Monday, May 5, 2014

Why does going to Amherst require a transfer?

This past weekend, I was trying to get from Hartford to Amherst. This should be a reasonably easy trip. Peter Pan Bus runs about a dozen trips daily between Hartford and Springfield and eight between Springfield and Amherst. I went looking for tickets, figuring that at least some of those trips would be through-runs. All required a transfer in Springfield. Which is odd, because to get from Boston to Amherst—except for Fridays and Sundays during the academic year—also requires a transfer.

It got me to thinking: can you get to Amherst by bus from anywhere further than Springfield without a transfer at the decrepit-if-soon-to-be-replaced Springfield bus station? It took some reverse engineering of the bus schedules (update: somewhat available in PDF here), but I am pretty sure the answer is a resounding "no."

This is a problem. It is an issue both in the time it takes to get from Boston to Amherst with a transfer—generally more than three hours for a drive which would be a direct trip of 1:45—and the psychological effect that people do not like to have to transfer (especially if they have to sit in a post-apocalyptic bus station that hasn't seen a broom since the mid-'70s). There is research about transfer penalties in transit, and it is likely that this carries over to intercity bus travel as well.

Amherst to Boston should be a strong bus market, even at times when undergraduate students aren't decamping for home on the weekend. Amherst to New York City should also be a well-traveled route. In theory, one bus could start in Amherst and run to Springfield an on to Boston in 2:00, with a transfer available to Hartford and New York. In practice, there is minimal schedule coordination, and every passenger is required to get off of one bus, wait in the bus terminal, and get on to another. By imposing a penalty which doubles the travel time (to Boston) and requires a transfer at a substandard bus station, it discourages students to use the bus system, and indirectly encourages them to use a private automobile, despite $4 gas and tolls.

I ran some times for these routes on a weekday (I chose May 7). The average transfer time for a Boston-to-Amherst trip is 27 minutes. The average transfer for a New York/Hartford-to-Amherst trip is 40 minutes. Departing Amherst the transfer times are a bit better: 20 minutes to Boston and 28 to Hartford/New York. But it still incurs a significant time penalty, and a significant issue of not having through service.

There is a model for this sort of service: Concord Coach's service from Boston and Logan airport to Portland Maine and beyond. Several times a day, two buses leave Boston and Logan and run to Portland. Since the market beyond Portland is smaller, only one bus is needed to go beyond Portland to Augusta and Bangor. The bus which originates in Boston runs through, and passengers coming from Logan and who wish to go beyond Boston get off one bus and on to the other. The transfers are timed for minimal delays, and the buses will wait for each other if one is running late.

Peter Pan's service could mirror this from Amherst. Buses would leave Boston and run to Springfield, a major market. At the same time, buses from New York to Springfield would be coordinated to arrive a few minutes before the Boston bus. Those passengers would still transfer if they wanted to travel beyond Springfield—and a few minutes of schedule padding could be built in due to traffic conditions south of Hartford (with more frequent service to come on the New Haven-Hartford-Springfield rail corridor, train services could be synched with bus service as well). Boston passengers to Amherst would stay on the bus, and instead of a 10 to 30 minute transfer, it would necessitate a five minute stop.

This would mean that passengers from Boston to Amherst would see significantly shorter trips and would no longer have to move from one bus to another. If Peter Pan claims that there is not enough demand for the market, perhaps it is because their service is substandard on the route. For anyone who has taken a bus to Amherst it's a joke that you have to change in Springfield. In reality, it is a detriment to service.

A major stakeholder in this should be the Commonwealth of Massachusetts. They operate, in Amherst, a major research and educational institution, which should have strong ties to Boston. There's no direct highway access to UMass; driving between the two requires an indirect trip via Springfield and Northampton, or a narrow road from Palmer. The state also, through it's BusPlus+ program, subsidizes Peter Pan routes, buying buses for the operator in exchange for Peter Pan's operation of commuter routes. MassDOT should, as part of this relationship, encourage Peter Pan to run direct service between Amherst and the three largest cities in the state: Boston, Worcester and Springfield. Not doing so is a disservice not only to the traveling public, but also to the state's major public university. It's nonsensical that you can't get on a bus in Amherst and get to Boston. That could—and should—change.


