RR503

Takes a certain type of transit nerd to be interested in planning subjects but somehow unaware of the access plight of satellite job centers. Some reading: http://library.rpa.org/pdf/RPA-Trans-Regional-Express-T-REX.pdf#page=27

Inter-agency politics aren’t even the reason this hasn’t happened — it’s construction costs and imagination! You’re connecting LIRR to LIRR through LIRR’s basement terminal at GCT here; nobody else need be involved. The only question in my mind is whether this is a good use of Atlantic’s expansion potential. There are many good arguments to be made for extending that line to connect with MNR or NJT rather than looping back on itself to foster through-core travel. A train from Valley Stream to White Plains would be quite useful.

Yes, that'd be what I'd do. There are also a whole number of plans with an extension via 10th Ave that'd be doable/potentially good. There is fast approaching a day when, even with CBTC, QB express will not be able to take more passengers. That day will bring ridiculous dwells at QB exp stations, unpleasant travel, etc. Deinterlining is the only realistic fix for that, and frankly doesn't end up costing riders that much time if done right, while also, through simplifying train movements, allows for greater overall capacity. We can also just build the passageway!

As hinted at in the above, it's really a question of dwell times. CBTC definitely will move trains faster given any dwell because you can close in at exactly the safe speed rather than via ST cutbacks (which, being a form of intermittent speed control, will necessarily have to enforce a lower closing speed to preserve safety), but that doesn't mean you're going to be flying into Roosevelt during the rush; CBTC is vulnerable to dwells too. Another angle I'd be interested to investigate is the way in which the interlockings at either end of Roosevelt interact with CBTC. CBTC treats interlockings like areas of fixed block control which actually is quite limiting to capacity; I'd love to know how that's working on the . I mean, no? If you ran more PM peak Concourse local svc you'd also probably be running more AM peak Concourse svc, which would mean you need more trains at Concourse to begin with.

"Organization before electronics before concrete" should be a saying we all keep in mind. The bypass doesn't do anything for capacity that deinterlining couldn't do, and moreover forces the preservation of all sorts of ugly merges. The constraining factor on Queens today is the number of Manhattan-bound tunnels; the Bypass doesn't expand that figure and thus doesn't really get to the heart of Queens' issue. I'd prioritize some new line across the river over it.

Not really, no. And for CBTC to be effective, you'd have to move the switches around. Ocean Parkway would face the exact same issue as Stillwell -- extremely restrictive signalling coming into the platform because of active switches lined against traffic at the end of the platform. Do not recommend!

Set included 4830. Was on the same trip as Friday AM.

A set of 42s ran this AM. On their last legs to be sure, but not gone yet.

Sad to see the 42s go. The railfan in me always preferred them to the 32s -- much better looking cars than their older brethren.

No disagreement from me here! Concourse should have a lot more service than it does today; it's a critical Lex reliever. Yeah, you cannot merge one train one minute behind the next. Just doesn't work! Hence the suggestion of timed connections at Newkirk. I certainly get your argument for KH meets, but do think Newkirk is better. For one, the calculation shouldn't be "how many stops do transferring riders get to skip" but instead "how many transferring riders can benefit from a zero penalty connection to a train that'll be slotted ahead of the one behind it/what's their savings" which is a calculation that heavily weights the connection point towards Newkirk. Secondary is the issue of track capacity in/around Dekalb Avenue. Scheduling an even 3 minute headway minimizes the chance of congestion through the crappy track geometry between Atlantic and Dekalb, and also minimizes the impacts of merge delays at Gold St. Newkirk allows a ~3 min headway; it's the one that should be done. Reconfigure, my bad. The point remains, though, that the benefit of sorting riders at _one_ (not particularly busy) stop simply does not justify the costs involved in extending round trip runtimes by 10-15 minutes and removing a crossover. The incremental cost incurred there is absolutely enough that it'd be taking away from service increases--that's a 20ish percent increase of runtime! Nobody is saying we should short turn instead of running to CI or instead of rebuilding CI. Both of those things should be done. However, the genesis of this conversation was the fact that even with moved switches, CI will still suffer from an extremely restrictive signal system design that prevents operation at turning capacities normal for similar ops in the system. I suggested that, if the needed turning capacity above and beyond what could be provided at CI, this would be a way to create it. This has nothing to do with wanting to create a relay op or the being "special" -- just a recognition of CI's limitations as a place to turn service.

