24 January 2020

Electric Timetable Contest

The coveted Takt Cup
Timetable planning has long been a staple of this blog, with the support of rapid prototyping tools like Richard Mlynarik's excellent Taktulator, a calculator for "Taktverkehr," the German term for clockface timetabling. While it may take a few minutes to learn how to use the tool, you can easily punch in a stopping pattern into the Taktulator to get an instant score, based on well-researched quality metrics and train performance calculations described here almost a decade ago. The service quality score is normalized so that the 2011 timetable, not much different from today's, earns 100 points.

Working back from its long term service vision, Caltrain has started planning for the near term timetable change that will occur with the start of electric service. Through a process of elimination, Caltrain has settled on two candidate service patterns, each with six trains per peak hour per direction, linked below in the Taktulator. You can verify that the resulting string line diagrams match extremely closely with the last couple of slides in Caltrain's presentation.

Two Zone with Express
Score: 123.3
Fleet: 13 EMU + 7 diesel

Distributed Skip Stop
Score: 124.1
Fleet: 13 EMU + 7 diesel
This timetable has a bit of a "can't get there from here" problem.

Can YOU beat those scores with a better concept?

Of course, scores depend on the assumptions you make. If you assume that the downtown extension is built into San Francisco Transbay, that all the diesels are replaced by EMUs, that dwell times are shortened by system-wide level boarding, that operating practices are reformed to allow better punctuality with less padding of the timetable, that terminal turn times are shortened to match foreign practice, and that a cross-platform transfer station is built in Redwood City with a short four-track section from just north of San Carlos into Redwood City (most of these contemplated in Caltrain's long-term planning), then you can set a sky-high score. In fact, using the Taktulator, you can even quantify the service benefit of each separate improvement. If we're allowed to dream, surely this is one of the most efficient:

Richard's Finest
Score: 230.2
Fleet: 16 EMU

Unfortunately, for the start of electric service in 2023, we'll have to settle for a bit less. There is no service to San Francisco Transbay, there is a fleet of 19 EMUs available of which you probably don't want to operate more than 17 at any given time, dwell times are still long (for simplicity, assume 45 seconds everywhere), timetable padding is ample (assume 10%), terminal turns are slow (assume 15 minutes), and there are no expanded stations or passing tracks. So, with those assumptions input into the Taktulator, can you beat Caltrain's score and win the coveted Takt Cup?

Please post your suggested Taktulator timetables and scores (and your supporting rationale) below in the comments. In your comment, use a clickable hyperlink, in the format <a href="your-taktulator-link">your timetable title<\a>, for brevity and clarity.

Here's my first entry for this contest, to kick things off:

Silicon Valley All Stop
Score: 126.9
Fleet: 17 EMU + 4 diesel

This improves on Caltrain's concept by admitting what census data and Caltrain's presentation tells us: all of Silicon Valley has enormous ridership potential, and running skip-stop express service south of Menlo Park is harmful to overall service quality. In short, the Baby Bullet is bad. This timetable also makes better use of the EMU fleet, as was intended when additional trains were ordered, by running 5 EMU + 1 diesel per hour per direction, instead of 4 EMU + 2 diesel.

Can you beat my score subject to the assumptions above?


  1. At first glance, Caltrain is downgrading Menlo Park from 3 tph today to 2 tph with EMU.

    I'm not convinced that Baby Bullet is bad. Some of the NB AM trains peak out in Sunnyvale. According the ridership data: 319 and 329 are already 120% over capacity in Sunnyvale which is the 3rd stop after Tamien and SJ. Next stops are PA and RWC, so there's big demand for service to fill 2 trains without stopping anywhere else.

    Perhaps the number of local services could be increased, but I'd start by splitting up the ridership of 319 and 329 across a 3rd train with similar patterns.

  2. I think that there will be an uproar if they get rid of express trains. Given the amount spent, increasing the best case SJ-SF time by 10 minutes is not going to fly. I like the skip-stop with express idea best. Something like this:

    Skip Stop with Express

    This offers 65-66 minute SJ-SF diesel express service (similar to today) serving all current bullet stations except 22nd/MP. Those two could be included with distinct AM/PM service patterns for express trains like today. A/B skip-stop EMUs take 71-72min SJ-SF

    Score is 124 as it has the connectivity problems common to the distributed skip-stop pattern. I don't know how big a deal that is though. The San Mateo Co. stations are so close together most people can choose one of two nearby stations as an origin depending on their destination.

    Since all express stations are also served by skip-stop trains, this service plan offers an easy transition to 4tph mid-day pattern by just stopping the express and spreading out the A/B skip-stop trains at 15min intervals. Evening/Weekend service would ideally be 3tph to line up with BART schedule.