  1. I would argue it would make more sense to have through-service from Boston to Albany, and another service Hartford to Amherst (or Brattleboro), that connect in Springfield. Simplicity is helpful. It could be timed better, as per your example, though. (Well, I'd really like to see the BOS-ALB rail upgraded to not-terrible, and the upgrading SPR-HAR rail continued north, but dreams are cheap.)

  2. There are only three Boston to Albany direct buses per day, and only one of them stops in Springfield. This is not as large a market as Boston-Amherst, because Albany is mostly tied to New York City. Springfield is a bit out of the way for Boston-to-Albany buses, which is probably why several bypass the city.

  3. Wait, what? When I attended UMass Amherst, there were multiple direct runs between Boston and Amherst on any given day. Some stopped in Northampton, Springfield and/or Worcester, but it was pretty rare that I ever had to change buses. We were fond of calling Peter Pan "the bus company that won't grow up", but I'm surprised to see their service levels have gone that far downhill in the 25 years since.

  4. Wow, interesting. I wonder if anyone has any Peter Pan schedules from the '80s. It would be interesting to see. (For what it's worth, Concord Coach has doubled service between Boston and Portland in 10 years; I don't think that's happened with Peter Pan to Amherst.

  5. In the vague remote future when the state has a proper regional rail system, a direct train from Boston and Worcester to Amherst via Palmer would be possible. The connecting track at Palmer actually only connects in that direction and the current Vermonter service has to reverse direction there. In theory it's possible to take the Lake Shore to the Vermonter, but that requires passing Palmer, changing trains in Springfield, backtracking to Palmer, and waiting for the train to reverse there, which is so thoroughly ridiculous that I doubt anyone actually does it.

  6. @Anon, I saw that. In fact, it's only an hour longer to take the LSL to Springfield and change to the Vermonter to Amherst via a 20 mile backtrack to Palmer than to take Peter Pan. It would be very, very interesting if the state looked in to better Amherst-to-Boston rail service, especially since Amherst is slated to lose service when the reroute the Vermonter to the Connecticut River line. Especially with the DMUs the state is planning to buy. One issue is how curvy the main line is west of Worcester. It's never likely to be able to support high speed operation, but even if most could be operated to 79 mph it would beat highway traffic, which has to deal with congestion at the Boston end and narrow, slow roads from Palmer to Amherst. If speeds could be increased to allow for an average of 60 mph (the line is almost entirely grade-separated), it would take 85 minutes to get to Palmer and probably another 30 to get from Palmer to Amherst. That's two-hour service, not much worse than a direct trip by car.

    1. The problem is that the line is quite curvy, so increasing speeds will be hard, and there are some pretty severe restrictions once you get west of Worcester (30 and 40 mph curves), and the overall speed limit is 60 mph. The other problem is that west of Worcester, the line is single track, which probably limits the capacity for new passenger service, but I'm sure a Friday/Sunday only DMU run would be possible as an experiment once the DMUs arrive.

    2. Line speed west of Worcester is 60 mph, but could probably be upgraded. While there are some curves and grades, they don't come anywhere near the severity of the curves west of Springfield on the climb over the Berkshires. There are about 5 miles of restricted curves between Worcester and Palmer, all of them in the 45 to 55 mph range. Check out the track chart: there are a few 3˚ to 4˚ curves, but also a 14 mile section of mostly tangent with no more than 2˚ of curvature (miles 62 to 76). No reason why portions of this couldn't be upgrade to class 4 or 5. West of Springfield is where you get the 4-6˚ curves, many reverse curves and some 1-2% grades. That's what slows service down.

      The Amherst branch has a 55 mph line speed with some restrictions, but not a ton.