Quick Note: Are HSR Transfers Acceptable?

When SNCF built the first TGV line, it did not have funding to complete the full line from Paris to Lyon. Instead, it built two thirds of the line’s length, with the remaining third done on legacy track at reduced speed. The travel time was 4 hours; when the full line was completed a few years later, it was reduced to 2. The one-seat ride remains the TGV’s current operating model, to the point that one unelectrified branch got direct service with a diesel locomotive attached to the trains at the end, and was only electrified recently.

In Japan, transfers are more common, because of the different track gauges. At the outer ends of the Shinkansen, it is common for people to transfer to a legacy express train at the northern end of the line, though on two branches JR East built two Mini-Shinkansen lines, regauging or dual-gauging legacy track to make TGV-style through-running possible. In Germany, the entire system is built on transfers, typically timed between two high-speed trains.

I mention this because the California HSR activists are talking about the possibility of transfers as an initial phase. Some politicians occasionally hint about forced transfers at San Jose, even though it is relatively easy (in fact, planned) to electrify Caltrain and run trains through to San Francisco, but more intriguing is Clem Tillier and Richard Mlynarik’s proposal about running to Livermore first:

This is predicated on prioritizing the San Francisco to Los Angeles connection. It has nothing to do with Sacramento or the East Bay… those are just the cherry on top. Focus on the cake, not the cherry.

LA – Livermore HSR 2:06
Transfer in Livermore 0:10
Livermore – SF Embarcadero BART 0:57
TOTAL SF-LA via Altamont/Livermore BART 3:13

LA – Gilroy HSR 1:57
Transfer in Gilroy 0:10
Gilroy – SF 4th & King by Caltrain 2:00
TOTAL SF-LA via Pacheco/Gilroy Caltrain 4:07

It’s simply not a contest. Even for San Jose, LA – SJ downtown times would be approximately equivalent via Livermore BART once BART to SJ is built. So let me reiterate: No other alternative, least of all Pacheco, provides such a “Phase Zero” access to SF.

The one possible problem: Livermore’s quality of service will be low after BART goes there. From their 1982 opening until 1985, the Tohoku and Joetsu Shinkansen only served Omiya, located 30 km north of central Tokyo; however, Omiya was already connected to Tokyo by multiple high-capacity rapid transit lines, and an additional line was built at the same time as mitigation for the line’s construction impacts.


  1. political_i

    To build BART to Livermore would cost the same for the HSR overlay to San Jose, you might as well build the HSR to San Jose and force the Caltrain transfer at that rate.

  2. Miggle

    The airline industry is probably the most comparable mode to intercity high speed rail, and transfers are ubiquitous there.

    With regard to the California system, I’m sure there have been studies to model how passengers would arrive at the stations. If a significant proportion of travellers are going to be transferring from BART or Caltrain already, then a forced transfer in the inital phase should be tolerable.

  3. Miggle

    Whether the BART Livermore extension provides a transfer to ACE, and hence to San Jose for HSR passengers, would be another important consideration.

  4. Miles Bader

    Shinkansen lines terminate in major hub stations that allow transfer to a vast number of different lines serving a huge number of destinations. This is beneficial for passengers. Such transfer points occur at a place where passengers naturally disperse, and so essentially mirror travel patterns. They make sense.

    That is very different from terminating in a station out in the boonies, and forcing a transfer onto a slow train with a very different character. That really doesn’t help passengers, nor does it represent anything more than “We decided to save money by dumping you here, haha, sorry about that.”

    This is as much a symbolic issue as it is a practical one — if people really do end up taking BART from the boonies into SF (as the “final design” — as a temporary stopgap, it doesn’t matter), it feels cheap, unfinished, bungled. Granted that this is true of much rail travel in the U.S., but CAHSR was supposed to be different — it was supposed to be “the one America finally did right.”

    Oh well.

  5. Alai

    HSR could, I guess, get built without upgrading Caltrain, but doing so would have enormous benefits for the bay area. If you add that it’s no more expensive than building BART to Livermore, well, it’s no contest.

  6. jim

    For transfers not to reduce ridership, you need either that you can achieve DB levels of schedule discipline and that your potential riders believe that you can achieve DB levels of schedule discipline or that you run such short headways that it doesn’t matter if a rider arrives at the connection point late; there’ll be another train along in a few minutes.