    At the LPMG meeting, the presenter pointed out that express service resulted in crowded trains as service is differentiated and more popular trains are fuller. However, I think the diesels are going to have a lot more seats than EMUs to start with so maybe having some differentiation is a good idea.

    1. Trying to post this again - iPad had problems with the links before.

      One idea I had was a timed BART-EMU transfer at Millbrae, which requires a 15min gap at Millbrae but means express no longer needs to stop there. This yielded this:

      Millbrae Timed Transfer with Express

      With Clems 10% padding and 45sec dwell assumptions this gets a score of 127.4 assuming 4tph electric / 2ph diesel as Caltrain does in their proposals. (probably a good idea to start with).

      If I use Clems assumption of a 5 electric / 1 diesel I can add a couple of stops to the EMU express and get this:

      Millbrae Timed Transfer with Express EMU

      This gets a score or 128.8 and maintains an express time of 65min SJ-SF with all Millbrae stops timed for transfer to BART at 15min intervals.

    2. A Caltrain official last week said that post-electrification there will only be 4 diesel-hauled round trips under the SF-SJ wire per weekday. These will continue providing one-seat peak-period, peak-direction service to Gilroy using Bombardier cars.

      However, until all currently-funded EMUs are in service, all trains will continue to run on “diesel schedules” so that they can still be run as diesels.

    3. @RealityCheck - It makes sense that they don’t change the commute-time schedule until they have enough EMUs to run a 6tph schedule reliably. However, I would hope that they would start by using the EMUs for off-peak and midday local service and modifying that part of the schedule. Should be able to cut 15-20 mins off the current atrocious 96min SF-SJ time.

    4. Wow, looks like jpk122s has both hands on the Takt Cup. Note the model doesn't award points for timed transfers with other systems, so extra credit for creating the regular 15-minute takt at Millbrae BART!

      I fixed the fleet assignment, and it looks like this can operate with 17 EMUs and 4 diesel consists:

      Millbrae Timed Transfer with Express EMU

      128.8 is the score to beat! Can anyone break 130?

  3. Meanwhile in HSR land:

    1. So exactly what's new here. Republicans still want to kill it and Ralph still wants to sell papers.

    2. Though I do find the DTOIG report interesting in how it admonishes the FRA more than it does the CAHSR.

    3. Complete report:

  4. It is unclear why Santa Clara County (AKA Silicon Valley) would be interested in pissing away another $60M/year on this POS.

  5. Is there any plan to build pocket truck for turn around in middle point? Judging from traffic density, it is better to have turn around pocket track south of Sunnyvale or south Mountain View. With limited resource of train-sets but need more frequent service north of Sunnyvale, 25~40% of train can be terminated @Sunnyvale and immediate trun around to northbound. Making all stop of south of Sunnyvale will provide frequent service (every 15~20min) for Lawrence, Santa Clara, Diridon and Tamien.

    1. Caltrain doesn't do plans or planning! So stop wondering if they do. This is America, and we employ Americans to do things The American Way.

      I messed around with turnback locations south of Redwood City and north of San Jose (Palo Alto, Mountain View, Sunnyvale) years ago, and I found it hard to find a location that makes timetable+infrastructure sense -- let alone the political issue of turning just short of car-centric hicksville San Jose, soi-disant Center of the Universe.

      But for areally easy and very feasible little near-term project that makes tons of timetabling sense, adding a (non-pocket, so not ideal, but much easier to build) siding track just south of Redwood City station, on the west side of the tracks, between Jefferson and Maple Streets, in the "City easement on JPB-owned right of way" that is Pennsylvania Avenue behind the Redwood City library would be a truly useful thing!

      With a facing (SB track to NB) crossover (a cheaper, low-speed one) just south of the existing platforms and north of Jefferson, terminating "San Mateo shuttle" trains can cross over and park without blocking the main lines for more than a couple minutes, and without more than a couple minutes of non-revenue movement time. (Doing so 1200m further south using the existing and useless four-track section at Redwood Junction would take six or seven minutes longer.)

    2. Eastern side of the existing tracks, not "west". Sorry for the typo.

    3. BART's Yellow line do turn around at Concord station in the weekday peak period. They are 100% American. Why Caltrain cannot do?

  6. Breaking News:

  7. OK, there are a number of things to tease out, or, to be cynical, a number of different ways to game the Taktulator score.

