Science tackles the Red Line and why buses travel in packs

Adam, thanks for posting this.

This actually looks like a pretty interesting read for the weekend. Let's hope that the T's operations people think so as well.

"Equal headway instability" has a nice ring to it, too. So nice, in fact, that I fully expect that UHers far more creative than I am will have loads of fun using the phrase in double entendres of outstanding quality.

It doesn't look like they took much effort in their model. It's a good start, but including passenger needs (like certain stations have higher off-rates/on-rates than others, which causes part of the problem) would give them a more accurate picture of how to fix the problems. Also, one of their conclusions for maintaining headway was to send trains even if people are still standing on the platform trying to get on. That's just not going to fly for all sorts of reasons from safety to human nature.

One solution I've always figured should be put into play is *intentionally* expressing some trains. By now, the T should have a good understanding of which stops have more people than others at what times of day/week, which stops most people want to get to, and which trains often end up being the cause of headway instability because of regularly occurring delay-causing events like the ones identified in this paper (like heavy passenger load, etc).

So, if Train A is maintaining headway from BC to Blandford and Train B is starting 10 minutes behind it at BC, but would normally get slowed down and lose headway to Train A due to passenger load. Instead of then having to express from Packard's Corner to BU Central when Train C finally catches up, why not start Train B at 12 minutes and tell it to only hit major stops, like Chestnut Hill Ave, Washington St, Harvard, Packard's Corner, BU Central, then Blandford. With the pausing inherent at certain intersections, like the road crossing at Warren and Packard's Corner, it'll never actually catch up to Train A but by missing half of the stops intermittently it'll still service a large chunk of people with Train C to pick up the rest (starting with a reduced headway of say 7-8 minutes at BC).

Instead of normal service for 1/2 the line, skipping 5 stops in a row to "catch up", and then hitting the last 2-3 stops before going underground, this would make more sense, and maintain headways easier and more regularly for all passengers to understand.

This study is like an economic study - it reads well on paper, but will never work because people are people and don't behave like models.

Effectively the problem is everyone will pack on the first vehicle to arrive, slowing it down, and thus slowing every vehicle down. skip-stopping or express trains don't solve the problem: They just re-locate the problem, with the added issue they you are really denying service to people (while some people would get 'better' service, the skipped people get no service)

The solution on paper is to 'shut the doors' after X seconds, even if you strand people at the station. Vehicles are no longer bunched up and you stay on schedule. Trouble is a) people won't like that - don't forget these are the people who refuse to let passengers OUT before they board - and now you want them to not board at all??). And b) you have to have a good system. What good is it to closed the doors on people when the next train/bus is really 5-10 minutes out??

A better solution is to have an active map at every station. If the platform is packed, and you can SEE on a map the next train is 60 seconds away, a lot of people will sit back and wait on their own.

Science is all fine and well for explaining the Red Line. But I'm pretty sure the Green Line can only be explained by a malevolent higher power.