Parable of the Polygons
December 8, 2014 9:13 AM   Subscribe

A playable blog post about triangles, squares, bias, and society from Vi Hart and Nicky Case.
posted by cthuljew (22 comments total) 53 users marked this as a favorite

Spoiler: A tiny preference for diversity keeps segregation from skyrocketing, even with fairly high preferences not to be a minority in one's neighborhood, but lots of shapes are unhappy/unsatisfied, especially if you start with one shape in the minority.
posted by straight at 9:31 AM on December 8, 2014 [1 favorite]

I noticed also that it's very hard to make anyone in the minority happy when there is less empty space on the board. Pretty obvious metaphor there.
posted by cthuljew at 9:33 AM on December 8, 2014

So in the "I'll move if <10% or >80% of my neighbors are like me" case, the desire not to be in a completely homogeneous neighborhood allows diversity to spread until the segregation level is 0. But notice what else has to be done -- the shapes have to accept being a minority themselves. With a 10% threshold, a triangle only needs one triangle neighbor to be happy even with seven square neighbors. Earlier, the threshold was 33%, so a triangle would need three triangle neighbors and five squares.

When I run a simulation of "I'll move if <30% or >80% of my neighbors are like me", segregation hovers at around 15%, and shapes are constantly moving -- the desire to not be a minority conflicts with the desire for diversity. A square might want to gain one triangle neighbor, but that triangle would then be unhappy surrounded by squares.

Maybe high segregation is acceptable as long as the triangles and squares aren't killing each other? But that might not be possible either...
posted by Rangi at 9:37 AM on December 8, 2014 [2 favorites]

I modeled my city: 69% "square", 5% vacancy rates . Mostly what I'm getting out of the simulation is that the rate of segregation will never change.
posted by maryr at 9:51 AM on December 8, 2014

In one of the demo's I had unhappy squares but no place to put them that would make them happy. So I thought "I just need to move some of these already happy triangles to make some room". Boom, ethnic cleansing.
posted by benito.strauss at 9:52 AM on December 8, 2014 [3 favorites]

(You may want to try a new board a few times - I'm finding if I start low at those rates I stay low, if I start high I stay high.)
posted by maryr at 9:53 AM on December 8, 2014

Fascinatingly, it does seem possibly to make all of the minority happy while the majority still flits around unhappily. Huh! (I don't think the simulation allows anyone to be happy completely surrounded by their own shape)
posted by maryr at 10:00 AM on December 8, 2014

This was really great. Lebensraum was never so satisfying
posted by liliillliil at 10:52 AM on December 8, 2014 [1 favorite]

I think this is an allegory for MetaFilter moderation.
posted by GenjiandProust at 11:22 AM on December 8, 2014 [1 favorite]

I'd like just one more slider - the ability to adjust triangles' and squares' preferences separately. What happens if triangles only like to be around other triangles, but squares like more diversity?
posted by wanderingmind at 11:48 AM on December 8, 2014 [5 favorites]

I also would want wanderingmind's extra slider...
posted by subversiveasset at 11:54 AM on December 8, 2014

I would like to draw attention to the ecstatic green pentagon at the bottom of the page.
posted by evidenceofabsence at 11:56 AM on December 8, 2014 [3 favorites]

This post and the one after it about Dixiecrats have some great synergy
posted by boo_radley at 12:35 PM on December 8, 2014 [1 favorite]

This is pretty great, though it's hard to throw in the economic cost of moving once you're already established somewhere. But it's cute and interesting!
posted by corb at 12:42 PM on December 8, 2014

Would be interesting if the sliders adjusted a bell curve of the population, instead of everyone being exactly equal. That would be more realistic.

In a real society, you're going to have a few people more shapist, a few people who value diversity more, and the majority hovering in between. I wonder in this scenario if the shapists would create little enclaves, or if the diversifists would be able to win them out.

To take this to a whole other level, you could have the shapes preferences individually modified based on their experiences. If they're in a homogenous area, they tend to become more shapist. If they're in a diverse area, they become more diverse. Every time they have to move seeking homogeneity, they become more shapist. Every time they have to move seeking diversity, they become more diversifist.
posted by brenton at 1:24 PM on December 8, 2014 [1 favorite]

For the single slider demo, you can also remove segregation by making the polygons EXTRA shapist. Above about 81%, so many of them are unhappy that they'll just keep mixing around and segregation trends downward. Of course, that's a VERY mobile society, and everyone is miserable, so there's that...
posted by solotoro at 2:34 PM on December 8, 2014 [1 favorite]

This definitely got me thinking about my own behaviour. I'm a triangle and I currently live a neighbourhood where, of my nearest 9 neighbours, I'm pretty sure only one household in non-triangle.

I would definitely rather not live in a totally homogenized neighbourhood, but if that one non-triangle family moved out and was replaced by a triangle family, would I consider that an impetus to move? Probably not?

Similarly, about 5 years ago I lived in a neighbourhood for several years that was majority square, probably about 70% square, with most of the rest being triangle. I definitely preferred living in that mix to my current one, culturally. But, at the same time, when I walked just three blocks north, I would find myself in a 99%-100% square neighbourhood. As much as I would like to say I would be happy living there, I never felt welcome when I was walking through or shopping in that neighbourhood. If I found myself living there, I may well be uncomfortable enough to move.

So, I mean, I guess it shows that my preferences are somewhere around less than 99% triangle and greater than 20% triangle. That those two numbers don't add up to 100% sort of suggests that I am more shapist than I thought.

Interestingly, if those preferences are universal, they only really seem to work out in the simulation if there are a pretty even mix of triangles and squares.
posted by 256 at 2:40 PM on December 8, 2014 [1 favorite]

(Is it weird that I automatically saw "squares" as the obvious majority?)
posted by maryr at 3:09 PM on December 8, 2014 [1 favorite]

Curiously, if you have a finite box (as in the post), and you have >90% bigotry, a perpetual cycle of white flight occurs, leading to 1) everyone's moving all the time and 2) the segregation rate goes to 0. So sprawl helps to keep racism alive, I guess. Which, come to think of it, you know what happened in the South in the civil rights era?
posted by Maecenas at 5:02 PM on December 8, 2014 [1 favorite]

Weird, mine made a perfect map of Milwaukee.
posted by buriednexttoyou at 5:54 PM on December 8, 2014 [3 favorites]

Am I right that the system only counts six neighbors, and that for some reason, the up and down squares don't get regarded as neighbors?
posted by Conrad Cornelius o'Donald o'Dell at 9:46 PM on December 8, 2014

Curiously, if you have a finite box (as in the post), and you have >90% bigotry, a perpetual cycle of white flight occurs, leading to 1) everyone's moving all the time and 2) the segregation rate goes to 0.

But this is just an artifact of the random as opposed to "intelligent" movement of the shapes in the simulation's algorithm. Clearly it is straightforward to completely segregate the two shapes a move at a time; but the random jumps of a random shape at a time to any suitable space prevents the simulation from making any highly segregated "progress". I think.
posted by sylvanshine at 12:21 PM on December 9, 2014 [1 favorite]

« Older It's too early on a Monday morning for this hot...   |   LOST CAUSE: Dems, It’s Time to Dump Dixie Newer »

This thread has been archived and is closed to new comments