. On a train you can generally walk about whenever and however you want, there’s no such labor to passenger requirement, so a dedicated cafe/food area with lower staffing levels makes more sense.
Took Amtrak NYC to Buffalo a few times about fifteen years ago. Ignoring for a moment that all but once the train was between three and seven hours late, and that the ticket price wasn’t much less than the cost of air fare.
The restaurant car, on the trips where the train had one, was grossly overpriced and required eating within a particular time window. It seemed like a giant waste of space and money.
The cafe car sold boxed lunchables, snacks, and drinks almost three times the supermarket rate. For subsequent trips, I boiled some potatoes and eggs, the way we used to when traveling long-distance in Russia via train. They were better than anything Amtrak was selling. Also the train stations in Buffalo are basically in the middle of nowhere (Depew) or in the middle of a parking crater underneath a highway interchange (Exchange Street); neither had any place to buy food within walking distance. Maybe things have changed, but I refuse to take Amtrak principally because in my experience Amtrak can’t guarantee that you’ll arrive on the day promised, much less the hour. If I did have to travel Amtrak now, I’d do what I do on long flights and bring a sandwich or three.
Penn Station *had* a lot of food options.
They recently just closed all the retail on the LIRR Concourse to double the width. Presumably Moniyhan Shopping Mall will have retail space as well, but I don’t think it’s quite online yet, and like everything else in the new hall is not where most people are going.
Some more European data points: Renfe’s 100-meter CAF trainsets (Class 120) have a half-cafe, half-1st-class car out of 4. With Talgos, the cafe car only takes up 1 out of 12 cars — do you think this is a reasonable middle ground? The newest ones currently in service (the 200-meter Class 112) have 365 seats with 3 first-class, 8 second-class and 1 cafe car, with 3-abreast seating in 1st class and 4-abreast in 2nd. For the upcoming low-cost Avlo service, they modified 3 Class 112 trainsets by turning the cafe and 1st class cars into 2nd class seating, bringing the total capacity to 438. The funny part is, SNCF will start running Ouigos in Spain this summer and they will have a cafe car (only the upper level, I think) — their advertising says “In Spain we love bars so much, we’d rather take away the seats”.
The future trainsets will increase the capacity to 521 (AVE) and 581 (Avlo) by introducing 5-abreast 2nd class seating thanks to a wider carbody. Not bad for a single-level train, but still far from the Shinkansen…
My personal opinion is that cafe cars are fine when there is a clearish distinction between long- and short-distance trains, e.g. probably Spain and France, but not Germany or the Benelux. For a 5h trip I’d like to have the option of a hot sandwich/noodles/whatever, which I can’t get at my origin station because they’d get cold too soon, and on a moving train liquid food can be messy. (you could have a communal microwave inside the train, but ew). I also like to have a place to talk to the people I’m travelling with without the volume being rude to other passengers, but that’s maybe a Southern European thing, haha. However, for a 1h trip, I’d rather have a vending machine and a large bike rack.
Do you have data on the economics of onboard cafes in other countries? Renfe doesn’t only lose revenue from not having seats on the cafe car, but IIRC the actual bistro service doesn’t even break even. That reinforces your argument for getting rid of the cafe car, cultural aspects aside.
Even tho not necessarily all high speed trips in Germany are the range you describe, virtually all ICEs have total itineraries where one might get peckish – six hours is the lower end…
Yep, what I meant is that due to Germany’s population distribution, the same rolling stock has to cater to different needs, because (for example) the same ICE serves Munich-Ingolstadt, Munich-Nuremberg and Munich-Berlin passengers. I suppose that is also the case in Italy. Whereas in Spain with its donut-plus-Madrid population, the high-speed rolling stock is divided between AVE/Alvia (2 classes and cafe car) for long trips like Madrid-Barcelona, Madrid-Galicia and Barcelona-Andalusia, and Avant (single class and no bar) for short services like Madrid-Toledo, Barcelona-Girona and Granada-Málaga. The extreme diversity of rolling stock causes an operational nightmare for Renfe, though.
Spain could run their trains through Madrid. Atocha and Chamartin aren’t dead end stations.
Granted, capacity and reliability might be issues (how many standard gauge tracks link those two stations or rather are under construction?) But other than that there are only upsides to running trains Galicia-Madrid-Catalonia instead of all ending in Madrid
Yeah, I agree, but citing our former prime minister: “It’s very difficult todo esto”. The two-track standard gauge tunnel should be open soonTM but right now, even if the tunnel were open, there are too many constraints (apart from the demand mismatch between the Chamartín and Atocha lines): there’s limited rolling stock that can run on the currently unelectrified portion of the Galicia line; there’s also a limited quantity of trains capable of running on the French network; some trains cannot use the Guadarrama base tunnel, for safety standards reasons; and of course AVEs are captive to the standard-gauge lines. See what I mean by operational nightmare? Liberalization will also add some new issues — the new private operators cannot run trains on the Andalusian line because they can’t find suppliers and technicians for its vintage version of LZB, and there are regulations preventing competition on subsidized commuter services like Madrid-Toledo and Barcelona-Girona.
