←back to thread

357 points pyduan | 3 comments | | HN request time: 0s | source
Show context
pyre ◴[] No.8719432[source]
On the third board (the first one with a slider), setting the bias to 100%, segregation stays under 10% (I've run it for a bit and the max was 8%). Obviously we need to tell everyone be act like "I'll move if less than 100% of my neighbours are like me"! ;)
replies(5): >>8719508 #>>8719549 #>>8720076 #>>8720128 #>>8720557 #
lisper ◴[] No.8719549[source]
This is just because when a piece moves, the destination is random. If you allowed the bias to affect the choice of destination and not just the decision to move or stay put you'd get the result you expect.
replies(1): >>8719786 #
normalhuman ◴[] No.8719786[source]
Right, but you can't demand more realism for the results you don't like and accept the conclusions for those you do like.

For example, in a more realistic scenario people move for all sorts of reasons: leaving parent's house, changing jobs, marriage and divorce, etc. This stirring may lead to decreased segregation for slight decreases in bias, unlike what this more static model suggests.

replies(2): >>8719870 #>>8720159 #
1. lisper ◴[] No.8719870{3}[source]
This is not intended to be an accurate model of what actually happens in the real world. It's only supposed to test the hypothesis that minor but widespread individual bias can result in large-scale segregation.
replies(1): >>8719934 #
2. aaronem ◴[] No.8719934[source]
...which is then treated, in the remainder of the text, as though it were an accurate model of what happens in the real world.
replies(1): >>8721930 #
3. vidarh ◴[] No.8721930[source]
No, it is not. You may want to reread it. Including the part where they explicitly state that the real world is a lot more complex, and where they give references to more in depth treatments.