    To start with, the "Richard's Finest" scheme above isn't that way because I was gaming the software, but because I was considering all of
    * Actual real-world peninsula demographics and land use;
    * The distances between existing station stops, and the way in which closer stops lower average speed more than wider stops (which breaks down pretty well into "north of Redwood City" versus "south of Redwood City";
    * Copying existing successful practices from elsewhere in the world, which is the only thing any public transportation anywhere in the USA should ever do;
    * Giving a damn about operating cost, measured as a good proxy by the peak number of trains required to operate the service plan;
    * Attempting to do facilitate that whole log-bygone "Caltrain-HSR compatibility" thing, by slotting in (or "blending", as the idiots have it) the inter-regional trains into out Caltrain-y regional service pattern with as little infrastructure and as little disruption as possible.
    * Considering analogous service demand and service levels (hey everybody, BART isn't running 8tph from Fremont or Richmond to SF, even with FOUR ideally-located CBD stations!) and so going for realistic and adequate (4tph corridor-long SF-SJ) service and contained trains-in-service count (ie operating cost)

    So "Richard's Finest" wasn't an attempt to game the Taktulator metric and win the nine-years-in-the-future Takt Cup, but an attempt to give a shit about regional transportation quality and the regional environment, by pretty much copying what other people elsewhere have been doing, increasingly successfully, for decades.

    That all said, it's instructive to examine a number of scenarios and see what their effect on the Taktulator score is. Some of these scenarios may reveal more about defects in the scoring system than anything, certainly. (It's "special pleading" when you do that, but "insight into algorithmic failing" when I do, OK!)

    Starting with "Richard's Finest", which assumes
    * level boarding (what sort of corrupt sleazebags would not have left us with corridor-wide level boarding by 2010, and done so at negative cost, given the reconstruction of nearly every station platform over the period of Caltrain's existence)
    * non-third-world schedule adherence (7% padding) and operating practices (under 15 minutes for a train to reverse direction ... oooooooooh futuristic)
    * non-excessive peak service (4tph to every station, with 6tph north of Redwood City to serve the dual purpose of filling in close-spaced local stops and spreading demand over more trains to the 40000lb gorilla) of Peninsula transit demand, downtown SF)
    * Oakdale Station exists because of real-world job and residential density
    * Hayward Park and Broadway stations do no exist, for the same reasons, along with proximity to Burligame and (new) Hillsdale stations respectively
    * leaving space for 2tph HSR trains
    * trains run to Transbay (again, the 40000lb gorilla)

    We can see what happens to Taktulator scores when we mess with any of these factors... [but it's really hard to do so in one blogspot.com comment, because long URLs count against a 4098 character comment limit ...]

    1. Starting from "Richard's Finest", which I'm going to designate "2+4 +DTX +Oak -HP" (2tph local north of RWC; 4th limited north of RWC local south of RWC; downtown SF extension exists; Oakdale stop exists; Hayward Park stop discontinued), we get what I'm going to regard as a "base" score of 230 while requiring 16 trains in revenue service.

      To see what drives this, note that a realistic, once-near-term, now-post-global-ecological-apocalypse 2tph HSR slots into this really nicely.

      If we want to game the taktulator score, and acknowledge the actual real world reality that high speed rail in California isn't going to happen within many of our lifespans, and certainly not within the lifespan of Caltrain's not-in-service-for-years-to-come electric fleet, we can up the local service to an unnecessarily high 4tph (""+4 +DTX -Oak +HP"), with 4tph local service to Hayward Park (unrealistic, but we're gaming here), and overall goose the score to 252 at the cost of 20 peak trains. WINNER!!!

      [... to be continued, in sub-4096 byte chunks ...]

    2. [... contining, tiresomely ...]

      OK, so trying to tease this out, how much does the Taktulator (incorrectly, hah!) reward 4tph local north of RWC? Looks like ""4+4 +DTX +Oak -HP" about 12 sweet-sweet Takulator Cup points (not that these factors are additive!), score 242 with 20 trains.

      Compared to "base", how does (incorrectly!) making 4tph stop at Hayward Park station within the city of San Mateo, literally within sight of Hillsdale Station and barely out of sight of downtown San Mateo, score? ""2+4 +DTX +Oak +HP" gives score 239 (feels like a Taktulator scoring, but I'm playing along) with the base 16 trains.

      How much would the (unrealistic, for a long long time, given how things work around here) non-existent Oakdale station in SF contribute compared to base? "2+4 +DTX -Oak -HP" scores 221 with 16 trains. The magnitude is surprising, to me at least.

      And the 40000lb gorilla question: what does the quite literally criminal PBQD/WSP+MTC(=Heminger)+SFCTA+PCJPB+TJPA conspiracy to fuck over the SF downtown extension cost the world? "2+4 -DTX +Oak -HP" score plummets to 139, for the same 16 train revenue fleet requirement. "Death is too kind a fate for any of those responsible", as they say. Far too kind a fate...

    3. [... continuing, tiresomely ...]

      But we haven't begun to plumb the depths of agency and consultant outright corruption and incompetence yet. Not begun.

      Because our "base" scenario assumes the ability to reliably make station stops of 20 seconds (you know, just like, oh, I don't know, what would an utterly insular and ignorant local yokel drawing a half-million-dollar public salary understand?; oh, I know!; how about BART!!!) or do something highly technical like shift into reverse in under 15 minutes.

