30 April 2010

Alternatives Analysis Analysis, Part 2

If you haven't run out of steam as you finally reach Appendix K of the recently-released preliminary alternatives analysis, you will come across an intriguing fact: by the year 2025, Caltrain plans for a best trip time of 65 minutes from San Jose to San Francisco 4th & King, and 70 minutes to the Transbay Transit Center.

The Baby Bullet we know and love does San Jose to San Francisco in as little as 57 minutes (by the timetable, which is generously padded at the end of the run to make on-time performance statistics look better). In practice, diesel bullets can and do make the run in 55 minutes.

How can this possibly be? $6 billion later the best commuter service will be 10 minutes slower?!?

How can a timetable, the very root of years of painstaking planning and carefully apportioned capital investment, the embodiment of a grand vision of improved commuter service, the next coming of the Baby Bullet, fall so terribly short of what could have been?

How many microseconds will it take for the entire NIMBY universe to latch on to this inconvenient truth?

Some observations:
  • The proposed timetable, towards which every Caltrain capital investment ought to be targeted in tangible and measurable ways, is presented almost apologetically, and seems almost an afterthought;

  • Caltrain plans to give up (that's right, give up!) commuter trains overtaking each other, despite the presence of a third and fourth track in what is nominally a "shared corridor", and henceforth operate on just two tracks. The reason offered is that overtakes are operationally unreliable--but that is only so because Caltrain's existing passing tracks are very short, and Caltrain's existing gallery cars turn station dwell times into a game of Russian roulette. By 2025 both of those reasons will have vanished.

  • Confining Caltrain to a separate, 2-track system does not take advantage of any synergy with HSR, and perpetuates a false choice between speed and frequency. Today this trade-off is biased in favor of speed, but the Baby Bullet limits throughput to 5 trains per hour. In Caltrain's proposed 2025 timetable, based on the BART model, the trade-off is biased in favor of throughput (one train every six minutes!) but average speed suffers just like BART's. In a fully HSR-compatible system with a mid-line overtake, this constrained trade-off would be removed entirely, freeing Caltrain to provide both speedy AND frequent service with cross-platform transfers.

  • The underlying assumption appears to be that with HSR in the mix, Caltrain will no longer need to provide express commuter service. Even as Caltrain is starved of operating funds and teeters on the brink of bankruptcy, it appears perfectly willing to cede its highest-yielding ridership to the HSR operator, with no guarantee that service will mesh together with convenient timed transfers and a common fare structure.

  • Some might say, if there will be a high-speed train every five minutes in each direction, how could there possibly be any spare track capacity to allow Caltrain to overtake on the high-speed tracks? This question is ill-posed: a portion of that ridiculously frequent HSR service would no doubt be used to serve intra-peninsula demand that Caltrain could not tap into if it were stuck on two tracks with slower trip times. The track capacity should be re-allocated to better meet demand.

  • The fundamental issue here is one of flexible and dynamic allocation of a scarce and valuable resource: track capacity. If you build two separate systems that are not designed to inter-operate (freely sharing tracks and stations), then you cannot adjust service patterns to meet the demand that actually develops. You're stuck with guessing what demand will be 25 years from now, and quite literally casting it in concrete. The outcome, more likely than not, is sub-optimal, with demand under-served and resources under-used.
Caltrain's entire strategic service plan ought to be re-oriented toward the following specific interoperability goals:
  • The same platform height as HSR, so station platforms can be allocated to actual demand, making optimal use of a scarce resource;
  • The same train control system as HSR, so commuter trains and high-speed rail can freely mix and track capacity can be dynamically allocated to actual demand, again making optimal use of a scarce resource;
  • An improved timetable that provides both speed AND frequency through the implementation of a mid-line cross-platform overtake.
With those features, a good case could finally be made that the high-speed rail project would produce tangible benefits that trickle down in measurable ways to each and every community along the peninsula rail corridor. Sadly, right now it looks like those $6 billion will buy us 10-minute slower commuter service. This is simply unacceptable.