    So LA to SF via Livermore and BART might work, assuming BART runs sufficiently short headways, but SF to LA won’t. What happens if there are BART delays?

  7. BBnet3000

    “Even for San Jose, LA – SJ downtown times would be approximately equivalent via Livermore BART once BART to SJ is built.”

    Unless BART changes its service pattern, and i doubt they would, this would require another transfer, and one that backtracks no less. Meanwhile, most people presently take the bus from Emeryville to SF when taking Amtrak rather than transfer to BART at Richmond (which is a good end-of the line transfer for BART).

  8. Alon Levy

    @political_i: just out of laziness (and lack of time to check the EIR), do you know what the travel time would be for Altamont through San Jose?

    @Jim, Miggle: each BART line runs on a 15-minute takt, except in the evenings and on weekends, when it switches to a 20-minute takt. In contrast, ACE runs three roundtrips per day.

  9. David Alexander

    At 4 tph, Livermore isn’t that bad, but I’m far more concerned about the out of the way location when compared to the mode that we’re competing against. In other words, we’re looking at an HSR station in what’s admittedly an exurban, arid area of the Bay Area, but the airports are short drive or trip away from the core. While the time spent traveling may be similar, there’s just that psychological barrier of having to sit on a BART train out to the middle of nowhere, and the use of an alignment which makes running future service somewhat more complicated due to the need to bifurcate service to San Jose and San Francisco. I’d much rather see the line run via Pacheco, and have some upgraded & electrified Caltrain service…

    • Alon Levy

      The split problem with Altamont can be mitigated in several ways. San Jose, a giant suburb (it has more employed residents than jobs, though by a tiny margin) that’s unlikely to have high ridership, could be cut from the line or demoted to a shuttle. It’s this possibility, or more precisely the psychological perception that San Jose is on a branch, that led San Jose to lobby for Pacheco. Another solution, favored by the Bay Area technicals, is to run trains with timed HSR-HSR transfers, using the fact hat Altamont offers good high-speed commuter service between the Sacramento Valley and the Bay Area. Drunk Engineer once proposed LA-SF and Sac-SJ services with a timed transfer, but complete imitation of the ICE pattern would have services alternate, so that you could get your one-seat ride at a lower frequency.

  10. EngineerScotty

    Since someone mentioned the airlines, a few thoughts:

    * Baggage handling is generally eliminated at transfer points–airport staff (usually) ensures routing of checked baggage. I’m assuming that at least some fraction of HSR users will be travelling with nontrivial amounts of baggage–how this is handled could make or break things. In particular, forced transfers to inter-urban services such as BART, which don’t make traveling with baggage convenient, would be a problem. (By the same token, many air travelers will drive to the airport and rent cars at the destination rather than use public transportation for the same reason–taking a suitcase on the subway is a pain). If CAHSR is only going to make it as far as Livermore, they will probably need a big parking lot and a rental car agency or three at Livermore to handle those passengers for whom a connecting ride on BART is impractical.

    * Airlines transfer windows often are measured in hours (or significant fractions thereof)–while this is good if the first leg of the flight is late (it give margin of error), nobody likes sitting around an airport in a strange city. A key selling point of HSR is that travelers’ time won’t be wasted in this fashion–that transfers can occur in a matter of minutes rather than hours. To the extent that HSR starts to accrue the disadvantages of air travel, it will become less attractive in comparison.

    • Miles Bader

      I do not know about European HSR services, but the Shinkansen does not have checked baggage. From what I can see, most people use local transit (maybe with taxi as the final leg) to get from their house / hotel to Shinkansen stations.

      [I feel compelled to note that this isn’t a counterexample against my earlier argument against using BART from Livermore as a permanent solution — Shinkansen terminals are far better located, and Japanese local transit is infinitely better than what’s in the bay area.]

      This is obviously to some degree a matter of what people are used too — Japanese have done this for a long time (and there’s a wider availability of alternative methods for shipping luggage, e.g., one often sees people shipping golf clubs from the local convenience store!), so it’s not a big deal. Maybe Americans expect to have more luggage. But I will note that there seems to be a trend these days towards lighter luggage in the U.S. even on airplanes — many of my friends now try as hard as they can to fit everything into their carry-on…

      • EngineerScotty

        The fact that US airlines are now charging to check even a single bag, whereas in the past each passenger got 1-2 checked bags gratis, may have something to do with that. (And even though it can be annoying, unbundling of services is probably a good idea).