      So we have to re-baseline, for our super super super Special Needs cases. (The ones who've spent a couple billion of your dollars, with the result being the Caltrain you suffer from today, ... and forever. They're truly special, our Caltrain consultants and agency staffers!)

      Let's forget about running trains to downtown SF (not going to happen), let's forget about HSR (not going to happen), let's ignore level boarding (not happening, by explicit design of Caltrain staff who are on the public payroll), let's be all "realistic" and "grown up" like rocketeer Clem would have us all be, and see where the Taktulator takes us.

      Descending through "2+4 +DTX +Oak -HP" (score 230, 16 trains) to "2+4 -DTX +Oak -HP" (score 149, 16 trains) to "2+4 -DTX -Oak -HP" (losing Oakdale) score 129, 16 trains, a new low baseline but still with futuristic level boarding), we wonder how much worse do things get given Caltrain's explicit multi-decade policy decisions of deliberately sabotaging service modernization?

      Throwing a minimum 45 second station dwell time, and upping scheduling padding to 10% and turnback times to 15 minutes (worse than the 19th century, way to go, American Commuter Railroading Expert Professionals!), we end up at "Redwood 45dwell 2+4 -DTX -Oak -HP", plummeting to a score of 113 while increasing fleet requirement to 17 trains.

      But we're not done ...

    4. We're not done, he said... waiting for the next episode

      With our third world assumptions, can you beat 126.9 ?

      (Also re: Oakdale... the surrounding population density exceeds 22nd which is a healthy "bedroom" stop for Silicon Valley workers. However, my model has the feature that it counts all people as equal regardless of socio-economic background or race, so it may see more potential in Oakdale than our unequal world will allow to be realized. Still, Oakdale is an important location that deserves new service!)

  8. As discussed here on this blog long ago, the constant warning time (CWT) grade crossing protection system Caltrain (and, apparently all other RRs) uses is incompatible with 25kV AC electrification. Denver RTD, using the same (Balfour Beatty) electrification contractor, has for years had major ongoing problems with this issue (e.g. regulators threatening to shut down service and the 24/7 use of human "flaggers" at grade crossings to allow operations). And because FRA says RTD's crossings do not provide CWT, RTD is seeking to receive an FRA waiver so as to be able to implement train horn quiet zones.

    With that background, I am disappointed to learn that Caltrain's is now seeking regulatory approvals to also use a non-CWT post-electrification solution for its grade crossings. They're calling it "dual speed check" (2SC):

    Consistent Warning Time (CWT) for Grade Crossings

    The Electrification contractor is moving forward with design using a dual speed-check solution which apparently will satisfy FRA and CPUC requirements.

    The JPB and its contractor met with the Federal Railroad Administration (FRA) and the California Public Utilities Commission (CPUC) on September 19, 2019. The FRA requested a test plan for a complex, multiple crossing installation such as San Jose. FRA still needs to decide if the proposed dual speed check solution (2SC) is "new and novel technology;" a HQ decision is expected soon.

    According to Caltrain's Director of Operations & Maintenance at the most recent Citizens Advisory Committee meeting, it measures train speed and accounts for maximum possible EMU acceleration as follows:

    • measured speed is 40 mph or more, the crossing activates as though the train is going a constant 79 mph
    • measured speed is under 40 mph, the crossing activates as though the train is going a constant 59 mph

    While a simplistic binary scheme like this ensures an accelerating EMUs can't "outrun the gates" ... it also implies:

    • they've given up on a more elegant solution providing warning times based on actual train speed
    • unpredictable, inconsistent warning times
    • needlessly increased warning times (& therefore gate downtimes) for all trains going less than 79 mph
    • trains going 79 mph will result in the shortest warning times
    • the slowest moving trains will have the longest warning times and net gate downtimes
    • speeds over 79 mph will be impossible without first reworking or eliminating all affected grade crossings

    We've probably got some of the best and brightest engineers here, and after years of knowing about this issue (among the top electrification project risks right from the start!), knowing all that Denver went (and is going) through, this ham-fisted solution is what they come up with?

    1. Predictor: Integrated crossing warning system providing constant warning-time feature without requiring inter-connected wiring.
      Siemens predictors are built on an electronic, microprocessor-based modular system designed to reliably detect an approaching train and start the crossing warning system.

    2. From a cursory skimming of the Siemens Grade Crossing Predictor 4000 application guidelines, it appears it can do everything that Caltrain might need to achieve constant (or nearly so) warning time crossing protection and traffic signal preemption for bidirectional running at or near station platforms and everywhere else.

