Wednesday, 19 June 2013

Collective Motion, Heavy Metal, and exercises for the interested reader

In this video, Professor Moriarty summarises a paper in Physical Review Letters, entitled  "Collective Motion of Moshers at Heavy Metal Concerts". I suggest watching to the end.

The paper in short form: - and more from the researchers:

Abstract: Human collective behavior can vary from calm to panicked depending on social context. Using videos publicly available online, we study the highly energized collective motion of attendees at heavy metal concerts. We find these extreme social gatherings generate similarly extreme behaviors: a disordered gaslike state called a mosh pit and an ordered vortexlike state called a circle pit. Both phenomena are reproduced in flocking simulations demonstrating that human collective behavior is consistent with the predictions of simplified models.

Those of you who can enjoy the mathematics may also enjoy the source code for the simulations, available at github and linked to in the short form article, and anybody can play with the simulation, here.

On reading the reasearchers' summary and looking at the short form paper, I'm not yet convinced that this research really predicts or explains anything, but I'd like to know if it could.

In my view, moshing is a kind of social dance. Obviously you could throw yourself around to heavy metal music all alone, but the true experience, I would have thought, is the one where you get to bounce off other people in the presence of a band.

I have a lot of questions about lots of social dance phenomena that might be answered by research of this kind. People argue about them endlessly, with no conclusions, and that makes me think they must be simpler than they appear to be.

For example:

How simple can a simulation of a (social, progressive, partner) dance be and generate a plausible simulation of a dance floor? How about an orderly line of dance in a specific direction? What variations can you introduce and still have it work? What variations can you introduce and have it fall apart?
I say that lighting - or rather, its effect on the ability of dancers to accurately estimate the distance and velocity of other dancers - matters. Can we support or falsify that with evidence?
I also say that layout matters. Hard or soft edges to the dancefloor matter. Irregular corners matter. A dancefloor that is too big in relation to certain characteristics of the dancers, will always be disorderly. Can we model, support, or falsify any of this?

I say that flocking behaviour matters a lot: and that, if everyone does it just enough, you still get order, but if some people do it not at all, others have to do it more to create order. I say that this interacts with the size of the floor compared to the number of dancers. Can that be shown?

What rules or conditions do you actually need, and what rules are unnecessary? That seems to me like it should be an answerable question.

It would also be very straightforward to Dance Your PhD afterwards.


C said...

Ok in this order and fairly quickly change these settings to max

Flock strength
Fraction Red
Box Size

If you imagine you're looking down from the ceiling at one of the corners of Eton, with the black area representing the mental boundaries of the lanes and the clumps of red balls representing the dancers' movement. (It's takes about 30 secs for the outer lane to get going)

I think Flocking represents how contained the dancers are. I think Damping represents how much they care / are aware of the lanes.

Now change Flocking and damping to minimum - instant Negracha!

msHedgehog said...

What values are you using as maxima? In Firefox I get number fields, not sliders, and in IE it doesn't even run.

C said...

That's strange in Firefox I get sliders.

Flock strength - 5
Damping - 5
Fraction Red - 1
Box Size - 100

Flock strength - 0
Damping - 0

msHedgehog said...

If I set Fraction Red to 1, there are no black areas because all the balls are red. And there's no order of settings because none of them apply until you click restart. Which initial condition did you start from?

C said...

Um I get a real time change when I change the other settings without having to click restart. Try changing the box size? Hmm actually restart completely mucks it up. The 100% red might be a 3 card trick

The initial values are
Soft sphere 100
Flock strength 0.55
Noise strength 0
Speed 1
Damping 1
Box Size 28.2
Delta t 0.1
Frameskip 2
Fraction red 0.15
Colour scale 25

msHedgehog said...

Hmm - fraction red = 1 should make them all red. But if I change them slowly in the order you said, it doesn't redistribute the balls when you change the box size, and it ignores the fraction red, so it ends up with the black balls all in one corner, simulating a lane between them and the wall. It seems like a bug rather than a product of applying the settings. There seem to be many ways of getting chaos, though :)