  1. Indeed, what "is all this talk of going out of business a first step to a California High-Speed Rail Authority-Caltrain merging into one operating entity?"

    That's the first I've heard of it. I'm tempted to say that Martin Engel is smokin the ganja, unless someone else has heard something similar...?

  2. Maybe they figure that for SF-SJ travel people would take HSR, using Caltrain only for shorter-hop trips.

  3. Alon -- that seems obvious to me, the HSR tracks should have plenty of spare capacity to run SF-SJ express commuter trains.

    But maybe I'm just missing something obvious to the trainspotting crowd.

  4. The great thing about mainline rail is exactly its flexibility in terms of interoperation. You can have local and express commuter trains, intercity trains, and even freight trains all sharing the same tracks. But there's always a tradeoff between flexibility and capacity: the more different types of traffic you have, the less capacity you have, and the more likely you're going to have to resort to things like using sidings for overtakes. But I think the Caltrain corridor is also quite far from this point.

    I also think that giving up express service is a mistake on Caltrain's part. This includes not just the SJ-SF market but also Mountain View and Palo Alto to SF and Millbrae. Why not follow the model of the East Coast Mainline, where 125 mph intercity service shares the express tracks with 100 mph commuter service just fine, with some 16 tph in the peak direction using just the express tracks, in addition to the local and limited-stop trains on the local tracks.

  5. Some might say, if there will be a high-speed train every five minutes in each direction,

    There might be some day far far off in the future 12 high speed trains an hour passing through Fresno. There aren't going to be 12 long distance trains passing through Redwood City...ever... Not even if automobiles and airplanes are banned.

  6. Alon wrote: "Maybe they figure that for SF-SJ travel people would take HSR, using Caltrain only for shorter-hop trips."

    Except that SF-SJ is a minor part of Caltrain's ridership.

    For that matter, San Jose is a small part of Caltrain's ridership.

    People have really strange ideas about San Jose and its importance. Really, really, really strange ideas.

    The facts are that Caltrain could terminate all its service at Sunnyvale and retain 86% of its passengers!

    San Francisco to San Jose express service serves very few people in reality. If push came to shove, it could be pushed overboard and it wouldn't cause a splash.

    Even if that one and only one origin-destination trip were faster, it would still be a small deal. There is and would remain far more bang for the buck in speeding up the trips that more people people do take.

    Downtown San Jose simply isn't a major employment center compared to points further north in Silicon Valley, and not even three more decades of black-hole Redevelopment Agency spending, a baseball team, a football team, a BART line, and a unicorn breeding park is going to change that.

    Faster non-stop trains SF-SJ isn't going to change that.

    Caltrain's core ridership is southern- and mid-peninsula to the SF CBD (San Francisco actually has a developed, non-ghost-town Central Business District) and SF and northern-peninsula to office parks in the northern two-thirds of "silicon valley".

    (Lots of the limit on this last market is down simple geography: Caltrain is competitive where destinations are close to stations, and not sprawled out tens of miles on either side. That's the case where development is bounded by the Bay and the Coastal Range.

    Further south, eg for office park sprawl stretching along 85 through northern annex San Jose former wetlands to Milpitas, or for office park sprawl stretching west to Cupertino and Los Gatos, the connecting (shuttle)bus transit time from any Caltrain station makes a walk/bike/drive+train+long bus trip completely uncompetitive with either drive alone, carpool, or dedicated "Google Bus" type employer door to door service.)

    This isn't slagging San Jose. It's just geographical and demographic reality. There's no there there! And the there there is isn't realistic to get to from the train line.

    SJ-SF expresses will run empty, while SF-intermediate points trains will be required to run slower.

    Screw you, High Speed Rail, screw you, Caltrain, and the horse you rode in on. I'm your customer (remember them?) and all you make me want is see you die.

    I ride Caltrain nearly every day to Mountain View for work, but when I have to further south for meetings the trains are nearly completely empty. The jobs aren't in San Jose, and those that are aren't in downtown San Jose and aren't anywhere near the stations. That's not going to change!!!!!! and it hasn't changed in the 17 years I've been employed at different companies in different cities in Santa Clara County.