Hopefully through-running will become the norm when the network converges to standard gauge + 25 kV AC + ERTMS and the rolling stock becomes more homogeneous. We’ll have to wait to ¿2030? to see if that homogeneization actually happens and wheter it entails changes like eliminating the cafe cars or not.
When they introduced the double deckers InterCity in Switzerland, the trolley cart was a logistical puzzle, as the through passage is on the top floor, and so there needed to be a lift to bring that thing up, and a special announcement for the lower floor when the thing was on the upper floor. I was under the impression that the trolley depended on the restaurant carriage for things like water.
The trolley cart and restaurant are two different operations.
I worked as a trolley cart steward for SSG, eons ago, when coffee and hot water was held in big thermos. More or less the only thing I could get from a restaurant car was replacement of the hot water (yeah, coffee cream and sugar too).
I’m agnostic about eliminating cafe cars. Alon does raise many great points, though. The car could accommodate more passengers, the concession items tend to be unprofitable, and most of the items are just convenience store snacks anyway.
The cars could be eliminated on the condition that passengers could get the same items within the station, or the station itself is within an area with a convenience store or fast-food option within 500 meters of the platform.
Tech also allows order-by-app food to be delivered to prospective passengers. This allows the radius of options to be expanded to a few miles near the station, and the food package can be delivered in a designated station area or an Amazon-style locker.
China has been using online ordered food delivery to trains since 2017. Passengers order from local restaurants 1hr before passing by a station. The food is then delivered by courier to the train staff within the 2min station stop window. The train staff then distribute the food to the specific passengers.
This creates even more choice than any on board Cafe car or food cart can offer. You can sample cuisine from different cities along the route.
@Tonami Playman
That does sound pretty cool. A much more efficient version of what happens on the Trans Siberian in which local food vendors swarm the platforms at the rare stops (even in the middle of the night). Just please don’t tell me it is by Uber Eats!
OTOH, presumably some of the long-distant trains are express or run long stretches without stops? What next? HSR-drones that drop the food to some elevator on the train roof?
The Chinese entry in the dick measuring contest of speed, G17, makes no stops between Peking and Nanking. That’s ~1000km over 3h15min or so. It would be interesting to see how a drone would meet a train going at 350kph.
But in any case, their security practices operate like Penn Station where passengers slowly file through security checks and onto the platform. It feels hard for me to say a 2min station stop would be reasonable elsewhere besides West Taiwan or NY Penn.
Penn Station has no security checks! Amtrak tells passengers to queue single file for a ticket check – even though there’s also a ticket check on the train – but if you really want you can get on via other parts of the station, which technically belong to other railroads’ turfs and don’t even advertise that you can go downstairs and take an Amtrak train.
Shinkansen trains have platform access control but have 1-minute stops. The difference with European trains as I understand it is 2 doors per car vs. 1, level boarding, and maybe wider vestibules, I’m not sure. TGVs have horrifically obstructed internal circulation since they intend to run nonstop, and then at the intermediate stations they stop for 5 minutes so that people can go outside to smoke. ICEs might actually be able to get similar dwells to Japan, I’m not sure – they have no level boarding, but they have easier passageways and I think the busier stations have 760 mm platforms so it’s only 2 steps up to the train.
The Shinkansen doesn’t have platform access control in the sense CRH does; sure there’s a separate ticket zone from the local lines, but once you’re past the gates, any platform is fair game, much like non-Amtrak bits of Penn as you say. For a discussion pertaining to any potential US HSR networks I’d expect them to follow this paradigm.
IIRC CRH requires you to board trains from specific gates depending on platform and train car much like an airport.
@Yukka “The Chinese entry in the dick measuring contest of speed, G17, makes no stops between Peking and Nanking. That’s ~1000km over 3h15min or so. It would be interesting to see how a drone would meet a train going at 350kph.”
All the CRH trains have at least a half cafe car. The trains serving that route have 48 seats in the cafe car compared to 90 seats in other cars as can be seen here. So they could still use the cafe car. The food delivery to train platform is a complement to the existing cafe car food service.
Also on the same route they are now running 17 car 440m long CR400AF-B trainsets with 1283 total seats. Still 40 seats less than the 16 car 404m long N700 Shinkansen with 1323, but the Chinese trainset would have 1325 seats if they eliminate the cafe car. This comes from dedicating 2 cars to airline style business class seating which the N700 lacks.
The app for ordering food is the same app used in booking train tickets. So no Uber eats.
Some 100 series shinkansen trains had double decker cars, assigned with a cafe on the lower level and green class seating upstairs. There were also other forms of 0 and 100 series trains with varying configurations of restaurant cars.
The current E5 and E7 series shinkansen also have Gran Class, a kind of real first class with at seat service. Then again those are in cars 10 and 12 respectively, which means one could theoretically park the food service carts in the employee areas of the train, probably within the very long streamline noses.
@yuuka: “Some 100 series shinkansen trains had double decker cars, assigned with a cafe on the lower level and green class seating upstairs. ”
Zut, Alon, those Japonais heh? They must have picked up that weakness from les Francais. They have broken two of Levy’s cardinal rules: cafes/restos on trains and duplex trains! They’ll eventually come down with a bad case of Paris Syndrome when they realise what horrors they have inflicted on themselves, non?
