Fantasy T maps

[ritchiew]

I've been to Munich, and I think it's a good parallel, and also demonstrates why I don't think giving the regional rail and subway lines equal thickness is actually reasonable. In the central tunnel S-Bahn trains come every ~2 minutes, it's the most frequent rail line in the city. Outside the central tunnel, each line only comes every ~20 minutes or so, half of the average frequency of an individual U-Bahn Line (Or 1/4 the frequency of the common 2-line U-Bahn sections).

I see. I forgot the Munich branches were so infrequent.

I have service in the central tunnel every 2 minutes, service within 128 every 5-8 minutes,

the portions within 128 are a comparable weight to the subway lines, and the portions outside 128 are noticeably thinner.

This mostly makes sense. But if Regional Rail trains are running every 5-8 minutes inside 128, to me, those are subways. So I don't know why they get a "comparable weight" to the other subways. They should probably just be the same. They're subways.

Then noticeably thinner lines past 128 makes a lot of sense. Like, with some label just saying "to Newburyport," maybe?

 

[TheRatmeister]

This mostly makes sense. But if Regional Rail trains are running every 5-8 minutes inside 128, to me, those are subways. So I don't know why they get a "comparable weight" to the other subways. They should probably just be the same. They're subways.

The same reason it's not shown as one color in Munich, it's because it's a collection of services with distinct endpoints that, for the most part, only share relatively short sections. (Although with the service pattern I described it's not that short.)

 

[JeffDowntown]

The same reason it's not shown as one color in Munich, it's because it's a collection of services with distinct endpoints that, for the most part, only share relatively short sections. (Although with the service pattern I described it's not that short.)
View attachment 75012

I have to think that thru-running Regional Rail in a future NSRL scenario is going to need to define specific endpoint pairs as "lines", as done with the S Bahn and RER in Europe. Otherwise interpreting the schedule would be a total nightmare for riders.

 

[F-Line to Dudley]

I have to think that thru-running Regional Rail in a future NSRL scenario is going to need to define specific endpoint pairs as "lines", as done with the S Bahn and RER in Europe. Otherwise interpreting the schedule would be a total nightmare for riders.

The 2003 NSRL Major Investment Study took a very rough stab at trip pairings for the 2-track and 4-track builds in its Operations Study chapter. The results were...not very elegant at all.

Given that there's an inherent mismatch in south vs. north endpoints, I don't see how you run the whole system with fixed endpoints. Some things...like pair-matches with the Old Colony lines and Fall River vs. New Bedford to north...are going to have to vary if you see value in giving them rotation thru to northside instead of consigning certain endpoints to an always-surface turn (and thus having a caste system of run-thru haves and have-nots). Plus for schedule-keeping on a legacy system that's being retrofitted for run-thru instead of having the system grow originally as run-thru, you have to have flexibility to short-turn things that are running hopelessly late before they drag down every interacting schedule. So for that reason too you might only see "advertised" run-thrus to the opposite 128-sited short-turn with anything beyond that being non-guaranteed. Sort of like how SEPTA will only plot a precise schedule on any Regional Rail line through Center City, with the opposing endpoint being up-for-grabs. Although obviously our frequencies would be much higher so the transfer penalty for reaching a "non-guaranteed" destination isn't going to be all that onerous.

End-to-end ridership on a unified system is not going to be that large because the schedules are just too excruciating. It's all about where you draw the line across the divide for where the most inter-region ridership is going to be, and guarantee that. "495 to opposite-side 128" guarantee seems like it would be the organizing principle that satisfies the lion's share of demand vs. threading the needle guaranteeing (and failing pretty often) true endpoint-to-endpoint service. And that pretty much means you only need to match mainline-to-mainline consistently and not so much branch-to-branch.

 

[JeffDowntown]

The 2003 NSRL Major Investment Study took a very rough stab at trip pairings for the 2-track and 4-track builds in its Operations Study chapter. The results were...not very elegant at all.