    What the @#((*!@#!^@%$@@$#$$ is anybody at Caltrain thinking? Is anybody doing any thinking?

    Electrify or die! The sky is falling! Death before Altamont dishonor! Stand and deliver! Your money or your life! Slow trains are all you suckers deserve! Hand over a billion bucks now or we'll kill this puppy!

  7. That's not going to change!!!!!! and it hasn't changed in the 17 years I've been employed at different companies in different cities in Santa Clara County.

    If someone had looked at the ridership in 1960 they would have determined that a bus once and hour would have been adequate. The concrete they pour between now and 2020 is going to be around in 2130. None of us knows what downtown San Jose or downtown Atherton for that matter, is going to look like in 2075. Which is why they should design it so any train can stop at any station. . . which is what I think you are trying to get at...

  8. Anonymous: be careful with your eyeball estimates of ridership. If all baby bullet stations get the same ridership, then the train will be 20% full between Mountain View and San Jose, and fully half the train will get off at Mountain View, which will make the train look pretty empty. That's always true for the last segment of any outward trip (as opposed to one where everyone is going to the last station). But I think you're basically right: San Jose just isn't that huge of a market for Caltrain. It's somewhere after Palo Alto and Mountain View, and perhaps also Menlo Park or Redwood City or something. Nobody needs a nonstop service between SF and San Jose, and no matter which station HSRA choses as their Peninsula stop, the others are going to get screwed over. Besides, there's no guarantee that it won't cost something like $40 one way from SJ to SF. After all, that's the price of an HSR ticket from Providence to Boston (while the commuter rail is a more reasonable $7.75).

  9. Daniel Krause01 May, 2010 13:49

    San Jose is planning to emphasize large amounts of office development over the coming decades, some of it downtown and some at satelite office areas such as the future Alum Rock and Berryessa BART stations. In the future, it is likey that significantly more people will be commuting to San Jose.

  10. Anonymous: Technically, express commuter trains running on the HSR tracks probably would stop at Millbrae and Redwood City/Palo Alto/Mountain View, but I think your point is still valid.

  11. something doesn't smell right here.

    CalTrain has been saying for several years now the the baby bullet saved their gooses; that it is the prize of their line.

    These numbers don't bear any of that out

    I also think one should look at the 2009 EIR for electrification, and see starting on pages 112 and especially table 2.9-3, that cost savings projected by using electrical power, are more than offset by higher maintenance needs. Hardly sounds like a good prescription for getting out of deficit operations, which is what they have been preaching.

  12. Anonymous pointed out very important fact. I would like to suggest some idea.

    Since Caltrain's most competitive market is north santa clara county (Sunnyvale, Mountain View, Palo Alto) to Downtown San Francisco, Caltrain should focus on express service between these point.
    I would suggest to reduce the train between Sunnyvale to San Jose. In the rush hours, 2 Baby Bullet per hours should be enough to serve San Jose.
    Since no local train between this section, Baby Bullet also stops Santa Clara and Lawrence. They can transfer to local train at Sunnyvale or Mountain view.

    Caltrain should concentrate their resource to the section which have potential and more profitable.
    From Sunnyvale, there should be 4 train/h. In addition to Baby Bullet, Local-Express train should be run. This train makes all stop until Redwood City and then express to SF. From Redwood City, another 2 train/h will stop all station to SF.
    With this idea, number of train is reflecting passenger density and maintain certain level of sear occupancy thought out the section. There are siding tracks north of Redwood city. Sunnyvale have space for additional tracks on the north bound platform.

    Add 2~3 stops into Baby Bullet makes SJ-SF travel time to 64 minutes longer but only 20% of the rider are affected. If Baby Bullet runs every 30 minutes, instead of current 17~43 min spacing, convenience will hide travel time increase. At the same time, Santa Clara and Lawrence get the benefit of express train and more frequency.

    Caltrain still have enough room to improvement in their "Product".

  13. The EMUs used on Japanese commuter lines have absurdly low maintenance needs. Their MDBF is in the seven figures, and they're built specifically to be low-maintenance. What is Caltrain doing wrong if it projects maintenance needs to be so high?

