This is probably off-topic, but are they using different colors for male and female escalator users? Does it matter? They already have different graphics, but for some reason felt the need to make it really clear that there are different gendered people here, riding escalators? Very strange.
Also, it doesn't seem to address the "total time for the ride", the minimization of which is of course the reason why people want to walk in escalators. That's the egoistical reason, to minimize overall throughput is perhaps somewhat less interesting to each individual.
That said, optimizing the system as a whole for throughput is a good goal so perhaps it's time to change escalator culture!
Optimize for latency of one class vs. optimize for total throughput... Both are valid choices in different contexts... Welcome to network QOS management !
From the foot of the elevator people walking will reach the top faster. However, navigating from the back of the queue at peak time to the start of the escalator is often time consuming.
It would be interesting what the times were between those navigating the queue and walking up vs during the trial where everyone stands but the queue to the escalator is shorter.
This. Systems Thinking teaches us that optimising a small part of the system (walking up the escalators) can have unintended consequences (bottlenecks/queues elsewhere) that result in the time through the whole system to be slower.
Apparently you haven't been on the London Underground at rush hour. The escalator halls aren't long enough for an orderly queue to form on one side when a tube offloads its passengers.
The original problem they are trying to solve is a big crowd of waiting people develops at the bottom of the escalator, and since London Underground passageways are quite narrow this blocks back until the platform gets overcrowded, which is unsafe.
Then don't you think they should encourage everyone to keep moving instead of being lazy and standing there. Escalators, like moving walkways, are designed to move people faster, not for people to just stand there like sheep.
Do they have a parallel set of stairs next to them? That usually seems to limit crowding, since the bigger the queue for the escalator gets, the more likely newcomers are to just take the stairs instead of join the escalator queue.
Given the age of the system in many stations there is room for only two escalators. Further older stations are very deep resulting in extremely long sets of stairs.
At some busy stations (e.g. Canary Wharf during the morning rush hour) you'll find a big cluster of people at the bottom waiting for the standing side while there's space on the walking side. Obv. it gets dangerous if a trainload of people can't leave the station before the next trainload arrives. (Or more cynically, if they can achieve the same throughput with fewer escalators they'll save money on escalators).
Yeah, but it's only dangerous because those people choose to stand there waiting instead of using their legs. It's their choice to be lazy and those create a danger for everyone else. We need to stop rewarding people who can't do anything for themselves.
As a user of the system, I am not interested in throughput, I am interested in latency. Standing-only might have better throughput, but it reduces latency for the users who hurry and are willing to walk.
Very common places and times, actually. The tube, like everyone else, builds capacity for peak load. That means there almost certainly will be periods every day when the system is at or near capacity. That's when decreased throughput turns into increased latency via queuing. The only time that's not true is when the system has infinite queues or never approaches peak load, neither of which is a very useful assumption.
> That's when decreased throughput turns into increased latency via queuing.
Only for the people who stand, not for the people who walk, because the people who stand and queue up leave passage for the people who walk to pass through.
I can't comment on arbitrary metro systems around the world, but where I live this always happens. When demand is high, people who stand queue up for the escalator, but there is always enough space for people who walk such they are unimpeded.
Very rarely demand is so high that people queue up for the trains themselves (only saw this twice in my life, I think), but even then, there isn't any congestion on the stairs.
In my experience - many systems all over the world, though none as a regular commute since 1995 - the clot of people queuing to stand often severely impacts ability to reach the walking side. An even bigger impact is people who aren't good about standing all the way to the right, impeding progress for the walkers on the left. This phenomenon seems to get worse as queues get longer, though that might not be the case in London with its notoriously strong social norms. In both cases, though, a lack of sufficient throughput does impede walkers as well.
You can get past onto the walking side - the standing side hits capacity first and so only people who aren't in a hurry are delayed. (I see this every morning at Canary Wharf)
Exactly. In my life I have never saw a congestion on the escalator. At rush hour, poeple who stand, queue up, sure, but they never queue up enough in order to block the passage of people who want to walk.
Even on New Year's Eve, where demand is highest and the trains are overloaded, I haven't seen congestion on the stairs.
They're also assuming people are willing to stand tightly packed, rather than one on every second stair. I've never seen an escalator, no matter how dense the crowds, with strangers standing on the same side of adjacent steps.
Not quite, because this is based on an actual real-life test.
If people are currently leaving a single spare step when standing, but the people walking have to leave more than one spare step (or because there aren't so many, more than one step is left) then there's spare capacity that's not being used.
As the article clearly stated, they did leave people a choice. Some of the tube stations have more than a dozen escalators, and only some were changed to standing-only. It makes perfect sense to optimize for throughput generally, but leave some lanes for those relatively few people who need/want to walk (TFA also discusses the difference in this willingness for escalators of different lengths).
In most places it's not a question of throughput, it's a question of allowing those that are really in a hurry to go faster, so it's not like societal normals should change due to this study conclusions.
The more people willing to walk the lower that % becomes.
You could probably sway a lot more people to walk instead of stand by pointing out (truthfully) that it's an easy habit to get into to promote fitness.
It would be a good use for spare advertising space not being used to encourage tourists to go see Wicked or Mamma Mia.
I was just about to comment: And so far, not one mention of the health benefit of walking (well, actually, ascending stairs) vs standing still. Then saw that you had it covered. :-)
Indeed. Or just get rid of escalators entirely. Stairs are much cheaper and can accomodate more people. And maybe it'll dissuade people from running around with more luggage than they can carry.
Lifts are of course perfectly acceptable for the diagonally challenged (wheelchairs, prams, whatnot).
