The Washington, DC region is great >> and it can be greater.


Metro will look at the "invisible tunnel" "later this year"

The idea of a virtual tunnel between the two Farragut stations on Metrorail was the subject of several posts on CommuterPage and here about two years ago. The idea is to allow passengers to exit one Farragut station and enter the other within a set period of time, counting the whole thing as a single trip rather than two trips.

Given that Metro is currently making significant changes to its fares, I suggested to Interim General Manager Sarles that now would be a good time to also incorporate this long overdue idea for several reasons:

  • Metro staff is already working on making many changes, so it would be more efficient to implement now than come back and make more changes later.
  • Many customers are understandably unhappy about fare increases. Implementing the invisible tunnel would provide a positive change in operations that would add convenience for some of your riders to help offset some of the negative PR that comes with fare increases.
  • It's a no-brainer: It decreases congestion at Metro Center, provides a speedier ride for some customers and has absolutely no downside whatsoever.
Mike Russo, Metro's Assistant Chief Engineer for Automatic Fare Collection Systems, told me that they support the idea, but it "will not be an easy programming effort" because of memory limitations in the faregates, but that they hope to "re-examin[e] the Farragut transfer concept later this year" after the current fare changes are done.

Here is the complete text of his letter:

Dear Mr. Offutt:

Thank you for your June 20, 2010 email message to Richard Sarles, General Manager of the Washington Metropolitan Area Transit Authority (Metro), regarding proposals to allow seamless transfers between Farragut North and Farragut West Metrorail stations. I have been asked to reply.

Metro is continuously seeking opportunities to improve the travel experience for customers. The proposal for transfers between the two Farragut Square stations is worthwhile and may well be implemented in the future, with appropriate upgrades to the existing system.

As you may be aware, Metro has hired an outside contractor to implement the fare changes recently approved by the Board of Directors. Part of that contract is a requirement to deliver a fare system that can allow rail-to-rail transfers, a capability that should include the "seamless" transfer you envision between the Farragut Square stations, allowing customers to continue on their journey while being charged as if they never left the Metrorail system.

Unfortunately, we have not had the opportunity to test that specific functionality, and preliminary indications are that it will not be an easy programming effort. Among the many concerns we must address are the memory limitations of the faregates, since the fare tables that go with this transfer structure are quite large.

This is not at all to suggest that the Farragut-to-Farragut transfer cannot be implemented in the future. The inclusion of that requirement in the contract is evidence that Metro intends to make the change because it would benefit both customers and the system. However, at the moment we are focusing all available resources on major, Board-mandated changes to the fare structure: adding "peak-of-the-peak" charges, implementing passes on Smartrip®, and meeting the IRS requirement on SmartBenefits® (separate benefits for parking and transit). Any one of these goals would be challenging, but we must complete all three, plus minor adjustments, before a rapidly approaching deadline.

We look forward to re-examining the Farragut transfer concept later this year, after we have had sufficient time to observe the latest fare changes in full operation and make any needed adjustments.

We appreciate your inquiry and your patience as we work to implement a fare structure that is reliable and applied fairly for all customers.


Michael Russo
Assistant Chief Engineer
Automatic Fare Collection Systems

Steve Offutt has been working at the confluence of business and environment for almost 20 years, with experience in climate change solutions, green building, business-government partnerships, transportation demand management, and more. He lives in Arlington with his wife and two children and is a cyclist, pedestrian, transit rider and driver. 


Add a comment »

Makes too much sense.

by Martin on Jun 29, 2010 3:12 pm • linkreport

Well, implementing this would be cheaper than building an actual tunnel, so hopefully they stick to this and get it done. I'm once again pleasantly surprised by Metro's responsiveness. Keep it up!

by Teyo on Jun 29, 2010 3:13 pm • linkreport

What about an invisible tunnel between Chinatown and Metro Center? Or a real one (as was proposed a few years ago if memory serves...)

by Jason on Jun 29, 2010 3:19 pm • linkreport

Sounds like one of Metro's contractors isn't holding up its end of the contract. I don't get how or why this should be any more complex than a Bus->Rail transfer, which the current system seems to be able to handle without a problem.

