Penn Station Can Handle the Load: New York Is Ready for Through-Running
52 comments
·January 16, 2025bobthepanda
crazygringo
It's been that way for decade(s?) in New York. Presumably because otherwise too many people try to take it a stop or two for free before the conductor comes around to check tickets.
It's not a thing at any other Amtrak station I've been to, where the next stop often isn't for 45 minutes or more.
And no you can't have conductors in every train car. That's way too expensive and not needed for the rest of the 12 hour journey or whatever it is.
rayiner
Amtrak is run by the dumbest people in the country. Another gem:
> Today, Amtrak schedules the Acela, which travels express between Washington and Boston, to overtake the slower, local Northeast Regional at Penn Station. The organization claims that this requires scheduled dwells of 30 minutes for Northeast Regional trains
But why would anyone competent want to work for an organization like this?
thinkingtoilet
Do you have any idea of the history of why Amtrak was created? What limitations both funding and regulations are put on it? It's a god damn miracle it runs so well with what it's given. The dumbest thing is that we the people don't fund it and invest in rail, not people doing their best with what limited resource they are given in a country that is half a century behind when it comes to trains.
jazzyjackson
the freight companies lobbied for passenger travel to become the government's responsibility, IMO the american people were robbed of the value they were entitled to when the right of ways were granted to the railroads back in the 19th century, but in the 60s and 70s railroads were going bankrupt left and right, establishment of Amtrak was little more than a bailout for the railroads. The government could be asking a lot more of the railroads to run shorter trains so they could fit on their sidings and allow passengers to pass and remain on time. IME the #1 issue with Amtrak outside of north east corridor is most stations are served by long haul routes that cannot be relied upon to show up within 6 hours of their schedule, because any delay causes them to lose their slot and freight companies can tell Amtrak to pound sand, their trains are too long to pull over.
jazzyjackson
Total price segmentation, slowing down the northeast regional so people pay out extra for the Acela, nevermind that the northeast regional trains can run just as fast as acela for the vast majority of the route.
emchammer
Amtrak train crews are gems though.
jjice
I haven't been on an Amtrak in about six months, so it may have changed, but where were you boarding? Was it New York? Boarding in NY is slow and structured, but when I've boarded in Back Bay (Boston), Portland (Maine), and Providence, it's the better way that you've mentioned.
They make everyone rescan tickets for the North East regional in NY also, if you're just passing through. It's a bit annoying, but I wonder if the traffic getting on and off is too great in NY to be able to do that. I have no clue, though.
Ericson2314
Yes their practice elsewhere is better. Amtrak seems to think that lazzaiz-faire platform ingress/egress doesn't scale to NY Penn, but that's exactly backwards — it's not letting people circulate freely which doesn't scale.
Also, fun fact, if you don't go in the main hall Amtrak waiting area at NY Penn, you can board the platform whenever you want. But it's hard to figure out what platform/track to go to in advance — hiding that information is how they discourage this.
craftkiller
laissez-faire
nicwolff
Semi-secretly you can skip the lines at the escalators in Moynahan's big train hall and just go down to the mezzanine and right to the platforms.
Ericson2314
Yes the bad airline mentality goes a long way to denying these things are possible.
Once one accepts that people are going to constantly leaving and entering the platform, that all ticket checking must happen on the train to not impeed circulation, running way more trains needing fewer platforms is revealed as (a) possible, and (b) the right way to do things.
tdeck
I recently took Amtrak across the country. In both Seattle and Chicago I showed up about 10 minutes before my train departed. Sure, they checked my ticket before I got onto the platform, but the experience was absolutely nothing like taking a flight. I wonder if it was something about the particular day or station you were at that made this worse.
jackcarter
It's been this way for over a decade: https://www.vox.com/2014/3/31/5563600/everything-you-need-to...
renewiltord
This method of boarding at DC famously resulted in 100+ passengers waiting upstairs as the train left without them https://wtop.com/travel/2024/12/passengers-frustrated-as-amt...
screye
That's how the TGV and subways do it. You scan in.
It's unconvienent if a human needs to read a ticket. But tap-in or scan-in systems are pretty fast.
Gibbon1
I remember reading with horror that the California high speed rail will have the TSA doing security. That's absolutely last thing the systems needs.
TuringNYC
I ride into NYC 3x/wk via Penn Station. I realize ETANY is not affiliated with NJTransit, but I dont see how a system with the level of incompetence that NJTransit has could possibly handle more complexity.
NJ Transit can barely seem to handle service w/o Through-Running, so any discussion of expanding service seems premature. Here are some highlights:
- Inability to tell consumers ahead of time that trains will be stopped (even though they know well in advance.) Now, entire private WhatsApp groups have been set up where commuters warn each other of stopped trains and clogged stations. This leads to people coming to Penn Station only to find out trains are not running/cancelled/delayed. This is with a hub/spoke -- imagine if they expand beyond Penn Station into CT/LI.
