Parable of the Polygons

[clyde]

I can't quite wrap my head around it yet, but it seems like an interesting demonstration of one part of the way society organizes itself. It's basically a simulation of a self-organizing group based on a rigid definition and requirement of homogenity (and later diversity) for the bordering neighbors of each autonomous unit. It's certainly an over-simplification, but the demonstration does get one thinking about how we self-organize based on our similarities and differences. I've heard that sponges do something similar, but far more extreme. Their expectations for homogeneity is much higher than most of ours. 

 

http://ncase.me/polygons/

 

[Merus]

I love simple mathematical models like this, because I believe, generally, that the 'simplification' allows us to understand the dynamics of a more complex situation. We can make segregation happen in a model with one parameter.

 

There's another simple mathematical model I love where by adjusting the distribution of the units in the grid, they were able to make units that would try to flip themselves to be unlike their neighbours to always follow their neighbours, which they called the 'hipster' model.

 

I feel like showing system dynamics through simple rules and gameplay systems are something games can do really well, and should be doing a lot more of. I think little racist triangles and squares are more illuminating than stories about an individual person's struggles with racial bias, because plenty of people will write off one person's experience in a way they can't when they've had to internalise the behaviour of racist shapes.

[clyde]

I like thinking about it not only in terms of racial diversity, but also how I like to surround myself with people that enjoy art.

[Gwardinen]

Although it's worth noting that when you get to the final sandbox mode, if you make one group significantly larger than the other (say, to reflect white and non-white populations in western countries), it becomes much harder to resolve to everyone's satisfaction.

 

I'm not at all trying to say this is a bad message, or a poor way to communicate it - actually I think this is a great way to communicate such a thing! - just that there are even more problems than the post itself is able to tackle.

[SuperBiasedMan]

This is fantastic

 

Although it's worth noting that when you get to the final sandbox mode, if you make one group significantly larger than the other (say, to reflect white and non-white populations in western countries), it becomes much harder to resolve to everyone's satisfaction.

 

I'm not at all trying to say this is a bad message, or a poor way to communicate it - actually I think this is a great way to communicate such a thing! - just that there are even more problems than the post itself is able to tackle.

 

and I absolutely agree (tried the same thing) I think this is ALSO why games/interactive media are great at demonstrating the result of a system's behaviour. You can use a procedural system to demonstrate something that you didn't intend but is an equally valid point.

 

This is the kind of thing I'd like to see more of in games and also to put into what I make. I do wonder, would this have more success as a game where it includes this stuff as a theme/plot/message as opposed to directly communicating that you're making this point. I often think that the people who most need to be convinced of these issues are the most likely to turn away when they can see it's explicitly making a statement. But maybe they'd take more from it if they're playing a systemic game about the poor triangle and its relationships with the other shapes.

 

(yeah obviously it still needs to be a well made point)

[tberton]

I quite enjoyed this, although something I think it doesn't really convey well until the end: the point of the simulation isn't to show how people segregate themselves, it's that small biases lead to big biases, that those biases have lasting effects even after they've been eliminated and that biases in the opposite direction are needed to correct those lasting effects. It's visualized as "neighbourhoods" (and the research comes from segregation studies) but it's meant to apply more broadly than that.

[Bjorn]

Although it's worth noting that when you get to the final sandbox mode, if you make one group significantly larger than the other (say, to reflect white and non-white populations in western countries), it becomes much harder to resolve to everyone's satisfaction.

 

I'm not at all trying to say this is a bad message, or a poor way to communicate it - actually I think this is a great way to communicate such a thing! - just that there are even more problems than the post itself is able to tackle.

I immediately made the same change in the sandbox, which I suspect is an intended consequence of the author's. It actually saves them having to outline it for you. You can ask yourself, "Hmmm, here's the demographic of my city, let's see what happens." And be unprepared that it devolves to being unsolvable with real world numbers.

[iax]

If you're interested in agent-based graphical simulation of complex systems, download NetLogo and check the included models and their descriptions https://ccl.northwestern.edu/netlogo/models/index.cgi e.g. you can find the model discussed here under social science / segregation.

 

NetLogo itself is an integrated modeling environment with it's own simple programming language so you can easily modify the models or create your own.

[Dewar]

I guess being a mathematician made me just see the problems with the simulation, such as there are edges which behave differently than the center, but I get the point at least.