Given that there's an inherent mismatch in south vs. north endpoints, I don't see how you run the whole system with fixed endpoints. Some things...like pair-matches with the Old Colony lines and Fall River vs. New Bedford to north...are going to have to vary if you see value in giving them rotation thru to northside instead of consigning certain endpoints to an always-surface turn (and thus having a caste system of run-thru haves and have-nots). Plus for schedule-keeping on a legacy system that's being retrofitted for run-thru instead of having the system grow originally as run-thru, you have to have flexibility to short-turn things that are running hopelessly late before they drag down every interacting schedule. So for that reason too you might only see "advertised" run-thrus to the opposite 128-sited short-turn with anything beyond that being non-guaranteed. Sort of like how SEPTA will only plot a precise schedule on any Regional Rail line through Center City, with the opposing endpoint being up-for-grabs. Although obviously our frequencies would be much higher so the transfer penalty for reaching a "non-guaranteed" destination isn't going to be all that onerous.

End-to-end ridership on a unified system is not going to be that large because the schedules are just too excruciating. It's all about where you draw the line across the divide for where the most inter-region ridership is going to be, and guarantee that. "495 to opposite-side 128" guarantee seems like it would be the organizing principle that satisfies the lion's share of demand vs. threading the needle guaranteeing (and failing pretty often) true endpoint-to-endpoint service. And that pretty much means you only need to match mainline-to-mainline consistently and not so much branch-to-branch.

Short turn destinations can be defined pairs as well. The key is you don't want to make the whole mess "purple line to somewhere, at sometimes".

 

[HenryAlan]

As a rider, I'd be pretty okay to just go catch the next train at my local station, then make a downtown transfer if the destination isn't the one that I need. Keep in mind, the vast majority of trips will involve the terminal zone as one of the destination pairs. It really wouldn't matter to me if some trains servicing my neighborhood went to Newburyport, some to Fitchburg, or wherever.

 

[JeffDowntown]

As a rider, I'd be pretty okay to just go catch the next train at my local station, then make a downtown transfer if the destination isn't the one that I need. Keep in mind, the vast majority of trips will involve the terminal zone as one of the destination pairs. It really wouldn't matter to me if some trains servicing my neighborhood went to Newburyport, some to Fitchburg, or wherever.

But you need an easy way to know which train is which. Defining lines by destination pairs does that (that is how S-bahn lines are designated in Germany, although they share lots of overlapping tunnel/right-of-way). And many people would wait 15 minutes to take a direct train to a destination versus a connection. People quickly learn that to go directly between X station and Y stations, they need an S-8 or S-9.... Skip the S-3 or S-4 that come through, because they will need a transfer.

 

[ritchiew]

As a rider, I'd be pretty okay to just go catch the next train at my local station, then make a downtown transfer if the destination isn't the one that I need. Keep in mind, the vast majority of trips will involve the terminal zone as one of the destination pairs. It really wouldn't matter to me if some trains servicing my neighborhood went to Newburyport, some to Fitchburg, or wherever.

That's fair, and maybe that's how it ends up working due to infrastructure limitations. But if at all possible, it'd be much easier for riders if there were just a few clear, consistent routes, at least inside 128-ish.

And that's true that most people are going to/from downtown, but there are some pretty important exceptions. Mainly Back Bay. A rider from northside should be able to know really instantly "is this train going to Back Bay?" Same for Ruggles. Ideally, with NSRL and S-Bahn-like service, there will be other important non-downtown commercial/residential hubs, like Lansdown, Chelsea, Lynn, Quincy, etc.

 

[jklo]

And that's true that most people are going to/from downtown, but there are some pretty important exceptions. Mainly Back Bay. A rider from northside should be able to know really instantly "is this train going to Back Bay?" Same for Ruggles. Ideally, with NSRL and S-Bahn-like service, there will be other important non-downtown commercial/residential hubs, like Lansdown, Chelsea, Lynn, Quincy, etc.

A lot of those... I imagine North Side riders would just do what they do today, ie: get on the Green or Orange line when they get to North Station.

 

[ritchiew]

K, based on this discussion, I slapped together this transit map just for the S-Bahn-like network. It's super rough, but shows how I was thinking it would actually operate and how to best show that.

The four thick color lines work effectively like subways. In that region, trains come every 8-ish minutes or better. You get on one color line and you know where it goes for a long distance in either direction. For example, get on the Blue line at Back Bay, you'll know it goes at least as far as Beverly in the north and Canton Junction in the south. Get on the Green line anywhere and you'll know that train goes as far as Framingham in one direction and Reading in the other.

Past the thick lines, no guarantees. The train might branch there, so you need to check which way it's going. It might short-turn here and not go further. It might keep going to whatever further out terminus.

