Ants show that emergency exits can work better when they’re obstructed | Hacker News Reader

Ants show that emergency exits can work better when they’re obstructed

nautil.us

126 pointsdnetesn21 comments

I read the book about The Station nightclub fire tragedy on a long flight once. The book had a different reason for people clustering at a single exit. It wasn't because of a heard instinct, it's because in an emergency people rush to the exit they used to enter. In the short amount of time to think in an emergency, people do not look around for new emergency exits. They go with what they know: the way they came in. (I believe this is why flight attendants remind people that their closest emergency exit may be behind them.)

In a situation of panic this leads to overuse of the main entrance and underuse of emergency exits. In extreme cases, it can lead to a crush and cause the exit to become inoperable.

Another issue is that emergency exit signs are posted high, but those are the first areas to be obscured by smoke in a fire.

My take away was to keep egress in the back of your mind when in crowds and to make a mental note of emergency exits.

Planning for unreliable alternative exits is one of the fun aspects of cave diving: you virtually always exit by returning to your point of entrance, even if you're nominally closer to another means of egress, since you can't be certain that the "closer" exit is genuinely navigable. When diving a traverse between two points, the best practice involves diving as far as you can one way, leaving a marker, then entering from the other side on a second dive, evaluating your gas situation at the marker, and then either continuing or aborting the dive.

"Fun"

In the Southern Brazil (Kiss) nightclub fire, the security team didn't initially know the fire broke out, and since it's quite normal at nightclubs in Brazil to pay for all one's drinks upon leaving, they didn't want to let anyone out and thus locked the front door. There were other contributing reasons (lack of exit signs, people mistaking the bathrooms for exits, etc) but policy, I believe, was the main reason for a high death toll.

This is very interesting as the Wikipedia entry in English [0] never mentions this part of the accident. They mention the security guards locked the door, but never return to the reasoning, nor if the security guards were also charged with anything.

0: https://en.wikipedia.org/wiki/Kiss_nightclub_fire

The part about the paying upon exit is well-known in Brazil among Brazilians. The Portuguese version of the Wikipedia article you cited says the following:

"Dezoito profissionais que atuavam na Kiss, quando ouvidos pela Polícia Civil, disseram que não receberam treinamento para uso de extintores nem para evacuação em casos de tumultos ou incêndios, além do que os seguranças não tinham equipamentos de comunicação."

The translation of that being: "Eighteen professionals that worked at Kiss, when heard by the Civil Police, said they received no training for the use of fire extinguishers nor for evacuation in cases of stampedes or fires, in addition to the security guards not having communication equipment."

___

The two owners were taken into custody, as they were ultimately responsible for those they employed and the policies/upkeep of the club. The security team was just doing their job, according to club operating norms in Brazil, as they (initially) thought people were trying to leave w/o paying. In no time at all, though, there was a stampede and the situation became uncontrollable.

Another reason, which you briefly touched upon, is the fact that in large crowds, your view is largely obstructed by the people around you. You completely lose orientation and follow the movement/pushing of people as it's obvious there are people exiting that way.

Here's an analysis of crowd dynamics when exits were blocked off at the outdoor 2010 Love Parade festival in Duisburg, North Rhine-Westphalia, Germany

[0] https://www.youtube.com/watch?v=8y73-7lFBNE

Anybody able to find a video of the corner exit with no obstruction? Wish they posted that one, being that they claimed it was the fastest.

Seems logical that the two walls act to funnel people down towards the opening - a 90 degree-wide corner exit is halfway between a 180 degree-wide funnel towards a single door in a wall, and a 0 degree wide corridor that leads to a door.

Likewise the 'obstructed' wall entrance just turns into two corner exits, side-by side (I imagine the scenario doesn't work out well for the individuals caught directly behind the pillar, as the flow past each side of the pillar into each exit zone prevents them from getting into either flow effectively).

I'd bet that if you put another pillar behind each side of the pillar obstructing the exit you'll double the throughput.

And thus it follows that, by repeatedly adding pillars and doubling the number and halving the size/spacing at each step, you can approach infinite throughput!

I can see how 90-degrees corner helps funneling. I'd like to see an experiment with a 50-degrees angle, I'd assume people/ants crush each other, because they think it's fair to remain shoulder-to-shoulder with the same people.

no link, but there was a good article or documentary about 10 years ago about how you can improve throughput of an automatic-opening double door by putting a bollard in the centre. iirc it was over 2x increase.

basically people aim for the centre and people cluster up waiting for them to pass instead of having 2 way travel all the time.

If you are interested in this stuff you should read http://www.elsevier.com/books/stochastic-transport-in-comple...

Also: http://www.thp.uni-koeln.de/~as/ https://www.uni-due.de/zlv/portrait/schreckenberg.php

The TV show Crowd Control did something about this last year (http://channel.nationalgeographic.com/crowd-control/videos/e...)

This is very interesting, but the title is misleading.

The ants don't do anything better than humans.

They just act like panicked humans do, so you can easily test different theories about the best way to build exits.

We switched to the subtitle.

The new title is still misleading (although I'm not asking you to change it, it's the fault of the article).

They don't do better with it obstructed, they do better when there are channels leading to the exit.

I don't doubt you, but in cases like this it is helpful if someone can suggest an accurate, neutral title. We don't have the capacity to absorb all the content well enough to do that in every case.

I don't know why they're republishing this. This same source published the same article last year: http://nautil.us/issue/13/symmetry/want-to-get-out-alive-fol...

Interesting, but old news.

Reply on news.ycombinator.com