      I can't tell for sure, but do these "off-the-shelf" Siemens Wayguard GCPs work well with 25kV 60 Hz AC electrification? (It would seem so as long as you stay well away from using 60 Hz frequencies and since the photo on the product page shows a train under catenary passing through a protected grade crossing.)

      And If so, why weren't they used by Denver RTD, or by Caltrain?

    3. So... not exactly what it says on the tin, "Constant Warning Time"

      (Note the term "Consistent Warning Time" is an abuse of the acronym... the 'C' definitely stands for "Constant" in normal usage as one can quickly ascertain from the 20x higher number of Google search hits for "Constant Warning Time" relative to "Consistent Warning Time")

    4. Clem,

      Surely "consistent warning time' is more accurate? Looking at the Wikipedia page for "Level Crossing Signals", the SRI design isn't actually _constant_ time. If the train accelerates or brakes after the crossing signal starts, the warning time isn't _constant_. It's extrapolated, from the absolute value and rate-of-change of inductance, and a decision is made:

      "The absolute magnitude of this voltage and its rate of change can then be used to compute the amount of time remaining before the train arrives at the crossing, assuming it is running at a constant speed."

      Note the "assuming".

    5. @Reality Check
      Operating frequencies (156 through 4000 Hz) are available for use on non-electrified track
      Operating frequencies 645 through 4000 Hz should be used on electrified track
      https://assets.new.siemens.com/siemens/assets/api/uuid:f7c524f8-ee2a-4fe9-aa71-0daeb7074600/version:1565620909/sig000019a-1.pdf section 3.2

  9. Invensys Rail has integrated its Coded Audio Track Circuit (PSO 4000) and its GCP 4000 Crossing Controller technologies to provide an Occupancy Detector System that can be configured to provide constant warning time in electrified territory. This means that a coded, audio, overlay track circuit is available for electrified territory in an integrated crossing package—the “Crossing-In-A-Box” for electrified territory. Invensys’ crossing-protection products and systems are equally applicable to transit or heavy rail passenger operation where grade crossings are present.
    Q: why weren't they used by Denver RTD, or by Caltrain?
    A: NIH

    1. How can NIH be the answer, when everything else in the system is NIH ... and after getting an embarrassing and expensive burn on the years-long failed CBOSS misadventure, it just doesn’t compute.

      So the question remains: why didn’t Denver, and now Caltrain, just buy off-the-shelf CTW grade crossing predictors that work with 25kVAC electrification??

    2. Caltrain August 2018 Board Meeting:
      The only known device capable of supporting Constant Warning Time {CWT) in electrified territory does not
      require insulated rail joints:
      "The PSO 4000 couples to the track with o bandpass, low impedance connection-you don't have to have
      insulated rail joints on the track."
      http://download.siemens.com.au/index.php?action-filemanager&doc form name-download&folder id-563
      3&doc id=17039
      Please consider deferring your vote on item #7 (d) CHANGE ORDER FOR INSTALLATION OF INSULATED
      JOINTS until after confirming the requirement for insulated rail joints with VTA signal engineers (the VTA
      purchased a PSO 4000 for $38,688.32 last year).
      https://hsr.ca.gov/docs/brdmeetings/2018/brdmtg_081618_Comments_Received.pdf (page 5)

  10. Compare and contrast: Level Boarding versus Electrification.

    Comparisons made with the same boring timetable: every train running every 15 minutes, making every stop from SF Mission Bay to SJ Tamien (including Hayward Park due to Taktulator scoring counter-factuals, excluding Oakdale, Broadway, Atherton, College Park). Good enough for BART; better than Caltrain!

    Today's scenario: no level boarding, diesel locomotives, appalling 45s/stop dwell, appalling 10% schedule padding, wasteful 15 minute turnbacks, appalling 2-5 staff per train feather-bedding:
    Diesel, 45 dwell, 10% pad16 trains (81.5% utilization), 103 Taktulator points.

    With level boarding, diesels, unexceptional 20s dwell, same 10% pad 15 turnbacks, and now with the possibility of one-person train operation:
    Diesel, 20 dwell, 10% pad14 (84.8%) trains, 112 points.

    With level boarding, diesels, 20s dwell, reasonable 7% schedule padding enabled by level boarding, 15m turnbacks, possibility of one-person train operation:
    Diesel, 20 dwell, 7% pad14 (82.7%) trains, 114 points.

    So having gone to level boarding while keeping diesels would have gone from 103 to 114 Taktulator points, while saving two entire trains and crews (12.5% of fleet = $70 million!!) and enabling One Person Operation (with Fare Inspectors, not self-important radio-blaring 19th century relic "Conductors"). All this for the low low cost of less than zero dollars (as system-wide platform rebuilding over the last two decades would have been a sunk cost, and by "sunk" we're talking the Titanic.)