There are a lot of assumptions in this map that could be wrong but seem reasonable for a really rough sketch like this. I think those proposed short-turn points are right. Needham line is converted to OLX and GLX. Franklin Line runs onto Fairmount. Obviously any of that could change.

What I was hoping for initially upthread was something like this but 1) good, and 2) with all the other existing subways integrated into it as well. The result would be a new Boston subway map with eight lines: the current four plus the four new S-Bahn-like lines. Doing that in a clean and clear way seems like an undertaking, or is at least beyond me.

 

[samsongam]

What I was hoping for initially upthread was something like this but 1) good, and 2) with all the other existing subways integrated into it as well. The result would be a new Boston subway map with eight lines: the current four plus the four new S-Bahn-like lines. Doing that in a clean and clear way seems like an undertaking, or is at least beyond me.

Seems a little bit like you're describing my goal in my ongoing fantasy map, just without all the extra fantasy of SLX, Urban Ring, BLX etc. Maybe with a bit more color distinction between the S-bahn lines. But it's difficult because we have so many colors and connections already, it might become confusing quickly. I thought it would be better to brand them by numbers but sharing a color to signify that they are the same sort of service, similar to the green line branches.

I'm currently still working on making the thickness of the lines represent the different types and frequency of transit. But I could try to put togethr a more short term version, post-NSRL.

I think those proposed short-turn points are right. Needham line is converted to OLX and GLX. Franklin Line runs onto Fairmount. Obviously any of that could change.

I agree with your short turns, almost entirely. I think the pairings I disagree with a bit more. Wouldn't the Farimount line be most compatible with the Newburyport/Rockport line? They both branch in two at 17 miles from the central stations, they both have dense stops in developed population centers, and they both have more stops than the other lines on the trunk (potentially close to similar (11 stops to Beverly vs 12 or 13 stops to Norwood/Walpole) with concrete signs/plans for Everett Junction, Wonderland, South Salem, River Works and more theoretical but discussed infill stops at Revere Beach Parkway, Sullivan, and Encore).

 

[jklo]

Wouldn't the Farimount line be most compatible with the Newburyport/Rockport line?

It's not that really, just that Back Bay is very likely to be the destination... and One Seat is preferred. There def would be some grumbling if some NS trains go to Back Bay while others don't.

 

[bigeman312]

As a rider, I'd be pretty okay to just go catch the next train at my local station, then make a downtown transfer if the destination isn't the one that I need. Keep in mind, the vast majority of trips will involve the terminal zone as one of the destination pairs. It really wouldn't matter to me if some trains servicing my neighborhood went to Newburyport, some to Fitchburg, or wherever.

That's essentially the SEPTA Regional Rail model and it works well there, in my experience. Although, I personally like matching pairs for Boston for no well thought out reason.

 

[erf51trains]

Seems a little bit like you're describing my goal in my ongoing fantasy map, just without all the extra fantasy of SLX, Urban Ring, BLX etc. Maybe with a bit more color distinction between the S-bahn lines. But it's difficult because we have so many colors and connections already, it might become confusing quickly. I thought it would be better to brand them by numbers but sharing a color to signify that they are the same sort of service, similar to the green line branches.
View attachment 75075
I'm currently still working on making the thickness of the lines represent the different types and frequency of transit. But I could try to put togethr a more short term version, post-NSRL.


I agree with your short turns, almost entirely. I think the pairings I disagree with a bit more. Wouldn't the Farimount line be most compatible with the Newburyport/Rockport line? They both branch in two at 17 miles from the central stations, they both have dense stops in developed population centers, and they both have more stops than the other lines on the trunk (potentially close to similar (11 stops to Beverly vs 12 or 13 stops to Norwood/Walpole) with concrete signs/plans for Everett Junction, Wonderland, South Salem, River Works and more theoretical but discussed infill stops at Revere Beach Parkway, Sullivan, and Encore).

I have a question about the zones on this map. Why are some of the 128 P&R stations (Quincy Adams 128, Westwood 128, Dedham 128, Needham Crossing 128, Norumbega 128, Whopperton 128, and Reading 128) zone 3 while others (Hartwell 128, Mishawum 128, and Purchase St 128) zone 4?

