Skip to content Skip to navigation

For A Faster Commute, Stand (Don't Walk) On The Escalator, Research Suggests

For A Faster Commute, Stand (Don't Walk) On The Escalator, Research Suggests

Now, how to convince commuters?

Friday, August 26, 2016 - 15:30

Joel Shurkin, Contributor

(Inside Science) -- British researchers ran an experiment that found that more people can move faster on a subway escalator if everyone stopped trying to walk to the top.

Yet implementing the strategy could be tricky -- riders in London mutinied when transportation authorities tested a similar idea on a real subway escalator last spring.

In a blog post from April, two British researchers -- Lukas Dobrovsky and Shivam Desai -- wondered if it was possible to ease the crowding at the bottom of long subway escalators.

The researchers worked at the London office of Capgemini, a worldwide consulting firm that specializes in operations research, a branch of science allied with physics and mathematics that helps to find the best solutions to complicated decision-making problems. Studying queues is a common part of operations research.

Queues can get notoriously bad on the London subway system, known informally as "the Tube." It is the oldest system of its kind in the world, dating to 1863. Until 2012, it was also the largest, running 250 miles. (Beijing now tops it with more than 300 miles of track.) The Tube moves 1.3 billion people a year. On Dec. 4, 2015, during the hustle of the holiday season, 4,821,000 passengers used the Tube and congestion reigned.

Because many of the subway lines cross each other, some lines were dug under others, so that many of the stations are among the deepest in the world, now accessible only by escalators and elevators. The deepest is Angel station, constructed in 1901, with escalators 90 feet long.

Jams at the bottom of the long escalators in these stations are common.

The custom in Britain and in similar U.S. systems is that escalator users stand on the right and walk on the left. It is often different in other countries, but the custom is so entrenched in Britain, that according to The Guardian, those who violate the custom and stand to the left are sure to get a "tut, tut, tut" whispered in their ears from behind.

To gather hard data on escalator habits, Dobrovsky and Desai went to the Tube's Green Park station to observe the passengers. They stood around at the bottom of the escalators or rode up themselves, and timed trips. They found that those standing on the right stood at every other step while those walking on the left used every third step. In other words, the walkers used more space than the standers.

The people walking got to the top in 26 seconds; the standers took 40 seconds. If the walkers could be convinced to stand still while they rode up, the researchers thought, there would be a third more room for riders on the steps and the queue at the bottom of the escalator would be shorter.

The majority would have a quicker trip. That might ease congestion, they wrote in their blog.

The Capgemini researchers plugged their data into Simul8, an off-the-shelf simulation program, and ran several scenarios. They found that if the walkers stood still on the escalator everyone could get to the top in a total time of just under a minute at this particular station. While the walkers would have an extra 13 seconds added to their trip up, a faster-moving queue would help save time before boarding the escalator -- as much as 79 seconds, the researchers found.

Meanwhile, an executive of Transport for London (TfL), the government agency that runs the Underground, noticed that in Hong Kong people there did not walk up escalators and it seemed more efficient. Around this time, the research at Capgemini was published on the Capgemini blog.

TfL decided to try an experiment at Holborn station in the spring. Holborn is where the Central and Piccadilly lines intersect, and has escalators just under 80 feet long, the second longest in the system. Signs were put up telling every to stand still on the escalator.

It didn't go well.

"People refused to stand as told," said Dobrovsky. "People didn't like being told how to walk up the escalator." Social media erupted negatively.

The experiment ran for three weeks. TfL thought first of making it a fined offense to walk up the escalator, but that idea was quickly abandoned. They resorted to uniformed employees with "loudhailers" standing at the bottom of the escalator cheerfully asking people not to walk. They also tried employees in plain clothes deliberately blocking the left, even to the point of holding hands to make an obstacle for insistent walkers.

They found that when commuters complied, the concept actually worked, and the escalator was able to handle more people in the same amount of time.

As a compromise, one of the three escalators at Holborn was left to people who insist on walking, the other two banned walkers.

TfL stopped the experiment after a time and everything went back to normal, walkers and standers. They have not decided whether to try again. Richard Charles Larson, director of the Center for Engineering Systems Fundamentals at the Massachusetts Institute of Technology and an expert on queues, said he thought at first the research was silly, but then realized there was merit.

"The math seems right," he said.

The plan might work better if it was limited to times when it was needed, he said. Like streets that turn one way or change the direction of lanes during rush hour, if the escalators only used the walk-only regime during rush periods, people might be more willing to conform.

"Just make it time dependent. It would be much less of an issue."


1 of 5 -- This slideshow illustrates the results of Dobrovsky and Desai's simulation. Graphics by Abigail Malate, Staff Illustrator


Authorized news sources may reproduce our content. Find out more about how that works. © American Institute of Physics

Author Bio & Story Archive

Joel Shurkin, photo by Abigail Dunlap

Joel Shurkin is a freelance writer in Baltimore who has also taught journalism and science writing.