They notably discontinued the bilvels because they couldn’t build good bilevel EMUs.
It was hard to squeeze in all the electrical equipment on a bilevel HSR EMU while trying to improve performance. An E4 motor car weighs 60 tons, while the equivalent E5 single level motor car is 45 tons- an E4 trainset can only make 240km/h in revenue service. There were also issues with wheelchair accessibility, and I reckon car cleaning took longer, not to mention food trolley service.
The cafe cars were phased out as the speeds increased.
0/100 series operated at 200 km/h, and the Hikari took four hours end to end. They had full cafes.
The N700s operate at 285/300 and take two and a half hours end to end. They have the bento box cart.
If Chuo maglev opens with a 67-minute end to end, presumably they’ll have nothing. We shall see.
Oh no, it gets worse, the 200 series had similar double deck cafe+seating cars as well.
Then you have the E1 and the E4, the E4 at 817 pax for a 200m long train making the highest capacity HSR train in the world. But I guess your Paris Syndrome comes from replacing the E4 with single decker 12 car E7s; then again, the E7s are faster so running more of them probably works out to be the same capacity as the E4. Not like it really matters considering Japan’s declining population anyway.
And while we’re here, don’t tell Alon that double deck Green cars still operate in commuter trains.
There’s a difference between doing a thing, and that thing being a good idea. For example, the NYC MTA said a few days ago that it removes station benches for cleaning. Does it make sense to unbolt a bench from the station floor, load it onto a utility train, and take it to a depot for cleaning, instead of just wiping the bench down? No. Is it happening? Yes.
@homer2101: “Does it make sense to unbolt a bench from the station floor, load it onto a utility train, and take it to a depot for cleaning, instead of just wiping the bench down? No. Is it happening? Yes”
Sack Cuomo NOW!
Speaking of the double deck green cars on commuter Rail, do they only stop at stations with multiple platforms like shinjuku and Tokyo? I would imagine them throwing a wrench into the schedule if they had to stop 2 track, 2 platform stations also.
Yes, the trains with double deck green cars stop at most of the stations, they are found on the long distance suburban services such that run on the Tokaido, Yokosuka/Sobu and Joban lines, to serve passengers who want a seat for their 1 hour plus commute. These green cars are extra fare and have lower capacity (no standees) so they don’t cause extended dwell times.
Even so, most of them only have a single leaf door. The Chuo line doors are double leaf, apparently to keep dwell times lower especially since it’s a 30tph operation.
Still not an issue, as most green car users are getting on at the start of run in outer suburban/exurban locations, and getting off in the urban core. Almost nobody is going to board a green car at Yokohama or Kawasaki(28 minutes and 18 minutes out from Tokyo respectively) on a service bound for the capital.
Not a fan of the idea of getting rid of café cars. Particularly for longer intercity journeys. If you have HSR city-pairs where the door-to-door travel time is marginal with flying (i.e. in the 2.5-4 hour range) then the principle way to win modal share is through the comfort advantage. A big part of that, of course, is the fact that you have a seat from one city centre to the other (rather than spending the time dealing with airport transport, security queues, etc.), but it also comes in the form of more comfortable seats (so no airline-style scrimping on these) and having an in-train dining experience that is superior to the tray meals dished out on plane. Buying food at the station doesn’t cut it for journeys longer than a couple of hours, because you might not be hungry at 5pm, but at 7pm or 8pm you definitely are.
And I’m unconvinced about the capacity argument. How many lines outside of Japan are selling out 16-car trains running 15tph? It’s only at that point that junking a cafe car for capacity reasons should come into consideration (since the only alternative is building a second line). Until then you can always increase frequencies and train lengths to accommodate any capacity crunch.
HS 2 projects 13 400-meter tph and 4 200-meter tph departing London when the full system is open. My ridership formula says it won’t need that, but Britain already overperforms that formula and I suspect my formula just lowballs short-distance ridership, which Britain will have in droves since London-Birmingham is less than an hour.
The LGV Sud-Est has 12 400-meter bilevel tph at the peak, with a confusing array of service patterns none of which has more than 2-3 peak tph.
And right now the busiest ICE segments, Frankfurt-Mannheim and Frankfurt-Cologne, are projected for just 4.5 tph in the D-Takt, but that’s in context of some really sluggish speeds on key connections and no Frankfurt through-service, and if those are fixed you can plausibly create enough demand for 12 tph on Frankfurt-Cologne.
Those service patterns are with café cars though, and presumably they aren’t (or won’t be, in the case of HS2) consistently selling out either (I think you mentioned TGV has a 70% occupancy rate). You could, I suppose, have café-less sets for short journeys like London-Birmingham or Paris-Lyon as a minor capacity gain, the question would be whether it’s worth having tailored rolling stock for this purpose. I suspect not. Just let people have comforts like enjoying a nice dinner at a table while the countryside rolls past, and stop trying to turn intercity rail into Ryanair.