    [ ... continued]

    1. [... continued]

      Now contrast level boarding alone with the "benefits" of Caltrain-style electrication. (And by "benefits" I mean "higher operating costs" and "huge consultant payouts" and "hundreds of millions more costly than open global market procurement" and "lipstick on a pig".)

      Electrification without level boarding (Caltrain's plan for us, forever and ever), appalling 45s/stop dwell, appalling 10% schedule padding, wasteful 15m turnbacks, appalling conductor feather-bedding:
      Electric, 45 dwell, 10% pad14 (84.3%) trains, 114 points.

      Hey, exactly the same outcome in Taktulator score and fleet size as level boarding with diesels, but a billion and a half more expensive!!!!!

      In other words, the system is working exactly as intended.

      For reference, combining electrification and level boarding -- something nobody over the age of 40 will ever live to see at Caltrain -- results in score 127, fleet 13 (79.7%). All with a "dumb" all-stops adequate-headway nicely-simple timetable.

      The explicit decision of Caltrain's staff and consultants to never consider nor undertake level boarding at any time in the last three wasted decades, nor to realistically plan for level boarding at any time in the next decades, has directly cost the public a billion and a half dollars.

      This makes the CBOSS catastrophe -- exact same cast of criminal staff and consultants responsible, exact same lack of prison time served and contractor bankruptcies resulting -- look like a bargain! Hooray for CBOSS!

    2. Blogger.com software is b0rked, even more hopelessly than usual -- links in last post should have been (still can't get them through):

      Electric, 45 dwell, 10% pad14 (84.3%) trains, 114 points.

      Hey, almost exactly the same outcome in Taktulator score and fleet size as level boarding with diesels, but a billion and a half more expensive!!!!

    3. [Google-owned blogger.com link posting screwup continued]

      For reference, combining electrification and level boarding -- something nobody over the age of 40 will ever live to see at Caltrain -- results in score 127, fleet 13 (79.7%).

  11. For those of you interested in CAHSR from Bakersfield to Palmdale
    a draft Environmental Statement was released by The Authority.



    Ralph V had a column about it. From his summary:


    Bullet train plan for Tehachapi passage would cost $18.1 billion over 82 miles

    FEB. 29, 2020

    .... route that would include 9.3 miles of tunnels and 15.8 miles of elevated structures, representing about 30% of the entire passage between the two cities.

    ... this stretch of the route, as planned, would take out the R. Rex Parris High School in Palmdale, the Lancaster Community Homeless Shelter, the Solid Rock Bible Church, eight motels, 253 residential housing units, 311 businesses and 175 farm fields, according to the environmental documents.

    ... a section of the iconic Pacific Crest Trail, which runs from the Mexican to Canadian borders through mountain ranges, forests and deserts, would have to be relocated.

    ... the cost of the Bakersfield to Palmdale section shot up from $7.7 billion in 2012 to the current $18.1 billion, an increase of 135%.

    The route would include one viaduct extending 4,600 feet, nearly a mile, and another soaring 220 feet, roughly the height of a 20-story building, above the ground, according to the rail authority

    1. I wonder, since Prop 1a will no longer be in play and 18 billion is a lot of money, if it's more prudent to now focus on 1) partnering with Virgin to do the tunneling section from Palmdale to Burbank and hold off on the Bakersfield to Palmdale section for now, and 2) build the line up from Merced to Sacramento before going to Gilroy and the Bay area. This would leverage the existing infrastructure and ridership sooner at a lower future investment cost. I think a return on investments should be the new number one priority; enough of the prop 1a rationalizations.

  12. Why the US Sucks at Building Public Transit (including HSR)

    So, it turns out, the answer is multi-faceted ... but I think this piece does a pretty good job of hitting many — if not all — of the key facets. See if you agree ... because I can't immediately think of what, if anything, it missed touching on.

    1. I think that the abject lying and purposeful ignorance by politicians means government can never earn the trust of the taxpayers. Private industry chases money, and can always be expected to do so, because if it doesn't it, disappears. The government can go bankrupt, but never disappears.
      Willie Brown, former SF mayor, San Francisco Chronicle - July 28, 2013
      "News that the Transbay Terminal is something like $300 million over budget should
      not come as a shock to anyone. We always knew the initial estimate was way under the
      real cost. Just like we never had a real cost for the [San Francisco] Central Subway or the [San Francisco-Oakland] Bay Bridge or any other massive construction project. So get off it. In the world of civic projects, the first budget is really just a down payment. If people knew the real cost from the start, nothing would ever be approved. The idea is to get going. Start digging a hole and make it so big,
      there's no alternative to coming up with the money to fill it in."
      This quote was cited by Bent Flyvbjerg in his report about megaprojects.
      Flyvbjergs analysis:

      Rarely has the tactical use by project advocates of cost underestimation,
      sunk costs, and lock‐​in to get projects started been expressed by an
      insider more plainly, if somewhat cynically.