      • MobilMan

        Re: Baggage check-in for HSR. Simple standard: don’t do it. Another unnecessary layer of complexity. Too few people would use it.

        It doesn’t make sense except maybe for for those extended airline check-ins. Check your bags at the local central-station and ride to the airport like in Hongkong, which is another matter entirely.

        Transfers have to be very smooth: same platform, timed transfer, guaranteed level of service, different organizations that make nice with each other. Otherwise those ten minutes for transfer are a very sportive assumption. As far as I know only Switzerland has a nationwide ITT system. Germany’s connections can be very hit-and-miss. Up to 90 minute differences on the same relation because of the unharmonized timetable, local trains etc.

  11. JJJ

    Nobody likes to transfer. It’s not just about the 10-20 minutes wasted, it’s about the complete turmoil. Taking a nap? WAKE UP AND WALK. Watching a movie? PACK IT UP AND GET OUT. Have 3 kids and all their coloring materials + snacks + clothes all out? HURRY UP AND PUT IT ALL BACK IN YOUR BAG.

    Not fun. Makes the experience unpleasant.

    One seat ride = best choice.

    Just look at airlines. You pay a premium for direct flights. Want the cheapest flight between LA and NYC? Then enjoy Vegas and Charlotte on your way there.

    Or ask customers that currently use the San Joaquin + bus combo. The buses are comfortable, and actually quite fast. People dislike having to get everything together and move it all 50 feet.

    • JJJ

      Adding to this, a quite popular practice in europe which is very rare in the US is having trains decouple en-route, again, to avoid the transfers. Maybe a train has 10 cars. All ten will do stops A-B-C-D. At D, the first 6 cars disconnect go to E-F-G and the back 4 go to W-X-Y. No getting off the train (assuming you’re in the right section)

      • upnorth71

        Also present in Japan, with the aforementioned mini-Shinkansen which are coupled/decoupled from fullsize Shinkansen trainsets at junction stations, as well as Narita Airport Express trains on the narrow gauge network. It requires an adherence to schedule discipline that is impossible with current North American railroad operating conditions, though it is implied in any future HSR system that will be built.

      • EngineerScotty

        The Amtrak Empire Builder does this heading westbound, splitting in Spokane with one half the train heading to Seattle and the other to Portland.

      • Miles Bader

        I guess it can work, but it does add to the uncertainty, especially if the train has non-reserved seats (which I think is a good idea), and runs at high/over-capacity.

        I’ve personally almost been screwed multiple times by being in the wrong car of a split train (in all cases, I managed to realize what was going on and get to the right part of the train in time, tho), and being on a “splitting” train always makes me nervous, even if I’m pretty sure I’m in the right car…!

        • Alon Levy

          They can have station staff directing people to the correct car, as they do on iDTGV trains (station staff check your ticket as you board, replacing conductors, and make sure you board the car where your seat is assigned). It’s labor-intensive, though.

          How does JR East make sure people who reserve Komachi or Hayate trains don’t board the wrong train? Presumably for those trains it’s possible to paint the exteriors in different liveries; the only disadvantage of doing so is loss of operational flexibility, but in the specific case of the Mini-Shinkansen the rolling stock is already different.

      • Miles Bader

        Er, well your reservation tells you where and when to be waiting (and the car number of course), and the times are very reliable. People already take care to be in the right location on the platform.

        [I didn’t realize it before, but it seems the Tohoku shinkansen is all-reserved, so the additional burden for free-ticket holders isn’t relevant… :]

        If someone really makes a mistake, they do check tickets on the train …

      • Miles Bader

        … of course, accurately locating your seat isn’t always sufficient — sometimes on an uncrowded train, I’ll move to a different seat in a different car if I don’t like my proper seat (e.g., on a MAX train, the bottom-tier seats don’t have such a great view, due to low walls along much of the track!). Now that I think of it, I’ve never gotten quizzed by a conductor after doing that, even though they seem to be quite careful about checking tickets…

      • Newt

        Amtrak does this decouple on the Lake Shore Limited route and probably others. It isn’t always smooth for the usual Amtrak reasons but it could be done well on dedicated passenger rail.

  12. Oaktown to SF across the bay every day

    I tried to leave a comment on CAHSRBLOG.COM, but no luck.