NYC's had this capability since 1997.

by andrew on Jun 29, 2010 3:21 pm • linkreport

Does anyone have more information about the memory limitations of the faregates? How old is the current system, is it a legacy thing, left over from the 70s-80s?

by Steve S on Jun 29, 2010 3:30 pm • linkreport


It's more complicated than a bus<>rail transfer, which just has to have a piece of information like "look at last transaction, deduct $0.50 if less than 2 hrs ago."

The Farragut West<>North transfer requires that the faregate figure out what trip you took before, how long ago it was, what trip you're taking now, what the total fare is between the two endpoints, etc."

by MLD on Jun 29, 2010 3:44 pm • linkreport

If (PrevExitStation == (farragutNorth OR farragutWest) && CurrentTime-TimeOutOfLastStation < 30minutes)
1. Set current entry station to previous entry station
2. Set time in to time in from previous station entry
2. open fare gate

There- I programmed your gate.

by A on Jun 29, 2010 4:02 pm • linkreport

There should just be a free rail transfer with a short period, say 30 mins, to/from any stations.

Also, if you exit from the same station you entered within a short period, it should be free -- say, if you realize you're in the wrong place, or you forgot something in your office -- or if there are system delays and you decide to give up.

by Gavin on Jun 29, 2010 4:04 pm • linkreport

They would have to have pretty severe storage limitations for text tables to be an issue. As far as programming I don't see how it's different. Rather than "you boarded a bus less than 2 hours ago" make it "you exited a Farragut station less than 30 minutes ago" or whatever.

by Brian on Jun 29, 2010 4:05 pm • linkreport

If my human memory serves correctly the "memory limits" they refer to affects the amounts of passes that can be stored in the system. With the new SmartTrip upgrades they will be able to include a larger amount of passes and still maintain compatibility or something along those lines.

by Joshua Davis on Jun 29, 2010 4:38 pm • linkreport

I'd love to have a free transfer from Capitol South to Union Station.

by Daniel on Jun 29, 2010 4:42 pm • linkreport

What Jason said.

In these two locations where a real tunnel had been planned or proposed, the virtual tunnel makes a lot of sense and is a lot cheaper than a real tunnel.

by Stanton Park on Jun 29, 2010 4:48 pm • linkreport

Another thing to consider is that the 17th Street entrance to Farragut West is closed on weekends. That would make the "invisible tunnel" less feasible.

by Matt Johnson on Jun 29, 2010 4:48 pm • linkreport

@A - Don't forget the line that credits back the money that was debited when exiting the Farragut station in the first place, either immediately or upon exiting at your destination.

Could Metro install an "invisible tunnel" smart trip reader that was programmed specifically for people using the invisible tunnel and limited only to people using it for that purpose? Then the programming could involve no debiting upon exit and approved reentry. There would need to be some provision that if the next entry were not within 20 minutes or at the other Farragut station, you were charged maximum fare or at least the relevant amount for the trip you did take.

by ah on Jun 29, 2010 5:55 pm • linkreport

Also, if you exit from the same station you entered within a short period, it should be free -- say, if you realize you're in the wrong place, or you forgot something in your office -- or if there are system delays and you decide to give up.

I have always found it unacceptable that this has not always been true. Why could that not have been programmed even into the paper farecards?

by ah on Jun 29, 2010 5:57 pm • linkreport

Does the credit need to be applied at the point of the transfer? If the transfer only applied to SmartTrip pass holders, there is an easier workaround that could get this running without relying on turnstile memory. The (interim) solution would be:

An automated script would compare all exits from one station to the entries into the next for the follow X minutes. All transfers would be identified and credits would be issued back to those cards. Since Metro's computers aren't heavily used during non-operating hours, an existing server could run the off-peak tasks. It's a simple DB script...

by Yoav on Jun 29, 2010 6:04 pm • linkreport


Now, imagine you and a friend (internet, CraigÂ’s list) both ride every day. You get on at Vienna and travel to Metro Center and your friend conveniently gets on at Metro Center and travels to Vienna. You determine that if you meet at, letÂ’s say Rosslyn, he gives you his fare card and you give him your fare card, you would both have fare cards showing you leaving from the same station you entered.

To prevent this, BART charges a $5.20 "excursion fare", for which you get the fun of up to a three-hour tour of the BART system.

by Michael Perkins on Jun 29, 2010 6:13 pm • linkreport

The fact that the faregates themselves are calculating exit fares from in-memory fare tables is a significant indicator that things are not so simple as a naive observer (myself included in that list) might think. From the smell of it, it sounds like the approach is to pay normal fare on exit from 1st station, and then pay on exit from the 2nd station the difference: "entrance to exit" fare minus "entrance to farragut" fare.