  14. "San Jose is planning to emphasize large amounts of office development over the coming decades, some of it downtown and some at satelite office areas such as the future Alum Rock and Berryessa BART stations."

    San Jose is only "planning" such developments for the purpose of inflating ridership estimates sent to FTA for the BART-SJ extension project.

    I can assure you, as someone who works in the area, that the city has absolutely no intention of planning offices there. Indeed, some office parks are being demolished for conversion into low-density SFH.

  15. Dear Alon,

    Caltrain plans to operate an Amtrak type system under FRA operating rules to a Baby Bullet schedule. Forever.

    The particular trains may be superficially a little different, but the important points will remain.

    So why should MTBF change? Nothing else is changing.

    They'll be Unique American Designs. They'll share tracks with freight trains, which will determine every aspect of the tracks on which the operate and the rules to which they operate. They'll require a Unique American Signal System. They'll be parked out of service, often at constrained terminal station revenue platforms, most of the day. They'll be over-crewed by a factor or 2 or 3. They'll be defined to be double deck, because that's what we have today. They'll be maintained by shade tree mechanics, on a repair after in-service failure basis. They'll operate as energy inefficiently as possible. Timekeeping will be assumed to be, and as a result will remain, dismal, because that's what we do today.

    The "European EMU" business is just lipstick on a pig -- more of the same "baby bullet" but with different nose cones on the trains and energy supplied from an overhead wire instead of a diesel tank.

    You have to understand that the Caltrain people have experienced exactly one "success" in their entire sad careers: the baby bullet that made the front cover of the American Public Transit Association magazine.

    So the only thing they can possibly imagine in the the future is another Special Olympics triumph in the form of more Baby Bullet.

    So of course operating costs are going to rise. Of course timekeeping is going to be bad. Of course maintenance costs are going to rise. Of course overtakes can't be undertaken. Of course we need double deck trains. Of course detailed schedule for a 50 year capital plan should be done by one person using an Micro$oft spreadsheet. Of course it's OK to run one hour headways except at commuter train peaks. Of course freight design standards should dictate all infrastructure. Of course we need 12 platform terminals. Of course we don't need to run to downtown San Francisco.

    All this stuff worked for Baby Bullet, and look how much everybody loves that! Don't mess with triumphant, peer-lauded success.

  16. "The facts are that Caltrain could terminate all its service at Sunnyvale and retain 86% of its passengers!"

    Dear anonymous Caltrain sufferer,

    This is unsurprising, but thanks for the excellent, easy to verify and easy to understand single number that shows just how completely wretchedly passenger hostile and insider rewarding the entire Caltrain capital plan is.

    (Well, one needn't look beyond the San Bruno catastrophe to see that, but $400 million for one grade separation is peanuts compared to what is going to go down in negligible-ridership Capital of Silicon Valley.)

    This 86 number should be tattooed onto the foreheads of everybody at Caltrain and the JPB.

    But "facts are stupid things", in the words of that towering Californian intellectual Ronald Reagan. 14 track two level Diridon Pangalactic coming right up! Because, like, TOD, whenever. And satellites. And stuff. This time fur shure.

  17. http://www.caltrain.com/news_2010_04_30_san_bruno_grade_separation.html

    Here we go!

    Is it possible to switch to FFSS based on the phasing plan of San Bruno? Without completely demolishing the newly-built structure?

  18. Is it possible to switch to FFSS based

    Why would they want to? From the drawings in the community meeting presentation it looks like SFFS to me - platforms on the outside. Not ideal but workable.

  19. Well, yes, I don't see why they would want to. But would they be able to do so?

  20. What particularly bothers me is that if electrification is supposed to be the only way to save Caltrain, why aren't they putting all their capital funding in that direction? Even if they're counting on HSRA to step in with more funding, it still makes sense to get the project going as quickly as possible, and finish as much of it as possible with the money that they do have. Even electrification as far as Redwood City could be useful and would allow Caltrain to get by with a much smaller diesel fleet (probably half its current size, or less) if they restructured their service. And a Caltrain that has full service to Redwood City and only commuter hour express service beyond that is still preferable to a totally commuter-only or entirely bankrupt operation, which are the apparent alternatives.