    I think that the idea of a STARTUP PHASE going to Livermore is AMAZINGLY GOOD. Wow!

    It’s the most interesting thing I’ve ever read about HSR in California, because maybe something that people can USE can happen SOONER. Once people see it works, THEN they want more tracks and more trains, and pay money and tax for it.

    Also, for a person in Oakland, Berkeley, Walnut Creek, almost anywhere BART goes, and also near BART in San Francisco, it will still best choice AFTER HS trains go to San Jose or San Francisco. An easy trip, on a BART train with no passengers (Dublin trains are always empty except maybe two hours going TO SF in the morning and FROM SF evening), no need to commute to SF and then walk to Transbay Terminal … NICE! The Caltrain line and Transbay Terminal and San Jose is nowhere near the biggest section of the Bay area.

    I don’t also how people can think a pure HSR system can be built and start up all at once. 350km/h everywhere … Perfect HSR … cost is no object … who cares about local trains … who care about local towns? Nobody does that. So what pieces can be cut off and done SOONEST and with BEST COST to show that this isn’t all a big money laundering with no results. (Newspaper article today: $660 million spent by CHSRA until today. Nothing built. NOTHING!)

    Example from Spain:

    Wikipedia “Línea de alta velocidad Madrid-Zaragoza-Barcelona-Frontera Francesa”
    1. 2003: Madrid-Lérida max 200 km/h
    2. 2006: Madrid-Lérida max 250 km/h then 280 km/h
    3. 2007: Madrid-Camp de Tarragona max 300 km/h
    4. 2008: Madrid-Barcelona Sants max 300 km/h
    5. 2009: Madrid-Barcelona Sants max 350 km/h
    6: 2016: Madrid-Barcelona Barcelona-Sagrera

    Yes, some many delays and it cost much more than promises, but it started working years ago, and now it works, and it gets more trains and eventually they go on to France too. Not all at once!

    CAHSRBLOG.COM people think maybe in Spain they should wait until 2016 for perfect service to Barcelona, and just sit and look at the tracks since 2003 with no trains.

    • Alon Levy

      There are two issues in California that are not present in Spain and Japan, which is why I’m not completely sure that Livermore would work.

      1. In both Spain and Japan, it was certain the line would fully open. Spain had enough money to build to Barcelona, but just opened a partial segment to Lerida when it could. In California, the currently available money is not enough to get to both Los Angeles and San Francisco; best that can be done with the stated sources of funding, including from foreign governments, is to get from LA to somewhere in the Central Valley.

      2. Livermore is not Sants or Ueno. The closest equivalent in the Bay Area to Sants and Ueno is 4th and King, or maybe Downtown Oakland.

      • Oaktown to SF across the bay every day

        I don’t understand.

        So, if if only gets to Livermore, and they run out of money, then … SO WHAT?

        SO WHAT?

        Is that so very bad? Maybe 20 minutes more time to very downtown of San Francisco. OK. Not perfect, but OK. Faster to all of East Bay than the final result anyhow. More than OK! Maybe 30 minutes more time to downtown of San Jose … but same speed as final result to non-downtown of San Jose, to Milpitas, to Cupertino, to Mountain View, to Palo Alto, to Santa Clara. That’s OK!

        So we should build nothing forever because maybe the perfect future doesn’t come and maybe people think that is good enough? STUPID!

        Maybe “good enough” is OK.

        Do something OK, do it NOW. Otherwise … nothing.

        Question (serious!): is the goal to make only something perfect (meaning kill anything that isn’t perfect) or is the goal to make something people who pay taxes will use?

        Yeah. High Speed Trains to Transbay Terminal (I go there most days … on a bus!) AND to some place in San Jose AND to Livermore is perfect. But High Speed Trains to Livermore only sounds great for now. I like something more than I like nothing.

        • Alon Levy

          That’s not really what I’m saying… what I’m saying is that Phase 0, defined around what there’s money for, should serve a downtown area. If it’s impossible to get from LA to SF directly (and yes, it’s fine to get to a near-downtown station such as 4th and King, on the model of Sants and Ueno), then it’s better to serve just one. So the best Phase 0 should be from LA to as far north as they can get on the money then can scrounge together.

          At this stage it has to be at least (point with good connection to LA)-Fresno because of the existing work on Fresno-Bakersfield, but if Japan’s offer of 50% of the money materializes, then it could well be LA-Livermore. That would be fine; the main anchor of HSR would be directly served, and there would be good service to many intermediate points as well as a decent connection to San Francisco.