- Inability/Unwillingness to communicate sources of blockages. There are ways to bypass NYPenn/Secaucus and go directly to Newark (PATH train). But NJTransit wont tell you where the blockage is, so its impossible to work around delays
- Inability/Unwillingness to communicate which trains will depart first, when multiple trains are backed up and queued up. People guess and hope they choose the "next" train.
- Regularly cancelled trains, esp after 7pm. They randomly cancel scheduled trains. No point in a schedule if you wont follow the schedule.
I'd want to contain the chaos of NJTransit to NYPenn Station and not beyond. At most, a 2nd stop at GC (like with LIRR did). The system isnt currently mature enough to be granted more responsibility.
Mind you -- this would be valuable. Folks who move to NJ necessarily cut themselves off CT jobs (esp hedge funds, etc.) So of course, having thru traffic from NJ all the way to CT would open up huge pools of job applicants and job opportunities.
Projectiboga
These arguements are also good to suggest they should not get the planned expansion of terminus tracks.
Ericson2314
> I realize ETANY is not affiliated with NJTransit
We are equally unaffiliated with all the railroads and transit agencies :)
We do have membership in NJ are very much as in the transit going-ons through the tri-state area. A big unifying idea is that the economic geography doesn't care about political boundaries, and so the transportation planning shouldn't either.
------
> NJ Transit can barely seem to handle service w/o Through-Running, so any discussion of expanding service seems premature. Here are some highlights:
You do raise a good point that through-running does require some amount of competence --- simply have a more interconnected rail network (revenue service on both sides) inherently means there is more potential for failures to propagate throughout the network. But, we'd like to believe this is surmountable.
------
1. First of all, NJT and the MTA have done it before! See "train to the game". See:
- https://en.wikipedia.org/wiki/New_Haven_Line#Meadowlands_gam... - https://www.njtransit.com/press-releases/take-train-game-met...
Of course, running some special event service is not the same as doing through running day-in and day-out, but it is something to build on.
------
2. What's the alternative?
- If there is no new expansion and no through-running, than the billions spent on Gateway way be rather wasted. That would be a huge embarrassment to the agencies, and politicians that stuck their neck out for the funding, alike.
- If Penn Expansion is not pursued for bullshit reasons, but the kayfabe is dropped and its done for the honest reason that no one believes agency competency can be improved, that is also embarrassing. Would the station actually be funded at that point?
We therefore think even just changing the conversation to acknowledge that ops competency, and not station geometry, is the binding constraint, would be a major improvement.
------
3. We have time
In line with the above counter-factuals, the rubber only hits the road (excuse me, steel hits the tracks :)) once Gateway is done. For better or worse, that is a long way off. This gives the agencies time to get ready.
- The money that would be been spent on station improvements can be spent on NJT "tech debt" instead --- all the behind the scenes infra that enables higher reliability.
- Congestion pricing should be raised, and hopefully the next crop of NJ politicians will be more open-minded and accept some money for NJT.
- Penn Station Access sending Metro North trains to Penn station makes for a good opportunity to "train" the agencies through-running, prior to Gateway being finished. And don't forget "train to the game".
------
> Mind you -- this would be valuable. Folks who move to NJ necessarily cut themselves off CT jobs (esp hedge funds, etc.) So of course, having thru traffic from NJ all the way to CT would open up huge pools of job applicants and job opportunities.
Yes, it is a huge opportunity! Our main report (https://www.etany.org/modernizing-new-york-commuter-rail) talked quite a lot about that. I wish our politicians were less provincial about state boundaries, and better able to visualize just how impactful expanding commute sheds is.
------
A final disclaimer: I am far less knowledgable on train things than the other ETA members, so take this all with a bit more grain of salt.
Projectiboga
Through running is to avoid an upcoming real estate grab / bodogale where Stephen Ross's Vonado wants to let the state of NY use emmenent domain, skip city land use review and demolish an entire block 8th ave to 7th fromst 31st to 30th street. This is for him to build mega office towers and build more terminus rail platforms.
The idea of through-running is to not have trains parked like that in an over crouded midtown. That way they can reduce the number of tracks and widen the platforms. It has beem done with great success in many big cities already. The idea would be to have NJ Transit trains run to platforms at their rail yard in Queens. This is a large ripe for development area between Long Island City and Sunnyside Queens. It could also go farther to Port Morris in the Bronx and link with Metro North and the further North East rail corridor. Long Island Rail could go to a new terminus rail yard in NJ which could have a bus terminal to reduce the amount of busses into NYC. The main issue I can see is some of those commuter runs are too long for crew rotations and may require crew shifts who just do last stop in NJ through to the outer boro NYC rail yards.