The schedule version of that, yes. Let’s say NYCT expects trains to lose 8 minutes along Queens Boulevard. The supplement schedule they write will have an 8 min long hold at the south end of the line (that location being itself a massive issue given having trains run late except for the last few stops throws merges to hell) with the expectation that the will arrive at that station 8 mins late and thus will not need to hold. @Union Tpke yes! 11 St is where the gets added time for QB.

There’s incremental capacity available. will never run 30, but 15? Sure. 18? Maybe. Getting the infrastructure ready to support those 50-80% service increases would be good, and would also help prevent the ever present post disruption terminal holdouts. I disagree. Synchronizing them that far south means the will arrive at Prospect Park *just* behind the in front of the one it connected with. Tying them at Newkirk means that won’t happen, and moreover means that everybody from Newkirk south on the gets a cross platform connection with an express that’ll end up in the slot in front of the they boarded at their local stop. It works operationally and from a planning perspective; it’s what should be done and as you say, given the relative reliability of near terminal service, it generally _is_ done. The average passenger has made that mistake once and knows not to do it again. At any rate, the solution to a customer communication issue is....better comms, not an all out terminal rebuild — one whose final configuration would, by the way, be significantly more costly to operate than current service. Yes, my proposal for switches for short turn function got conflated with a proposal to turn the into a relay op at BB. Another reason to oppose relay, FWIW, is that it would prevent the construction/use of local short turn relays at the station, as s would be occupying the relay slots on the express.

Strongly disagree here. There's nothing inherently restrictive about the terminal layout at Brighton Beach -- you can move the crossover and achieve more capacity -- and you frankly do not need more capacity there. The will always limited to [capacity of Dekalb]- throughput, which isn't all that much. Like sure, terminal resiliency is good, but crossover movement + drop back crews could likely get you a theoretical capacity of 30tph, which will almost certainly never be needed. There's also nothing preventing operation of service such that leaves just after -- it's just most people prefer that the and align at Newkirk or so so that a maximum of people can transfer onto an express that will skip ahead of the local. On the legibility front, people transferring from usually do so at Sheephead Bay where everything is on the track that it's supposed to be on; the only people who really get subjected to the confusion are riders originating in Brighton Beach, and they have next train indicators in the mezzanine for their sake.

Yes, though increased weekend/midday running times are almost always reflected through end-of-line holds at gap stations rather than actual interstation adjustments. A fun bit of trivia: there's one extremely common runtime add that's done in an interstation because there are no gap stations convenient to the corridor that needs extra time. Can any of you name it?

Dyre-180 and Flatbush to Franklin are the same because...they are. There's no really big truth to be found in that. Northbound runtimes on the overnight from Flatbush to Franklin are 12; 13 during rush hours. Southbound you have a padding hold at Church, which blows runtimes up to 20. Honestly, though, padding isn't the best way to think about longer schedules. *To a point* longer schedules help make service better because your merge interactions work, your crews are in the right places, etc. Issue is a) when longer schedules aren't fine grained adjustments to segments with longer runtimes or that will be subject to some supplement/GO congestion but instead just padding spammed at the ends of lines or b) when you add so much runtime that trains run hot and end up getting held a bunch, lengthening actual runtimes. Really this all is just an argument for moving to runtime-based service metrics....

Thing with doing Bway Jct-Marcy express is through eliminating skip stop w/o eliminating runtime gains, you'd make boarding at stops in Woodhaven/Richmond Hill much more appealing than taking the bus to QB. Crescent-Bway Jct would make that even more true, but honestly is icing on the cake -- big ticket item is getting peak headways below 10 mins. Re: the , Queens Boulevard is certainly part of the issue, but it's by no means the whole issue. Beyond the runtime issues with the 11 St cut merge and approaching Forest Hills, QB's biggest impact on service is making its schedule incredibly complicated. You're not only weaving the through Astoria traffic, but you're now tying it to the IND as well -- makes it impossible to schedule even, hold-free service and ridiculously hard to implement any schedule remedy for service issues because you'd basically have to rewrite the entire B division timetable to do so.