      Hirschman observed that humans are “tricked” into doing big projects by their own ignorance.... He called this the “principle of the Hiding Hand.” John Sawyer similarly identified what he called “creative error” in project development as the “miscalculation or sheer ignorance” of the true costs and benefits of projects, with such miscalculation being “crucial to getting an enterprise launched at all.”

      It is easy to understand why Hirschman’s and Sawyer’s theories have become popular. The theories encourage promoters and decision makers, like Willie Brown above, to just go ahead with projects and not worry too much about the costs or other problems.

    2. I don't know if it's just politicians. Didn't Texas Central, a private entity, initially come up with an 8 billion estimate for their system. I believe they're currently running between 15 and 20 billion, depending on what you include. And they haven't even started construction yet. The US being the HSR novices that they are, are just now getting a grip on the realities involved.

    3. Another example, Boeing's miscalculation and screw-ups on one jet. 19 billion in loses thus far and they still don't have a plane for delivery.

    4. HSR has been done before. The US has designed and launched rockets for 60+ years.
      Yet ...
      "According to NASA's own independently verified numbers, SpaceX's development costs of both the Falcon 1 and Falcon 9 rockets were estimated at approximately US$390 million in total."
      For the Space Launch System (SLS) uber-rocket launch tower at Cape Canaveral
      -- NASA started with the Mobile Launcher 1 tower.
      -- NASA's original estimate for modifying the launch tower was $54 million,
      according to the report by Inspector General Paul Martin.
      -- In 2010, NASA then spent an additional $693 million to redesign and modify the
      structure for the SLS rocket.
      -- Now, in 2020, analysis finds that the total cost of constructing and modifying
      the structure, is "at least" $927 million.
      -- The SLS rocket development will have spent ~$18 billion by the end of 2020,
      while not having flight-ready hardware presently. SLS is estimated to cost
      $2 billion per launch.

    5. And SLS is designed, built, and tested by ... Boeing.

    6. To the article's underlying point, people don't want to commit to a single big city therefore those transit systems encounter problems whereas suburban commuters always seem to get what they want. SF is much worse than LA in this regard, and vice versa Metrolink is worse than Caltrain because San Bernadino, Orange, Ventura and LA counties do not cooperate in the same ways as SF, San Mateo, and Santa Clara do.

      Things do change. The combination of the 737-Max and Covid-19 debacles is a crisis for the airline industry, and it will be harder for them to justify a bailout when they were already bankrupt 15 years ago. UAL's backstabbing hasn't been forgotten, and even if they are granted secured loans this time the public's tolerance for it has waned. At the same time, there seems to be a much more diverse cast looking for at least regional rail, if not transit outright. Existing investments in Utah and Colorado are bearing fruit as systems slowly expand based on ridership and a desire to be connected into a larger commuter economy. This is the opposite of Europe and Asia where cities are far more compact, and only develop commuters after establishing robust local networks to tie into.

      Already this argument is playing out in NY where it's explicitly about airline bailouts vs rail, a major step up as 30 years ago it was roads vs rail and 40 years ago few wanted rail period. People are suddenly realizing the opportunity in what Amtrak has managed to preserve. Even in Washington state, home of Boeing, there is a serious debate over the cost of subsidizing Boeing employees vs building the Cascades into HSR.

  13. Breaking News: https://tjpa.org/uploads/2020/04/Item11_DTX-MOU.pdf

    1. And now for the pork: https://www.sfcta.org/sites/default/files/2020-04/SFCTA_Board_DTXAllocation_2020-04-14_0.pdf

  14. LA-to-Las Vegas Virgin Train Wins $600 Million California Bond Allocation

    Approved unanimously in Sacramento on Tuesday by a committee overseen by California Treasurer Fiona Ma, Virgin Trains-Brightline can sell up to four times the allocation amount, raising as much as $2.4 billion for the project. The company, which operates the Brightline rail service in South Florida, is also awaiting word on a $200 million private activity bond allocation from Nevada that would raise an additional $800 million. The U.S. Department of Transportation last month provisionally awarded the project $1 billion of private activity bonds, raising potential funding for the 180-mile-long project to $4.2 billion. That’s just $800 million shy of the railway’s $5 billion construction tab.

    “West Coast, we’re coming for you,” Brightline tweeted on Tuesday afternoon. “Today, we were officially approved for tax-exempt bonds for our privately funded rail line between Las Vegas and Southern California. This means on track to break ground by end of this year.”

    1. After Bakersfield and Merced, Palmdale to Burbank next for CAHSR?

    2. Berryessa-Santa Clara. First things first for the WSP/PBQD parasite DBA the "public agency" "CHSRA".

  15. It's obvious that Caltrain ridership won't be coming back for ... several years.

    Other than executing the agency staff and perma-temp contractors (an excellent idea under all circumstances) and mothballing the line, any rational person would be seek to minimize operating expense while maintining some sort of not-irrelevant and semi-useable service level.
    "Semi-useable" means headways of 15 to 20 or even 30 minutes, all day, every day -- in other words comparable to BART's worst service.