One compromise solution, however, might be, for systems like Deutsche Bahn which don’t have mandatory seat reservations, to have bistro seating, but to let passengers sit in those seats without ordering anything if there are no seats free in the rest of the train. This often happens informally on ICE trains, though it’s up to the discretion of the conductor (since it’s technically not allowed).
“having an in-train dining experience that is superior to the tray meals dished out on plane”
Yeah, but who gets their food from the plane anymore? We just buy our food at the airport, and eat it on board. It is better, cheaper, and you have a much wider variety of options. I think it only makes sense to have food onboard when you have a really long trip (12 hours). Otherwise a few snacks and a few beverages is fine.
Correction: the cafe car on the ICE 4 does not constitute an entire car; there are 20-odd first class seats in that car. It would be more accurate to say that the dining functions consume about 1/18, i.e. about 6% of the passenger space of the train.
I would note that the double-deck InterCity 2 trains do not feature a cafe car but rather only trolley service (provided by Skychefs), an omission that has resulted in considerable criticism from passengers and press.
I’m afraid that I must side with the critics on this post: elimination of this service would be poor policy and would have a negative financial impact. While your engineering viewpoint makes a sound argument, too little attention is paid to the effect that on board services have on demand elasticity — the cited Swedish study notwithstanding.
One possibility to keep the cafe section would be that it is also a contact point for other railroad “commercial” things (as opposed to railroad “operational” things which have to be dealt with by the train chief.
For shorter (or smaller) operations, a corner with vending machines (snack/drinks and coffee) does work out well. For example, the Traverso trains of SOB (8-car FLIRT with increased comfort) has such a corner in car 3 and 6. Also many Japanese trains do have some vending machines on board.
It would be great if there were a Bento/Ekiben culture outside of Japan, but that may take quite a bit of time. And Ekiben is the complete contrapoint to the chain stores “culture”, as Ekiben very often have a local touch (using local specialities). I would, in any case, prefer a local Ekiben store over the Brezelkönig or whatever they are called.
So on a longish (3hr +) trip there is a chance that you will want something to eat that you haven’t already planned out before the trip (or maybe you just got to the station in time to run to the train, or maybe you got on at some smol intermediate station like Wilmington or Route 128). If the proposal is to have people jump off the train at intermediate stations, grab something to eat, and then jump back on, (1) I’m not sure how you manage hot food with reasonable dwell times, (2) regardless of what you do you are increasing the dwell times. (Also on the NE Corridor you would really have to reconfigure the stations completely, since the food is nowhere near the platforms, and even fairly important stations like Back Bay and Baltimore Penn have nothing in the station but a Dunkin’ Donuts.) The cafe car addresses this problem, though you can imagine more space-efficient solutions like an on-board vending machine with a microwave or whatever.
Chinese high-speed trains (and to a lesser extent Japanese ones) has vendors that walk through the corridors and periodically sell food. Basically, they offer the same kind of food as the Swedish high-speed (X-2000) cafeterias (that are very far from old style restaurant cars, and barely have seating space, they would really not be a loss to get rid off). As long as there is space, a good tray to eat on, and some area where vendors can prepare hot food I think this is totally fine.
As long as you can get some coffee, fruit, sandwich, a glass of wine, and some basic option for hot food, and this is offered every 30/min, I think this would satisfy most people. As the logistics is easier, the quality can easily be better and cheaper than an airplane. Also It is not like contemporary European or American cafeteria cars are like the oriental express in any case, where you actually “want” to eat, if you can plan ahead you always buy a sandwich on a station instead. Only offering food on platforms I do not think is a good compromise. In particular, as the case you really need those food options is during train delays.
For high passenger systems, with multiple trains every hour and large volumes, I think a solution like Tonami Playman suggests makes lots of sense, and is easily combined with a system with walking vendors. I assume this is partly a reflection on that the Chinese system has some really long trips (e.g. Beijing and Guangdong though).
That said, if your train system really has train platforms that can handle 16 car trains, I think a cafeteria-car can often be a good idea. But if you run shorter trains due to platform constraints (more common than systems with capacity for 16 cars), I am very sympathetic to removing Cafeteria-cars and lowering prices instead. On legacy night trains and so on however, I think proper restaurant cars are essential on the other hand.
I agree it is dumb to dedicate an entire car to cafe service but I usually buy food or drinks on board because I am in a rush or then become hungry on board. Selling the seats in cafe cars like the UK does seems to be a good compromise between revenue and pax comfort. Some of Amtrak cafe cars offer 18 biz class seats. Even Amtrak looked at converting their Acela cafe seating into 27 biz seats although in typical Amtrak fashion, dropped a good idea. While high speed services run with higher load factors than intercity trains, only during peak hours and holidays will that extra capacity be sold. RZD leaves off dining cars on some of their overnight trains during holiday season since they need that space for additional pax capacity (although most high speed trains cannot easily drop individual cars).