Re: this express-on-the-local proposal, as the old saying goes, the difference between a local and an express on the local is one stops at stations while the other stops between them. You can make this operation work with super coordinated schedules and low frequencies, but on the , where peak headway is 3 minutes, skipping more than 2-3 stops will put you hot on your leader's tail. Merging is unfun, but if we're going to do an express, we should do it properly. To this specific proposal, assuming ~30 seconds of savings per stop, you'd be gaining about 3.5 mins of runtime against your leader. At 12tph frequencies, that leaves about 90 seconds of buffer between you and the train in front of you, or the absolute minimum headway achievable on the densest of fixed block systems -- assuming everything goes right. And FWIW, the Jamaica el hardly has the most capable signal system. You'd _maybe_ (and a strong maybe at that) be able to do express on the local from Crescent to Bway Jct (2 mins of savings/4 mins of buffer), but even then I do wonder whether the inevitable variability makes it a better alternative to just build out the express track above the current el.

The bumpers are definitely the bigger issue — they mean you have to crawl in on GT. To be clear, the suggestion is keep the current track arrangement, and just run VCP trains express from 96 to 137 via M track? Or am I missing something?

Oh, no contest on any of these points. Merely trying to elaborate on why this proposal hasn't been implemented. It's not so much about skipping more stops as it is allowing the to go to Lefferts without adding merges with the . We're obviously not getting a 4th track on the Liberty El or a subway extension beyond Euclid, which leaves us with the middle track on the El to work with. You won't be able to avoid merges in the off peak direction, but with this design at least in the peak direction you're avoiding adding those interactions. I infer you're talking construction logistics here? I don't think those would be _that_ difficult given that this would be happening under a grassy median and the parking lots that surround Grant Ave. Doing a subway connection as you suggest would be great, but would require a much larger commitment -- this is somewhat deliverable within a reasonable timeframe. runs 18tph into the core right now, but can only add 6 more thanks to the crappy terminal config at South Ferry. I'm sure that'll improve if/when they add CBTC there, but for now that's the number I'd work with. It's certainly enough to run express service, the questions I'd pose are - How does this affect loads/how can we mitigate those effects? - (as you noted) How does this interface with our service priorities for CPW? - How do we avoid 137 becoming the next Parkchester?

Definitely -- it's an idea I've toyed with too. Only qualifications I'd make is that I'd want to analyze the impact such a service pattern would have on crowding, as well as what the most efficient way to route short turns through that complex area would be.

I'm not sure that makes it much better--people who don't want an el because there's really no upside for them in supporting it aren't going to be all that moved by us telling them it'll actually be 20 feet further away from their front windows than they thought. It saddens me that this is how things work in this city, but it is, alas, the only way things get done it would seem. Just sayin', we already have an elevated structure through the area with decent ridership to boot!

True, but again am wondering whether at least 20mph can be squeezed out of the existing curve geometry. If you can get it up to that, returns to realignment start to go down a lot. It's not just politicians -- MTA doesn't like extension because they can't get 3 roundtrips out of crews if they do it. Not in any way an insurmountable issue, but it's just yet another little barrier between current svc and a extension. As for the proposal, I was suggesting a Concourse Line sort of arrangement -- s to and from the Rockaways run express in the peak direction, local in the other. runs local in both directions. It helps a lot if you don't want to have to choose between a Lefferts and a Rockaways train! Just get off at Rockaway Boulevard and wait. I would staff Rockaway Boulevard better so that fumigation is quicker. Other than that, really the only way is to ensure you schedule adequate space between s and s.

Those are really bad headways! The 's branches don't qualify as frequent transit, to say nothing of the fact that they are egregiously in violation of MTA service guidelines (not that those really are adhered to anyway). Thanks to the complexity of branching and the complexities inherent in end-of-day layup patterns, there actually are considerable periods in evenings throughout the week where headways on both branches go and stay above 20 minutes -- 24-30 min headways aren't uncommon. This merits a fix, and as it so happens this proposed fix could chip away at long commutes from the Rockaways. I say worth a look.

I'm still skeptical that, politically speaking, one can build a new el without any stops to provide community benefit... Also, Crescent is a legit activity center! Would be curious to know what could be achieved at the curves there w/ better superelevation and CBTC to address the constraints of fixed block control over curves.

You certainly _could_ do it, I just don’t think it’s worth the investment. You’re fixing an issue that’s only present for a few months of the year and could be largely mitigated by just improving ops practices w/r/t relays. Now, if this is a part of a larger plan to realign the middle track of the Lefferts El to feed the express tracks at Euclid instead of Pitkin so you can run to Lefferts deinterlined in the peak direction, I’m all ears...