    This requires one person train operation and minimizing the number of trains required to provide a level of service.

    This requies level boarding (at ~650mm, not unreal not-happening CHSRA pulled-out-of-their-own-asses "standard" high-level) and no diesels.

    (An assimption here here that the catastrophically mismanaged and unworldly over-cost electrication and EMUs are a sunk cost, and that retrofitting diesel-hauled Bombardier trains for OPO and level boarding is uneconomic, through as I pointed out earlier above diesel level boarding would have been equals to electrification with slow dwells, but over a billion tax dollars cheaper. Ugh)

    Scenarios involving 8tph and wheelchair boarding assistance and remnant diesels and deputy-sub-assistant-triplicate-conductors have no relevance.

    The number of trains required to provide service is the headline ... and always has been!
    Work back from that.


    20s dwell, 7% pad (level boarding), 15min headway, 13 one-person-operated trains score 127.

    20s dwell 7% pad, 20min headway, 10 OPO trains (painfully close to getting down to 9 trains, which could be done with 30 seconds less turnback deadtime) score 120.

    20s dwell 7% pad, 30min headway, 7 OPO trains score 110.

    Get raising and extending (incrementally to 300m = 2x6car trains) the platforms, now, using "infrastructure" capital slush finding.

    1. "(painfully close to getting down to 9 trains, which could be done with 30 seconds less turnback deadtime)"

      Make that 3 minutes less turnback time: 12 instead of 15 wasted minutes for a train in service northbound to re-enter service southbound and vice-versa. (Your run-of-the-mill non-19th-century non-US regional rail operation can and does do this in 4 to 6 minutes, reliably and repeatedly.)

      20 second stations dwells and 7% timetable padding (level boarding), 20min headway (as good as BART's worst level of service at any time on any line), 12 minute turnbacks can be done with NINE one-person-operated trains, scoring 120. 86.4% utilization! (Fedbruary 2020 Caltrain was 68.6% ugh.)

  16. OK, here's another entry in the lowest-cost lowest-fleet sweepstakes.

    Not assuming level boarding (just like Caltrain's professonal staff have, for three decades!)
    So retaining stupidly long 45 second station dwell times and 10% schedule padding baseline.

    10 trains, score 113. (10 minute turnbacks required for a 10 train fleet, but that's asking nothing.)

    The cost? A new crossover (and "control point") immediately south of Belmont station.

    The trick? Southbound "Santa Clara Limited" trains run "wrong track" from just south of San Mateo station (50mph crossover "CP Palm" about 0.4 miles south of the platforms) through Hawyard Park and Hillsdale and stop "wrong track" at Belmont, where they have a cross-platform transfer after catching up with a "Belmont Shuttle" that preceded it.

    Immediately south of Belmont station, on a new-build crossover, the limited train returns to "right track" and makes all stops to Tamien.

    Northbound "Santa Clara Limited" run "right track" the whole way. At Belmont there is a cross-platform transfer to the "Belmont Shuttle".

    The shuttle waits for the limited to depart, then departs from "wrong track" and crosses to "right track" at existing "CP Ralston" 50mph crossover about 0.3 miles north of the platforms. Alternately, the northbound shuttle could depart at the same time as the limited and run "wrong track" stopping at Hillsdale (new island platform configuration, so it doesn't matter!) and stupid should-be-killed Hayward Park (stopping "wrong track", but all trains in both directions serving Hayward Park use the same platform, so no confusion) and not crossing back to "right track" until "CP Palm" south of San Mateo. Either works.

    It's a hack, and the turnback should of course be at Redwood City, but it's super-cheap. Again 10 trains operate the service pattern.

    FSSF IS NON-NEGOTIABLE and has hugely negative cost for future build-outs, because of the flexible service it enables.

    1. For reference, the same service pattern (SF-Belmont shuttle, cross-platform transfer, 30 minute headways) but with level boarding (and one person train operation!) tightens up to 9 train fleet, 77.5% fleet utilization, score 122 (and even relaxes turnback times to 14 minutes.)

      9 trains, score 122.

      It did surprise me that the Taktulator scores a Belmont shuttle so close to the "obviously" superior Redwood City (so close, yet so far) turnback, and that sticking with30 minute (not 15) headways isn't a scoring disaster, but I'm just playing the game here. ("Gaming" here includes serving Broadway and Hayward Park, which in reality demand permanent closure with prejudice.)

      Level boarding, island platforms.
      Level boarding, island platforms.
      Level boarding, island platforms.