I'm guessing you haven't used London's tube much :)
Many of the stations are very deep - the equivalent of a 15 storey building. Changing to stairs would certainly lower throughput because that's a really long climb.
From [1], "Hampstead is the deepest station below the surface, at 58.5 metres (192 ft), as its surface building is near the top of a hill". So the deepest is about 60m, but from browsing around, it seems many are at least 40m deep.
This may be true for TFL, but like they say, the results depend on people's willingness to walk, and desire for personal space.
At peak hours/stations in the Seoul subway system, standing only would be a disaster. There are more than enough people willing to walk to completely fill the left side of the escalator. Everyone keeps up a steady walking pace while being essentially 1 person per step, so I'd guess it's at least a 30% improvement in throughput.
I think this estimate is incorrect. The article demonstrated otherwise. Also, people are willing to pack closer together when standing than when walking (and when walking than running!)
Walking does increase the overall speed, but the study shows it's not enough for the increased space required. At least for British personal space requirements.
Similarly, when driving, you need a certain amount of time between each vehicle. Yes, you can pack in with <1m space between bumpers in a stopped traffic jam, but you need many vehicle lengths to be safe at speed.
For each mode of transport, there is an optimal speed and packing density. But you can't change one (speed) without affecting the other!
For this to be true, everyone would have to be perfectly synchronized, up to which foot they were currently lifting, so that the foot could go precisely into the space left by the person ahead, landing next to their other foot. If this is what's done, I'd like to see a video of it.
From the late Mitch Hedberg. "An escalator can never break: it can only become stairs. You should never see an Escalator Temporarily Out Of Order sign, just Escalator Temporarily Stairs. Sorry for the convenience."
I find it interesting that most pieces on the topic consider only the throughput perspective while game theory seems more appropriate. People aren't homogenous "units of production/demand" as queueing theory suggests, there is a mix of incentives at play. Some people are OK with waiting in the queue and on the standing escalator, while others are in a hurry and really need a way to go faster. It is surprising that almost no one complains about dedicated mass transit lines on our roads ("we can increase vehicle throughput by removing them!"), yet most people (people in charge, too) apply similar arguments to escalator usage.
Are they serious? It's faster for who exactly? The people who were already standing instead of walking? Why should the people who are walking care, they want to get somewhere faster so they walk, if people want to be lazy and just stand there then they can wait, that's the choice they made. This entire article and the time wasted on it is idiotic.
60 comments
[ 4.5 ms ] story [ 35.3 ms ] threadAlso, it doesn't seem to address the "total time for the ride", the minimization of which is of course the reason why people want to walk in escalators. That's the egoistical reason, to minimize overall throughput is perhaps somewhat less interesting to each individual.
That said, optimizing the system as a whole for throughput is a good goal so perhaps it's time to change escalator culture!
There should be three classes:
1. Wants to stand 2. Wants to walk 3. Wants to stand but will walk if lack of standing space.
These might be in their simulation but it's not well explained.
But the simulation isn't the main point, there was a real life study!
It would be interesting what the times were between those navigating the queue and walking up vs during the trial where everyone stands but the queue to the escalator is shorter.
Only for the people who stand, not for the people who walk, because the people who stand and queue up leave passage for the people who walk to pass through.
I can't comment on arbitrary metro systems around the world, but where I live this always happens. When demand is high, people who stand queue up for the escalator, but there is always enough space for people who walk such they are unimpeded.
Very rarely demand is so high that people queue up for the trains themselves (only saw this twice in my life, I think), but even then, there isn't any congestion on the stairs.
Even on New Year's Eve, where demand is highest and the trains are overloaded, I haven't seen congestion on the stairs.
(Only about 2/3 of the total are shown)
If people are currently leaving a single spare step when standing, but the people walking have to leave more than one spare step (or because there aren't so many, more than one step is left) then there's spare capacity that's not being used.
I'm in favor of building enough escalators to give everyone a choice. In an emergency situation everyone will have to walk anyway.
You could probably sway a lot more people to walk instead of stand by pointing out (truthfully) that it's an easy habit to get into to promote fitness.
It would be a good use for spare advertising space not being used to encourage tourists to go see Wicked or Mamma Mia.
Lifts are of course perfectly acceptable for the diagonally challenged (wheelchairs, prams, whatnot).
Many of the stations are very deep - the equivalent of a 15 storey building. Changing to stairs would certainly lower throughput because that's a really long climb.
I find that incredibly hard to believe. Assuming 4 meters per storey, you are saying that the London underground stations are at a depth of 60 meters.
If true, that's really amazing.
[1]: https://en.wikipedia.org/wiki/London_Underground_infrastruct...
At peak hours/stations in the Seoul subway system, standing only would be a disaster. There are more than enough people willing to walk to completely fill the left side of the escalator. Everyone keeps up a steady walking pace while being essentially 1 person per step, so I'd guess it's at least a 30% improvement in throughput.
I think this estimate is incorrect. The article demonstrated otherwise. Also, people are willing to pack closer together when standing than when walking (and when walking than running!)
Walking does increase the overall speed, but the study shows it's not enough for the increased space required. At least for British personal space requirements.
Similarly, when driving, you need a certain amount of time between each vehicle. Yes, you can pack in with <1m space between bumpers in a stopped traffic jam, but you need many vehicle lengths to be safe at speed.
For each mode of transport, there is an optimal speed and packing density. But you can't change one (speed) without affecting the other!
For this to be true, everyone would have to be perfectly synchronized, up to which foot they were currently lifting, so that the foot could go precisely into the space left by the person ahead, landing next to their other foot. If this is what's done, I'd like to see a video of it.