Here is the group advocating this idea. Rethink Penn Station NYC https://www.rethinkpennstationnyc.org/
mncharity
> Through running is to avoid an upcoming real estate grab [...] and build more terminus rail platforms
Could one craft a gold-plated through-running transit hub proposal which supports the grab? Then transit improves, grab or no grab. Given the dominant power of real estate in NYC, the TFA had for me the feel of a proposal from engineering to a c-suite with big divergent incentives. Could one tease apart the "don't build badly" from the "don't need to build" arguments? "It's unnecessary and undesirable, but if it's done anyway, at least do it right".
Ericson2314
Also for reference, https://www.etany.org/modernizing-new-york-commuter-rail is our previous report that describes what through-running actually is, why it would be good, etc.
This report is more narrowly tailored on refuting Amtrak's grossly mistaken reasoning in their recent study.
pcl
“Through-running involves operating trains across Manhattan and through to the other side of the city, instead of immediately turning them back to the suburb they came from, as is done today.”
Ericson2314
Imagine how nice it would be to take a single train from Flushing to Newark!
Ericson2314
Since it's midway through the article, note that https://github.com/effective-transit-alliance/platform-crowd... is the underlying model we used for passenger circulation.
CSMastermind
This seems to be part of a debate I'm unfamiliar with.
I am however familiar with this style of argument. The page reads like a lot of motivated reasoning, where they have a conclusion set and they're selecting data, massaging statistics, etc. to fit into their narrative.
They might be right in the points they're making, I have no idea, but the style of the article definitely makes me skeptical.
crote
Why not do the same with Amtrak's reports?
From a European perspective a 15-22 minute dwell time sounds ludicrous. My local train station is pretty large and designed around through trains, and a 7-minute followup between departures (so between train 1 departing on track X to train 2 departing on track X) is routine here.
If it can be done over here, what makes Penn Station so special that they need 3x as much time? Why can't Amtrak do what other railway companies have been doing for years? Amtrak is already lying about other countries using headway-based scheduling, so can they really be trusted about the rest?
I have no idea if this ETA article is 100% trustworthy, but Amtrak definitely isn't.
Ericson2314
Thank you! We need a lot more Europeans berating us, truly! :)
Ericson2314
I mean... it's a rebuttal. See my comment below: we already put out a report saying we thought through running was the right approach, and now we're responding to Amtrak's report on the topic.
Even if there were passenger circulation issues problems (and our passenger model was half-written when we did the first report), there are other solutions like decking over tracks to join adjacent platforms into a wider platform. We're happy that do not appear necessary, but if it did it doesn't fatally imperil through running as being the best option.
Finally, the number of outright errors and dubious claims in the Amtrak study makes us think they are doing their own reasoning. If you really don't believe my previous two paragraphs, just think ask yourself motivated reasoning seems less suspect! :)
Through-running, unlike 10s billions for a station expansion, is a cheap experiment. We can just try it! If there are more passengers than expected (yay! Though) and safe circulation does become and issue, that's great new info and very little money wasted.
librasteve
through running is a highly effective mass transit approach as exemplified in Munich & Berlin (S-Bahn), Paris (RER), and now London (Lizzy Line)
renewiltord
This is clearly written for an audience that already knows what's going on. I don't, so I'm curious. Is the difference between:
outer terminus------>ny penn|alight|board new train|----->later station
outer terminus------>ny penn|stay on train|---->later station
EDIT: Thank you to the commenters (who I can't answer because I'm rate-limited).
crote
It's about how trains are running.
Consider a rail network connected like A-B-C . With terminating trains you'd have one train run A-B-A, and another train run C-B-C. With through trains you'd have one train run A-B-C and another train run C-B-A. Terminating trains have to stop and reverse, which takes quite a lot of time because the driver has to go to the other end of the train. Through trains can just continue in the same direction, so it is a lot faster.
Because a through train occupies the track for less time you don't need as many tracks to serve the same number of trains per hour.
null
jccooper
The former (terminating lines) needs more tracks/platforms at the station, because turning the train around takes more time than stopping and continuing (through running) and turning around at the end of the line.
Having recently ridden Amtrak, easily the worst thing about it is the fact that they've decided to go with an airline style boarding process where the platforms are treated as a secure area and everybody has to single-file get their ticket checked to get onto a platform.
The whole advantage of a train vs a plane is that a train has many doors, allowing a lot of simultaneous boarding to happen; and they also already have conductors who check your ticket to make sure you are getting into the correct car, and another ticket check once the train is in motion. It would be significantly better at major stations to just have conductors at every train car on the platform in parallel doing ticket checks, rather than just have one funnel.