          If there’s no money for getting all the way to LA, they could cut it at Santa Clarita or Palmdale and run through on electrified legacy track at low speed; Palmdale is still the preferred option and could well be the only reasonable option, but Santa Clarita’s legacy route to LA is much shorter and straighter. That would require at least some FRA dealings, but it’s doable at least in principle – whether the California HSR Authority is competent enough to push through the changes it wants is another question. At any rate, such partial opening would be analogous to running through to SF from SJ or Newark at lower speed on electrified legacy tracks, rather than to ending at Livermore.

  13. Paulus Magnus

    Politically I think it would be disastrous to do so unless full funding had already been arranged and work proceeding along the relevant lines. I think you’d still have some nasty politicking even with it, but that would alleviate the worst of it and allow the HSRA to ignore the rest.

  14. Mike Dahmus

    “The airline industry is probably the most comparable mode to intercity high speed rail, and transfers are ubiquitous there.”

    If you’re trying to win people away from the airline industry, repeating one of their worst practices is not the way to do it. Most people envision rail as being worth the trouble because you can go downtown to downtown WITHOUT having to transfer.

    • Oaktown to SF across the bay every day

      Huh? I make a ubiquitous transfer every time I go to Oakland airport or San Francisco airport (drive, bus, BART plus bus .. plane doesn’t come to my apartment!). I ubiquitous transfer twice every time I go to work and come home. No problem.

      In Asia and Europe, transfers are the BEST part about the train. Not just one line A-B like an airline, but A-B-B’-B”-C-C’-D-E-… transfers to other places everywhere!

      What’s wrong with airlines is you are booked on ONE flight. Also, there isn’t another flight in 30min, 60min, 90min, … On a train, get to a hub station, then if you miss transfer you wait 10min-30min for the next train and you go. (OK. Maybe not so great in some parts like Greece.) Maybe you stand because train is full one time in 10 or something, but no “rebook” d “cancellation fee”, just get on the train.

      If you don’t want trains to be as bad as airlines, don’t operate them line airlines. That means flexible ticketing (or not flexible ticketing), easy timetable (train to LA every hour, not 11:07 13:40 17:06 17:20 17:54), no messing around with XRAY shoes, and GOOD TRANSFERS.

      If I ride LA to Livermore on HS train and it is late (how I don’t know; no other trains, brand new tracks and trains, no trucks crashing trains, but maybe in California they find a way), then in 14 minutes there is another BART. No problem. BART will be most liely the most reliable part of the trip!

      If I ride LA to San Francisco instead, then I need to transfer to BART ANYWAY.

      “transfers are ubiquitous” ALWAYS. Transfers train-to-train are easy to make easy if architect isn’t done because a train station is maybe 100m across, while airport can be miles from one gate to another. I say GOOD transfers is one of the best parts of train, and a huge advantage over plane. I say no reason why HS-BART transfer in Livermore not to be GOOD transfers. Really hard to do bad with just 2 train lines, one which works good already.

      So I don’t get it. Maybe 100 people arriving in San Francisco from direct train to downtown SF have no transfer and a really short walk. (Maybe 3 people arriving in San Jose.) The other 600 people on the train make TRANSFERS, probably 200 to BART anyway, 100 to taxi. No problem. Same thing in Livermore: not perfect, slower, but not infinitely slower, so OK, and very OK for right now.

      Voltaire was a smart guy from the 1700’s: “Le mieux est l’ennemi du bien”.

  15. Oaktown to SF across the bay every day

    So. No more arguments from me. Thank you to Mr. Levy for the nice forum.

    I was very excited about this clever idea on CAHSRBLOG.COM (well from “Clem” in comments, not from crazy guys who own that), and thought wow this was the best thing ever. Eureka!

    But maybe if you only ride airlines, if you always expect trains to break down, if you think government always wastes huge money — so much you expect only for big big stupid stuff always wasting money and never want small OK but not perfect things — then you think that starting out OK and making it get better later can not be OK. Hard to imagine things different, maybe. Understandable.

    I’m sad only me and two other blog comment guys in California want to see OK HSR in California SOON.

    Meantime, I keep doing transfers every day!

Leave a Reply

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

WordPress.com Logo

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

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.