  21. As in, I don't think FFSS would be a good idea, but even dumber would be to have to demolish the entire station to rearrange things for FFSS.

  22. But would they be able to do so?

    Spend enough money they could do anything......

  23. Again, would they be able to do it without demolishing everything they just built?

  24. Again, would they be able to do it without demolishing everything they just built?

    Sometimes it's cheaper to demolish everything and rebuild than it is to try to save things. The fast side could be used unaltered, The slow side would have major changes. Building it SFFS in 2012 and ripping it all out in 2014 would be really really stupid.

  25. I doubt electrification would "save" Caltrain. I think the only thing that would save it is maintaining the proper level of funding for the system. Caltrain should instead be coming out and saying, "If the Bay Area doesn't want to pay for the system, then there will be no system."

    Either we as a society value good public transportation or we don't. It's that simple.

  26. What particularly bothers me is that if electrification is supposed to be the only way to save Caltrain, why aren't they putting all their capital funding in that direction?

    You're not the only one bothered by this.

    Caltrain claims to be on its deathbed, and is plowing a quarter of a billion dollars into a San Bruno grade separation that adds ZERO value for the train riding public. It won't reduce trip times, won't increase train frequencies, won't reduce operating costs, won't increase ridership or revenue. Worse, it will sabotage design options for HSR.

    This almost makes the Oakland Airport Connector look good!

  27. San Bruno grade separation that adds ZERO value for the train riding public. It won't reduce trip times,

    It certainly will reduce trip times... once Caltrain get level boarding.

    won't increase train frequencies,

    Frequencies are a political/economic decision and rarely depend on the infrastructure until the infrastructure is at capacity.

    won't reduce operating costs,

    If they go with electric trains it will. Going with electric trains doesn't depend on grade separations though.

    won't increase ridership or revenue.

    Again that's more a political and economic decision.

    Worse, it will sabotage design options for HSR.

    The drawings in the Community Presentation look like a standard local station on a four track railroad, two side platforms serving the local tracks on the outside. It implies SFFS which a system used all over the world by most operators. Horrible isn't it?

  28. If they go with electric trains it will. Going with electric trains doesn't depend on grade separations though.

    I see we are once again in violent agreement.

    The San Bruno grade separation can only be justified on tenuous grounds of safety, but at the going rate it would make more sense to spend that quarter billion dollars elsewhere (to save many more lives), unless the project had a realistic chance of saving 30 to 40 lives in the foreseeable future. Then it might make fiscal sense.

    I am singling out San Bruno as an example of Caltrain's misguided and haphazard capital investment 'strategery'. A better strategy than whatever they've got would be to establish a clear set of published, quantitative service metrics against which all capital projects would be judged, in order to direct scarce capital dollars where they are most needed.

    I'll have more to say on service metrics in an upcoming post.

  29. Caltrain is adding some signals in between current signals. Theolically, Caltrain can add more trains, since signal spacing are narrower.
    For trainset for such additional trains, Caltrain already have two set of extra trains which are idling both SF and SJ termical.

  30. Hey, just remember that worse and stupider design errors have been made in railways back in the 19th century (Parliament's choice of the initial London-Dover routing comes to mind, as do the 'gauge wars'), and things survived.

    Heck, for an extreme example look at the decision to use private railways with disconnected and competing stations, instead of a national public railway system.

    Not that they shouldn't fix the design problems, but it does look like it will be an improvement over the current state *despite* the design problems.

    "What is Caltrain doing wrong if it projects maintenance needs to be so high?"
    Is that just vehicle maintenance, or does it include track & catenary maintenance?

  31. But we're supposed to learn from past mistakes, not repeat them ad nauseum...

  32. From here: http://www.thetransportpolitic.com/2010/04/29/san-diego-plans-extension-to-its-trolley-network-mostly-skipping-over-inner-city/#comment-42693

    Based on the conversations I’ve had with people at FRA, the agency is more likely to consider such track sharing in the context of PTC than you might think. This administration recognizes the problems with the existing rules.

