The math is the article is a little sketchy. "But a 2002 study of escalator capacity on the Underground found that on machines such as those at Holborn, with a

The math is the article is a little sketchy. "But a 2002 study of escalator capacity on the Underground found that on machines such as those at Holborn, with a vertical height of 24 metres, only 40% would even contemplate it. By encouraging their preference, TfL effectively halves the capacity of the escalator in question, and creates significantly more crowding below, slowing everyone down. " It doesn't make sense that this would "halve the capacity" unless no one was walking. It seems that with a 40% stated preference for walking, it would be only a 10% loss of capacity due to preference differences. The description should really also include a description of how fast walking is compared to standing. Elevators move at around 0.3 m/s, people walk at about 4x that pace, so even doubling the spacing requirements would shave a fair bit off of that. I think this result is a due to it only being tried on one of three up escalators. By the assumption that there are 6 lanes, devoting 1/3 to walkers and 2/3 to standers can lead to greater efficiency if that more closely matches the actual preference distribution. By maintaining choice, and matching the available options to those desired by passengers one can optimize the results for both those who prefer speed and those who prefer not expending energy.