Personally, I rely on the Amtrak cafe cars for most trips, most frequently between NYC and either Boston or DC. When rushing to get to the station, often with bags, the last thing I want to do is to stop to get coffee, etc. and then have to juggle it as I struggle with my bags into the station, down the stairs to the platform, find a seat, stow my bags and settle in. The availability of food takes a lot of the stress out of travel, especially if the train is early morning or directly after work or meetings. While it’s not an economic argument, there is also a real pleasure in relaxing at your seat with a coffee and muffin or bagel in the morning, or a beer and a sandwich in the afternoon and watching the scenery pass, and it is honestly one of the things I most look forward to when taking the train in any country. The food in Amtrak cafe cars can use a serious upgrading, but that’s a different issue, though I do not mind paying a premium for the convenience of on-board food.
Cafe cars on Amtrak are often busy, but most people take their food back to their seats. There’s no reason why those tables can’t be converted to seats as is done on many Amtrak Business cars which would offset the cost of some of the seats otherwise given to food service.
For me, I like getting a fresh cold beer from the cafe car when I want to. Waiting for the cart the come by or running up and down the length of the train to find it isn’t great either.
However, I don’t think cafe cars need to be full length of the car. 1/3 – 1/2 length feels perfect as you only need to provide enough space to wait and table space to get organized.
In California on the double deck trains, the cafe is on the less desirable lower level location where it takes up maybe 25% of car capacity.
You can run a train with more passengers than seats. You can’t do that with a plane.
Sure, there’s obvious downsides and comfort losses to passengers going without seats, but a) those are borne by passengers who didn’t pay for first class and/or a seat reservation b) the alternative is not going on that train at all.
I think the comfort loss for every trip by losing the cafe car is greater than the comfort loss for those marginal pax on marginal trips where the train would have seats for everyone if it didn’t have a cafe car…
And then we’re starting to enter the territory of: who pays their fair share? First class or second class?
I don’t understand the idea that length is a constraint. If freight companies can run 112 car long trains, than Amtrak can push past 8. The cafe car does not need to be in the center, it can be off platform.
You cite the NE Corridor and Philly-NYC, but there are NEC trains that run 11+ hour.
One of the big benefits of trains is being able to arrive 3 minutes before departure time. Having to show up 20 minutes earlier to buy food hurts that.
The Swedish study you cite is apparently garbage. I know this because not a single person on earth, aside from an airline executive, would cite the new slim seats as a positive. Theyre less comfortable than the seats on my local bus route. Clearly the methodology is off. Since rail will always be slower than flying, comfort is important.
That being said, Amtrak does not run an efficient cafe car. That could be improved.
Freight trains are not constrained by the length of passenger platforms in stations. Passenger trains are.
The idea of putting the cafe car off-platform, at the beginning or end of the train, is a good one though.
It’s not, and if you want to do that you’d be better off just having carts instead, since you *can* park the carts in a service area that docks somewhat off-platform.
Cafe counters are (should be) in a place where people can actually reasonably walk to them, probably the middle – this is a bit of an extreme, but imagine if the aforementioned 17-car CR400BF had its cafe counter in car 17 (or even a potential car 18) and the poor bugger in car 1 had to walk at least half a mile roundtrip just to buy something. Or if that’s business class (top tier on CRH) they’d have at seat service anyway… from a cart.
You could have a 14 car train, and have a half business/half cafe car at each end.
Planes are extremely limited in size. Buses are limited in size. Trains literally are free real estate.
Throw in double-level trains and its even less of an issue.
Putting them off-platform only works if your tracks are significantly longer than platforms.
In many urban stations this is not the case, for instance if you look at the end of the platforms in Zürich HB (where many high-speed trais alight), the switches are literally at the end of the platform: https://goo.gl/maps/VUc7QwhmhjXDPHnP9
Alon understates the degree to which the Japanese have gotten rid of cafe cars. Its not just the busy parts of the Shinkansen network like the Tokaido, Tohoku and Sanyo, but also the smaller ones (Kyushu, Akita, Kanazawa etc). Not just that but even the longer conventional “special expresses” that the JR companies have, and I’m not just talking about the commuter ones. In particular I would point to the Inaho (a west coast express between Niigata and Akita) or the Tokiwa express that goes from Tokyo to Sendai. These are both 3 hour plus journeys on not especially crowded lines. Even the last remaining regular night trains the Sunrise Izumo/Seto don’t have cafe cars anymore (and they are as long as Shinkansen in no. of cars).
And the only 3 non-JR interregional express services are the same. Kintetsu runs trains between Osaka and Nagoya that compete with Shinkansen on the basis of cost, comfort and a terminal in Central Osaka, but they only have vending machines even on their most deluxe train sets. Kintetsu Osaka-Ise Express designed for tourists also no cafe cars or even trolly ladies (its Japan). And Tobu’s Aizu-Wakamatsu services compete with Tohoku shinkansen with comfort, scenery, cost and only a food/drink trolly. Chizu Express between Tottori and Kansai similarly no cafe-services. If cafe cars are the preference of customers these guys clearly don’t believe it to be so (and they got rid of them in the 1990’s about the same time JR did). Its not just forgone train seats, its about labour and facility costs too like adding 2 conductors for single car. I’ll wager dining cars also weigh more too and it means extra equipment to repair.