  33. I agree that Caltrain should adopt an NEC-style line with Sunnyvale taking the role of Jersey Avenue. The market is clearly skewed one way, and there is way too much reverse peak service currently.

    At this rate though, it may be better for CAHSR to take over the entire row/commuter services and abolish Caltrain. Incompatible heights is one of the worst decisions I've ever seen. Then again it's really hard to trust public agencies nowadays. All the competent people are probably in the private sector for good reason.

    In the end though, I just want BART off the Peninsula at all costs.

  34. The market is clearly skewed one way, and there is way too much reverse peak service currently.

    The market is nearly symmetric. The Peninsula is not the NEC: San Francisco is not as core-dominated as Manhattan and does not have as good a connection between the train and the job centers, and Silicon Valley is a far larger job center than Middlesex and Mercer Counties.

  35. I agree with this point: there is way too much reverse peak service currently.

    I suggest Caltrain to re-allocate some of reverse peak service into traditional peak service. Very few reverse peak train can fill the capacity. They can only filled with bicycle space. Lot of bicycle means longer dwell time.
    On the other hand, Major destination of "Traditional peak" is downtown SF, bicycle is not a requirement.

  36. Anon, the reverse peak trains don't require a lot of extra capacity. They free up terminal track space in SF, and Caltrain would have to pay the train operator no matter what. In contrast, peak trains require more equipment, more salaries to pay, and more track capacity on the line and at the terminal.

    The problem with Caltrain's peak service isn't frequency. It's speed, which is inversely related to frequency, and transit connections in SF.

  37. Sorry, I guess clearly skewed was a poor choice of words.

    What I meant is there's too much frequency of reverse peak. I personally believe those are the trains that Caltrain should have cut, not the midday ones. While the former would have maintained reasonable reverse peak service, the latter is particularly detrimental towards Caltrain's reliability.

  38. Caltrain's demand may not be symmetrical, but there's a lot more reverse-peak demand than in most other "commuter" systems. And with the way the normal and reverse peaks line up, with the reverse peak being a bit later, inbound peak trains get turned to become outbound reverse peak trains. The main reverse peak trains don't have any huge effect on reliability, because after they finish their runs they're parked for the afternoon or evening. It's really the regular peak that has the big reliability impact on the reverse peak. And the key point of the "86%" number is that demand is distributed approximately evenly at the south end of the line, with San Jose having a level of demand similar to other stations, rather than being the huge destination that SF is. As for talk of the NEC and Jersey Avenue: there's a simple reason for things being that way. New Brunswick is the end of the contiguous suburbanized zone, and from there it's over 10 miles to the next station at Princeton Junction. It's a logical place to have that break in service. The Peninsula really is somewhat atypical in that there's some 60 miles of unbroken urbanization along the line.

  39. Slower travel times are bad enough, but
    what really gets me in the proposed Caltrain schedule
    is that most trains stop short of the TBT.


    I don't think there's demand for 10 tph if
    the new trains don't stop at the TBT.

  40. "the latter [reverse peak service?] is particularly detrimental towards Caltrain's reliability."


  41. "there's too much frequency of reverse peak."

    Commuter railroading at its finest!

    All you need is a 30 platform station in the CBD to store all the trains that make one trip inbound, pile up there in the morning, sit around doing nothing all day, and then drain back out to the provinces at night. Plus, a big clubhouse (pool tables, bunk beds, cigar rooms, big screen TVs, etc) for the Amtrak guys and girls to while away the midday langeurs. Done and done.

    So simple. So elegant. Why don't the backwards peoples unworthy enough to have been born overseas understand the beauty of the timetables of The Founding Fathers? What's wrong with them?

  42. If anything, having people commute in both directions is advantageous. Like others have said, it prevents the trains from piling up at one end of the line during midday hours, and consequently, needing a place to store them.

  43. And it prevents you from running empty trains...