Again, my reading is that exit gates only have fare tables for the single destination. Adding free transfers means adding an additional full fare table (all farragut terminations in this case) - and you'd need another full table for any other transfer points. Why memory should be a problem when you can buy 2GB of memory for $3 I don't know, but my guess would be that the gates aren't able to address larger blocks of memory or have a very inefficient structure. Fixing that (if my assumptions are correct) involves real significant software and/or hardware changes across the system, and would be another issue entirely.

@A's approach with @ah's addendum (on entrance #2, reset the entrance point on the card and credit the first part of the fare) would seem to make sense, except that would quickly make it easy for one to make all Farragut-terminating trips free (run the card and don't go through the gate). I prefer the deduction approach's alignment with the actual process.

Bus transfers don't involve an "off" transaction and so are much simpler. In fact that could be a simplifying solution - if all transfers happen in the central zone, just charge a single simple "transfer fare" at the terminating station for people who transferred on, rather than ensuring that the fare is identical for folks who walk vs. change at Metro Center. No additional lookups required.

by Gary on Jun 29, 2010 6:51 pm • linkreport

@ Michael

Couldn't that be prevented by time limits

Lets say you walk through gates trains are delayed you have 20 minutes during non rushhour and 10 minutes during rush hour and 40 minutes when single tracking to leave back out and get a refund.

thats long enough that the PID would change atleast once and if you walked from one end of the platform to the other you have time to get back and nothing else.

Or they could just give you a certain amount of rides in a time frame. You could get on the train get off and get back on within 40 minutes without occurring a fare or give them half off the next ride.

WMATA is a threat to itself for short rides (1-3 stops) and especially if someone has errands to run. One time when I had to run around DC for some stuff I ended up paying around $11 in train/bus fare over 3 hours.

by kk on Jun 29, 2010 6:56 pm • linkreport

Maybe Metro isn't interested because now's not the time for them to be giving up revenue ... I.e., If this 'invisible transfer station' is going to save someone a significant amount of time over using an established real transfer station, chances are the person isn't going to mind paying another couple bucks for another fare. Letting them now make this transfer at no cost, would result in less revenues to Metro ...

by Lance on Jun 29, 2010 7:09 pm • linkreport

If you're a regular pass customer, this problem solves itself.

by Michael Perkins on Jun 29, 2010 8:05 pm • linkreport

Michael Perkins' explanation of why it shouldn't be free to get off at the same station you got on at makes a lot more sense than mine:

I discovered the charge when I used the metro tunnel as a way to get out of the rain for a couple blocks while walking west on U street: my presence in the system added to platform congestion, but I did pay for it, so I figured the charge was to discourage people from cluttering platforms unnecessarily.

by Lucre on Jun 29, 2010 8:18 pm • linkreport

@ Michael - I agree that possibility exists, which is why the "short period" in the quote has to apply to limit it. I'm just not that worried about a guy from Rosslyn and a guy from Foggy Bottom swapping cards/smart trips as they pass. And to get back to Vienna would go well beyond the short period (10 minutes, say) I envisioned.

by ah on Jun 29, 2010 10:05 pm • linkreport

+1 Daniel. It takes 15 minutes to walk to Union station from Cap South or 22 minutes by train. Which may not seem like much, but imagine when both stations have CaBi bike stations. In fact the CaBi could serve as a virtual tunnel in a few places around town.

On leaving the same station - once I went into Eastern Market and the train came and promptly broke down. They wouldn't refund me when I wanted to leave. :(

by David C on Jun 29, 2010 11:15 pm • linkreport

Silver Spring to Bethesda would also be a faster CaBi transfer than the equivalent trip by rail, although that's a bit of a stretch.

(Actually, this would make a great bus route, particularly if you made a special exception to allow a free transfer from the bus back to the other leg of the red line)

by andrew on Jun 29, 2010 11:55 pm • linkreport

How about an actual tunnel? Why is it so difficult to build anything anymore?

by Dan on Jun 30, 2010 8:26 am • linkreport

I'm with Dan on this. Why don't we all just pony up the dollars to pay for real tunnels? Hell, it'd even put people to work (where were the applications for stimulus dollars?). Hire guys with pick-axes and shovels. In the long-run it'd be the best option.

by rdhd on Jun 30, 2010 8:52 am • linkreport

Actual tunnels will take much longer to plan and build. It would have been impossible to use stimulus dollars for such a project, because it hasn't been designed or gotten its environmental clearances yet, processes which take years and are themselves very expensive.