But the really revealing case is JR Kyushu. JR Kyushu unlike the Honshu Big 3 has had to claw its way to profitability while maintaining a large ageing network in a rapidly aging region. Building a Shinkansen service at the same time to boot. And yet no cafe cars on their daily expresses to Sasebo, Nagasaki etc. Indeed even on the Nippo main line connecting Fukuoka, Oita and Miyzaki on the East Coast where it struggles to compete with indirect Shinkansen, Highway buses and airplane services. But here’s the interesting thing. JR Kyushu has been the most active promoter of special dining/hotel tourism trains, these are deluxe high fare affairs usually less than 4 cars and run most during major tourist seasons. Food is not their only special feature. They are not regular services. And they run on the struggling rural lines often with just diesel locomotives. So we have a measure of Japanese customers revealed preference for speed, frequency and reliability over dining cars.
We should recall that Japan has very high mode share, a very diverse series of company structures running railway, high fare recovery and for the relevant cases no direct operating subsidy. Furthermore these services removal occurred rapidly after JR’s privatization suggesting that dining cars are reliant on a soft budget constraint. Competition is fierce between airlines, shinkansen, distinct conventional railines, private cars and highway buses. Dismissing with “its just Japan” isn’t good enough. And getting better in station food is much easier to solve then building infrastructure to up capacity. And remember Japanese rail conglomerates have more expertise on food services than any other railway operators thanks to their retail operations. If anyone could make dining cars profitable it would be them.
Your last sentence points to the problem of the argument.
SNCF runs its TGV division profitably (at least some years) and that’s with cafe cars and the horreur of duplex trains, not to forget only 2+2 versus 2+3 seating, a much bigger country with half the population. Maybe they could make a bit more “profit” by scrapping cafes, or maybe it would stop more people using the trains. At any rate it points to the folly of considering whether everything in a complex operation makes a profit.
If you are saying that competition is so severe in Japan between trains and alternatives, then something is not right. Closing facilities on trains is hardly going to win back passengers if that is really the case. And it suggests that the so-called private trains might need some help to reduce flying which is horribly polluting. Japan is the world’s biggest importer of fossil fuels per cap. The French have just adopted the policy of eliminating domestic flying.
Finally, if Japan either can’t run an onboard restaurant or build its cost into ticket prices, it doesn’t mean the rest of the world has to follow them.
You’ve assumed that removing dining cars loses passengers. There is no evidence to suggest that. Japan is the only clear evidence we have on this because they used to have them and got rid of them. I stress the conventional lines I mentioned that compete directly with Shinkansen on comfort, if dining cars brought passengers I’m sure they would use them. Passengers made the choice not to pay for dining cars and prioritised. Dining are no just a loss making, but they are genuinely costly plus opportunity cost on crowded lines/times.
And its not just about profit, there are distributional concerns about prioritizing dining cars over capacity. There is a reason that the only dining cars left in Japan are luxury rural hotel/event trains.
Also when it comes to mode share France has very little to teach Japan unless they ban cars and coaches. And islands issue makes it a fuzzy comparison. And Japan’s decarbonisation efforts used to be quite French but then 2011 and the LDP threw a 10 year tantrum over having to abandon the nuclear scheme. Little to do with train policy. Dimissing the Japanese example because “Japan is weird” is a very Anglo-saxon solipsism.
But seriously we won’t know for certain unless somebody tests it elsewhere. All I can say is the Japanese got rid of dining cars pretty much everywhere and people take trains a lot so take your dismissive nostalgia for dining cars somewhere else.
It was you who talked about profit. And it’s the theme of Alon’s article. So now you discover some other points but I don’t find any convincing. FWIW, I don’t think it’s even an issue in France and Spain, or Eurostar; ie. there is no hint of any thought of removing restaurant cars and I don’t think there ever will be. I haven’t eaten enough on TGVs to know if what Yom Sen says is true; I think it is just something people say (do they finish their complaints about SNCF food with the Woody Allen jibe “and such small portions too”?). The French of course have absurd standards: take their trucker’s cafs which often are the best place to eat if you’re travelling (on the road). Just on bread, on a train clearly a baguette cannot be fresh from a boulangere’s oven with the only thing fresh about it being freshly thawed; OTOH someone in France buys all those freshly-thawed baguettes in the supermarkets.
Japan is weird which even its greatest fans would surely not demur. Some of the weirdness may be good, some not so good. In terms of transport, clearly they are too singular in geography and concentration of population–not to mention the difficulty of car ownership and usage–that it can’t be a useful model for many other countries. Obviously some things can be learnt but rarely can one just assume some practice there will be sensible to transfer elsewhere.
