4 ms·

Timed transfers are when the trains are timed to arrive at the station at the same time so no additional waiting is required to do a transfer. This allows tran

Timed transfers are when the trains are timed to arrive at the station at the same time so no additional waiting is required to do a transfer. This allows transferring passengers to simply walk off one train and onto a different one located across the platform. They are common on BART where many transfers are timed transfers, but are relatively uncommon on other systems, since timed transfers comes with a large decrease in timetable flexibility which ultimately results in less system responsiveness to loading, since if a busy line has a timed transfer to a less busy line, you have to run trains at the same frequency down both lines. (BART varies the number of cars in response to passenger loading, but has less flexibility to vary the train timing.) It also requires a high on time arrival rate, since the trains need to arrive within about 1-2 minutes most times so neither train has to wait for the other. BART's OTA rate is actually pretty good and most timed transfers have both trains show up within 30-60 seconds of each other. BART only has five different lines though, which helps make this work, and also increases the need somewhat, since transfers aren't otherwise as common.

jrockway 10y ago
Right, BART is somewhat famous for the timed-transfers on other parts of the network, but they don't do it for the direction-changing transfer at San Bruno, if I understand correctly. You wait 20 minutes at SFO, then wait 20 minutes at San Bruno, and then you're in Millbrae where you get to wait an hour for Caltrain ;)
- djcapelis 10y ago
  Those are the maximum waits you'll experience, but yes it is possible to have to wait that long during certain times. If you consider the timetable mechanics carefully, you'll realize why timed-transfers for trains going different directions right at the end of a line are much worse than creating a time table to allow timed transfers elsewhere. You have to completely change turnaround time at the end of the line, idling staff for longer or messing with people's mandated break times, or etc etc etc. OR massively increase service frequency, with the corresponding increase of frequencies on the other lines with timed transfer to the line you've changed frequencies on. SamTrans used to run a more frequent shuttle like BART service between SFO and Millbrae when they were more directly involved in operations cost and setting the BART timetable, but then they reduced their involvement. If San Mateo county reduces its involvement, it isn't incumbent on BART to turn its timetable upside down to try and make the transfers faster to serve people in their county and other counties that don't pay in to BART at a loss to the district and the counties that do pay for an effective transit system. It's not like they can just make a small adjustment to better serve those passengers. And it's not like the passenger volume between SFO<->Millbrae justifies the change. The passenger volume between every single other BART station (other than Millbrae) and SFO is served fine by the current timetable. Making a change here literally only helps people coming from Millbrae.
  - amluto 10y ago
    > And it's not like the passenger volume between SFO<->Millbrae justifies the change. Huh? I've taken Caltrain to Millbrae and Uber to SFO many times. The transfer right now is so bad as to be unusable. Why would you use the current ridership to decide whether making the transfer not suck is reasonable? It's like Sonic.net saying they'll bring me fiber if 50% of my community signs up for 3Mbps DSL first.
    - djcapelis 10y ago
      That connection used to have a much higher service frequency and the traffic didn't support the costs of operating it. I'm not judging by what the traffic is now, I'm judging by what the traffic was when there was a dedicated BART shuttle to serve that exact segment at high frequency. Further, the timetable changes to increase service between that single station and the airport are large, therefore a massive amount of traffic would be needed from one station whereas improvements to existing service benefit people coming from the other 43 stations. So no, it's nothing like your analogy. And, even if the traffic would be similar to have an operating loss ratio similar to the other lines, that still doesn't address the fact that fundamentally it would be to provide service for transit users who live in counties that voted to reduce their BART service and opt out of that transit system. So in your analogy it would be like a Comcast customer complaining that when they use their friend's sonic.net line, the speeds aren't to their liking.

jrockway 10y ago

Right, BART is somewhat famous for the timed-transfers on other parts of the network, but they don't do it for the direction-changing transfer at San Bruno, if I understand correctly. You wait 20 minutes at SFO, then wait 20 minutes at San Bruno, and then you're in Millbrae where you get to wait an hour for Caltrain ;)

djcapelis 10y ago
Those are the maximum waits you'll experience, but yes it is possible to have to wait that long during certain times. If you consider the timetable mechanics carefully, you'll realize why timed-transfers for trains going different directions right at the end of a line are much worse than creating a time table to allow timed transfers elsewhere. You have to completely change turnaround time at the end of the line, idling staff for longer or messing with people's mandated break times, or etc etc etc. OR massively increase service frequency, with the corresponding increase of frequencies on the other lines with timed transfer to the line you've changed frequencies on. SamTrans used to run a more frequent shuttle like BART service between SFO and Millbrae when they were more directly involved in operations cost and setting the BART timetable, but then they reduced their involvement. If San Mateo county reduces its involvement, it isn't incumbent on BART to turn its timetable upside down to try and make the transfers faster to serve people in their county and other counties that don't pay in to BART at a loss to the district and the counties that do pay for an effective transit system. It's not like they can just make a small adjustment to better serve those passengers. And it's not like the passenger volume between SFO<->Millbrae justifies the change. The passenger volume between every single other BART station (other than Millbrae) and SFO is served fine by the current timetable. Making a change here literally only helps people coming from Millbrae.
- amluto 10y ago
  > And it's not like the passenger volume between SFO<->Millbrae justifies the change. Huh? I've taken Caltrain to Millbrae and Uber to SFO many times. The transfer right now is so bad as to be unusable. Why would you use the current ridership to decide whether making the transfer not suck is reasonable? It's like Sonic.net saying they'll bring me fiber if 50% of my community signs up for 3Mbps DSL first.
  - djcapelis 10y ago
    That connection used to have a much higher service frequency and the traffic didn't support the costs of operating it. I'm not judging by what the traffic is now, I'm judging by what the traffic was when there was a dedicated BART shuttle to serve that exact segment at high frequency. Further, the timetable changes to increase service between that single station and the airport are large, therefore a massive amount of traffic would be needed from one station whereas improvements to existing service benefit people coming from the other 43 stations. So no, it's nothing like your analogy. And, even if the traffic would be similar to have an operating loss ratio similar to the other lines, that still doesn't address the fact that fundamentally it would be to provide service for transit users who live in counties that voted to reduce their BART service and opt out of that transit system. So in your analogy it would be like a Comcast customer complaining that when they use their friend's sonic.net line, the speeds aren't to their liking.