 

[samsongam]

I have a question about the zones on this map. Why are some of the 128 P&R stations (Quincy Adams 128, Westwood 128, Dedham 128, Needham Crossing 128, Norumbega 128, Whopperton 128, and Reading 128) zone 3 while others (Hartwell 128, Mishawum 128, and Purchase St 128) zone 4?

Thanks, this is a good catch, that is mostly an error on my part. I generally think that 128 is good boundary to draw an approximate zone boundary as a ring of stations that are going to have a different ridership characteristic with P&R. In my overlay map, I include Mishawum and even Anderson RTC in zone 3, so that's an error. But Hartwell and Purchase St are a little different.

Purchase St and the entire Peabody line stuff needs to be in one zone for local trips to be straight forward and 95 branches distinctively northeast after Reading meaning it is a significantly longer ride on rail to there compared to the South and West side of the system. Hartwell (and now that I think of it more, at least Westwood, Dedham, Whopperton, and Norumbega should probably be similarly aligned to Hartwell) is a little odd because it is quite isolated out there, and it brought up a question in my mind. Should we try to make purely P&R stops cost more since it encourages or at least maintains car hierarchy, which has shown to be a bit of a waste of space and costly at other P&R stops around the system. I think there's no way you get around having stations with large parking garages next to 128 but, should we make those stops (which are not dense enough to be walkable or feasible for car free living) cost a little more to get to? Will the $1.70 more in fare make it so that people would drive one stop further in where there isn't as convenient parking/highway present?

Like for me, Reading 128 is a separate class of station even though it is similar distance because it is feasible that someone (some day with such extensive transit improvements) could live in the village car-free and rely on the train to get to Wakefield/Melrose for their day to day. I would think that the T and Mass would want to incentivize that lifestyle over driving your car from the suburbs to the train. But idk, it's a bit of a tricky one to consider when you start bringing up money incentives, we know that right now the CR is so ludicrously expensive that people will take a way longer bus to a bus/train than take the direct train from Lynn/West Roxbury/Roslindale because it's way cheaper, but that's a much larger difference than $1.70.

 

[ritchiew]

I agree with your short turns, almost entirely. I think the pairings I disagree with a bit more. Wouldn't the Farimount line be most compatible with the Newburyport/Rockport line? They both branch in two at 17 miles from the central stations, they both have dense stops in developed population centers, and they both have more stops than the other lines on the trunk (potentially close to similar (11 stops to Beverly vs 12 or 13 stops to Norwood/Walpole) with concrete signs/plans for Everett Junction, Wonderland, South Salem, River Works and more theoretical but discussed infill stops at Revere Beach Parkway, Sullivan, and Encore).

First off, awesome map!

As for the pairings, yeah, I really don't know what's best. If you incorporate this into your work, pick what makes sense to you. But here's my reasoning and what I think are the important factors (and I'm doing this from memory, so I might have something wrong here).

• The biggest factor is what infrastructure actually gets built. Original quad-tracked NSRL plans had level junctions to connect the two tunnels, one south of South Station and one North of North Station. If that's the case, and you really want to cram lots of trains though, that can be a big constraint. You'd really need to analyze the network and minimize how much trains have to cross over each others tracks. It'd be complicated. With that kind of infrastructure, I doubt my above example works. Maybe you couldn't have reliable pairings at all, and it'd have to work like how Philly apparently runs theirs. I don't know.
• One version of NSRL (I forget which) incorporates flying junctions on the north side. That would allow trains to run any pairing without having to cross over other tracks. That would make any kind of pairing much more feasible. That's what I'm assuming will get built.
• The next highest consideration is just giving the most people useful one-seat rides. That would be convenient for the most number of people, and means fewer people taking up space on platforms as they wait to transfer. You could do some fine grained analysis to figure the best pairings for that. As a pretty good approximation, I figured that effectively means that the two busiest northside lines need to go to Back Bay, and the busiest also needs to go through Ruggles. Newburyport/Rockport is by far the busiest northside line, so I connected that to Providence. The other three northside lines are basically tied for last, so I didn't even really think about which other should go through Back Bay. Maybe a different one would be better.
• FWIW, past NSRL studies have assumed that Providence Line trains have to pair with the Lowell Line because that's where NEC Amtrak trains would continue on. This makes no sense to me. Maybe there is some logic there, but I don't get it.
• The things you're pointing out, like stop spacing and number of stops, I don't think are relevant criteria. Or at least I don't get it. What's the reasoning there?