Incidentally, are those Ekiben boxes hot when delivered to the customer? I ask because I found another unexpected weirdness when I worked for a month in Tokyo: at the big hospital I worked at, I stopped trying to eat at the main hospital restaurant because so much of the food was cold. Cooked in the morning and not kept hot or reheated when served. Now don’t tell me that isn’t weird. I’m not a fan of mounds of sticky glutinous Japanese rice at the best of times but cold, glutinous rice ….! I was working like crazy (24/7 as the maxim goes) and eating properly became a bit of a problem esp. as it wasn’t operating at night. Then eating out at a local resto at night with the nightshift guys in my lab also didn’t work out, as I found the food less than palatable –huge mounds of rice which I couldn’t finish (a cultural faux pas)–and the rest being extraordinarily bland, which they ate night after night. On top of the awful midday meal I didn’t last more than 3 nights. Any of the interesting Japanese food we hear about was at very expensive restos, or fresh stuff (fruit) at supermarkets at extraordinary prices. It might also explain the popularity of the usual American chains in Tokyo. Admittedly I had been spoiled at my then French institute which had one of the best CNRS restaurants of wide renown!
There is no evidence to suggest that *in Japan*. We have a sample size of 1, and there are a lot of things about Japanese rail that make it different, and so many confounding factors in Japan vs not Japan that who’s to say the difference in mode share is not despite the lack of cafe car?
A better comparison would probably be to airlines, where food is pure cost (see: that old tale about AA saving $40,000 by trimming an olive from the in-flight salad). Airports are also similar to Japanese train stations in how substantial and varied the food offerings can be. And yet except for very short hops where most of the flight time is takeoff and landing, pretty much no airlines have totally eliminated meals on board at all. Even Ryanair still offers cold sandwiches for sale, and all airlines will try to run a drinks service at least.
Kintetsu operates an Osaka-Nagoya service which competes solely on price/comfort, since it’s over an hour longer than the Nozomi.
The latest trains have dual snack vending areas at each end, plus a bench chillout area.
https://www.kintetsu.co.jp/foreign/english/hinotori/
You could extend the argument to other specialised cars, like the head one in the Swiss IC, which contains a playground on the first floor, and a buggy / bike garage on the ground level. The seating capacity is probably a third of that of a normal carriage. I would argue that they offer a strong advantage for the trains because they all offer facilities unavailable in car. If you have young kids, travelling in a vehicle with a slide and other kids is a strong argument in favour of the train. I suspect a similar argument can be made about the buffet wagon.
Good point about families. In a road trip, you an stop anywhere for snacks. On a train, its important to have snacks on demand. The cafe car also offers an excursion, and a good way to distract the kids.
Trains have a lot of strengths. Lets not erase them all and turn them into greyhound on rails. (Not that theres anything wrong with greyhound! But that should be a separate segment)
The old Super View Odoriko had a kids play area into 2020, although it’s absent from the new version, which seems more aimed at older adults and retirees. (Notably, the steps up to the high-deck “observation” end cars are gone, affording a worse view but an easier ride for those with impaired mobility.)
Kintetsu’s Shimakaze (introduced 2015) has a full cafe car, and keeps the high-level observation deck: https://www.kintetsu.co.jp/senden/shimakaze/menu/
I think the best policy is to configure the train service to your target audience, their expectations, and the train route and journey length- rather than a one size fits all policy. If your service is between two major urban centers under 3hrs. with a predominantly business clientele- you prioritize what they want, which may just be a small cafe space (modular) or trolley refreshment service (coffee, bagels, nuts, etc.), and devote most space to revenue seating. On longer routes with families, casual travelers and such in the mix, it warrants more services which may be a cafe space taking half a car length (12~15m) with the remainder being revenue seating. If we are talking about the more regional Amtrak routes, a semi-decent on-board concessions stand may be necessary given the dismal choices (or lack thereof) of dining/stores in the environs of a present or future Amtrak/Amshack station, located in a blighted section of town.
Kids are also a group where otherwise the train would be worse than driving.
Caltrain feels it is worthwhile to have fewer seats in order to accommodate bicycles. It has “bicycle cars” and has a policy of “bicyclists board first” (because Caltrain won’t allow bikes in the rest of their trains, so sometimes a bicyclist get’s “bumped” because the racks are full). Now, think of crowding at commute time …
Caltrain also takes 4 minutes to unload all these bikes at rush hour at Palo Alto because what’s level boarding.
(“What’s level boarding” is also Europe’s motto when it comes to intercity rail.)
If only they had thought of bicycles and wheelchairs when a lot of that infrastructure was built in the 19th century (probably even CalTrain/SP?). Notably, stuff built as new in the late 20th is level-boarding (eg. RER, BART). Unlike Shinkansen, TGV was designed to use the same gauge as the entire network, and indeed to be able to extend beyond the LGV deep into the regular network, and would have to use the same track and existing platform infrastructure.
Palo Alto is probably the busiest station other than the termini, and one wonders if the long boarding time is really for bicycles as cyclists tend to be pretty nifty in most situations? More likely it is the regular pax and perhaps wheelchairs etc. Or perhaps the train stewards are fastidious about ensuring all cycles are properly stowed (and unaccommodated cyclists ejected) before the train leaves? Does CalTrain run at more than 15tph?
@ michaelrjames,
“Does CalTrain run at more than 15tph?”
More like 6tph peak at the express stations pre-Covid.
The ADA was more than 30 years ago now. Every building needs a significant remodel at less than every 30 years to replace worn materials and equipment. All stations have thus been updated to level boarding by competent management and this is a non issue in the US. If that last if false someone screwed up badly and we need to go to the “need to remove incompetent management” article.