In the mean time, why not do the virtual tunnel?

by BeyondDC on Jun 30, 2010 9:00 am • linkreport

The WMATa timetables suggest Farragut West to Farragut North takes just 9 minutes. How long do you think it takes to go up two levels from the platforms of one station to the street, walk a block, wait for a walk signal to cross two streets, and then go down two levels from the street to the platform level of the other station? Probably 5-8 minutes. I don't think this going to offer enough time savings for the casual rider to care enough to take advantage of it. Is that a reason not to do it at all? No. But I can easily see why it shouldn't be high on the priority scale.

I just hope when the invisible tunnel is barely used that it doesn't become a political justification to never build the underground tunnel.

by Jason on Jun 30, 2010 9:05 am • linkreport

The invisible tunnel would be a perfect complement to Metro's invisible customer service, invisible safety plans, invisible improvements, and invisible accountability.

by Fritz on Jun 30, 2010 9:12 am • linkreport

@Jason: It could quite easily make a bigger difference during off-peak hours. If a rider knows that a train is coming (via a smartphone, most likely) and that he'll miss it if he goes all the way to Metro Center, that could save a good 15 or 20 minutes on weekends.

by Tim on Jun 30, 2010 9:48 am • linkreport

@Michael Perkins re: Vienna/Metro Center scheme:
But can't they do that already? You pay the minimum fare when exiting the same station you entered, right ? So under your scheme you could get from one station to any other for the minimum fare (and there may not be a time limit on this, I've never tried).

@Matt Johnson: re: Farragut West
I understand why they would only have one of the exits open, but why they 17th St one as opposed to the one on the square (which is closer to more buses)? Could this be changed if they get the invisible tunnel?

by Steven Yates on Jun 30, 2010 9:56 am • linkreport

I equate an "invisible tunnel" with BRT (there is nothing rapid about bus transit in most cases) and the Alexandria old Town "Trolley" (It's a friggin' bus, NOT a trolley).

But, my strong opinions aside, it would provide benefit just like BRT and the Old Town "trolley". An "invisible tunnel" would allow for the collection of real data on which better decisions can be made as to whether the large capital investment should be made for a real tunnel.

by EZ on Jun 30, 2010 9:58 am • linkreport

@Steven Yates:
I don't think any entrances should be closed at any time, but I understand that Metro wants to save money, and I'd take a closed entrance over service cuts or a major fare increase.

Irregardless, that's the decision that Metro has made. If Farragut West is going to have only one entrance open, it must be the 18th Street entrance. That's because that's where the elevator is.

If they opened the 17th Street (Square) entrance and closed the 18th Street entrance (because they can only have one open without increasing the cost to run the station), it would mean that mobility impaired riders would be unable to use the station.

by Matt Johnson on Jun 30, 2010 10:25 am • linkreport

I think the invisible tunnel is only useful if they speed up how long it takes to exit and enter a station. The escalators should be turned up to regular speed instead of the useless snails pace they operate at. There should also be clear indications on the pavement for where to walk to so people can actually get to the other station entrance.

by James on Jun 30, 2010 12:43 pm • linkreport

Metro Center <---> Gallery Place would be sweet and would create a "super station" connecting all the major lines. I know it would be expensive, but the rents for all of that guaranteed traffic that could be charged would be a huge win, too. Lets get some private partnership rolling here. This wouldn't require any major re-factoring of the fare system either.

by Eric on Jul 2, 2010 12:11 am • linkreport

Add a Comment

Name: (will be displayed on the comments page)

Email: (must be your real address, but will be kept private)

URL: (optional, will be displayed)

You can use some HTML, like <blockquote>quoting another comment</blockquote>, <i>italics</i>, and <a href="http://url_here">hyperlinks</a>. More here.

Your comment:

By submitting a comment, you agree to abide by our comment policy.
Notify me of followup comments via email. (You can also subscribe without commenting.)
Save my name and email address on this computer so I don't have to enter it next time, and so I don't have to answer the anti-spam map challenge question in the future.


Support Us