Caltrain has rebuilt almost every station in the system in the last 20 years, with plans for electrification and HSR integration pending funding. No level boarding to any platform height. Richard Mlynarik thinks they should all be fired out of a rocket into the sun for a reason.
@Alon Levy
Isn’t the main impediment that it is difficult if not impossible to upgrade one station at a time? That is, it is all-or-nothing. Plus the awkwardness, or impossibility, of mixing old with new trains on the same line. I imagine you can give examples of the contrary but the ones I can think of–such as the retrofitting of part of Sydney’s North-West-Metro to take the new trains–involved closing the whole line for at least one year while the old platforms were modified. They substituted a bus service.
Nope! The stations don’t even have level boarding to low-floor trains like the existing fleet. American low platforms aren’t 550 mm, they’re 8″ = 203 mm.
Moreover, in the Northeast, train floors are consistently ~1,250 mm; there are high platforms with ~level boarding, and low platforms at which trains have trapdoors that reveal stairs, manually operated by conductors. The LIRR and Metro-North converted all platforms (except the Waterbury Branch) to level boarding in a few years in the 1960s to accommodate the M1 trains, which do not have traps, but other Eastern railroads either are dithering (MBTA, SEPTA) or not doing this at all (NJ Transit, Metro-North on the remaining branch).
@Henry Miller: “Every building needs a significant remodel at less than every 30 years to replace worn materials and equipment. ”
Hah, I can’t help note that just yesterday a 36 year old building in Atlantic City got a very significant remodel … and yes, inevitably a metaphor for both America and Trumpism.
To add onto Alon’s point, more recently the LIRR either converted most diesel platforms to high-level or closed them if the patronage was too poor to justify doing even that in the ’90s, since the C1s and C3s also do not have trap doors.
The only remaining low-level platforms on the LIRR are at Belmont Park, and they built a bunch of wooden stairs on two out of four platforms instead of upgrading them because the station is only open during racing days.
Copenhagen S-Tog has bike cars and that seems to be fine.
Is there an ekiben software system I can buy from Japan and boot on my railroad, including integration with local take-out places in the more far-flung towns? If so, that probably makes sense for a new greenfield service.
If not, and if I’m trying to “imitate, not innovate”, I’d be much more inclined to buy whatever cafe car Stadler or Sharyo or whoever is selling.
I have no idea. But I believe there is such a system in China, and it delivers to the platform even with access control.
I don’t understand buying food in stations as an alternative to café cars. It’s already a much bigger market and the default solution for most passengers. Everybody knows that food in trains is often of bad quality with little choice at high price, “sandwich SNCF” in France is a kind of synonym for “expensive and crappy food”, so for sure conscentious passengers will have bought food before taking the train, but you still have a market for hungry and careless travellers.
I have been a long distance traveller for a couple of months in Switzerland, 1h40 door-to-door including 40 min on a IC train and I can attest that café-cars can be very popular with long-distance travellers who see that as substituting useless commuting time with social time with colleagues and “train friends”.
There is a difference between countries with or without compulsory bookings: If no reservation, every customer sitted in café-car leaves an additional available seat elsewhere in the train, so the only lost space is the part dedicated to service. This is not true if the seat has to be reserved anyway
Not everyone knows that?
Like Amtrak dining car food is reasonably good, not spectacular, but good, competently-executed made-to-order meals.
(On the other hand Amtrak long-distance snack counter food is awful plastic crap, though some of the corridor routes manage pretty well)
Also, some things you want hot or with ice, and I don’t know that I would trust a public microwave on a train.
Stale cold coffee is quite unappetizing. So is ice coffee where the ice has all melted.
I think of all the food-related items I’d trust on a train, the microwave would be high on the trust list. It’s no contact with the food itself and the contact with the container is no worse than the table you’re going to put it on.
I can barely trust work colleagues with a microwave, and I’m supposed to be able to trust the general public?
There are all sort of microwave faux pas, and ways to ruin a microwave/start a fire with one. I’d like to keep that cat in the bag.
Agreed that TGV food is awful but I’ve had an excellent meal on a Thalys (long time ago), some of the food on ICEs is decent (bad coffee but excellent beer) and the Swiss train food is not bad either — bland only because that’s how the locals like it. Surely, quality effects profitability?
Not sure why a restaurant car would be counted like it had no seats at all. In all examples I know the service area occupies only around half the car. The rest has seats or at least standing room.
The cafe car has two functions. The first is obvious: food & drinks, this can be replaced by a trolley service. The second is that it provides passengers with an opportunity to go stretch their legs, walk for a bit, stand around for a bit and then going back to their seats. This is much harder to replace, and is an important health concern during long train rides. Especially for people who are prone to embolisms.
Maybe the cafe car should make way for a gym car instead..
I find it odd that cafe cars are places where people hang out, leaving their ticketed seat empty. Then train policies are difficult for those with pets or bicycles. It seems like mixed use cars is maybe the better idea. There is definitely more money in rolling on muzzled dogs and bicycles than selling coffee or snacks.