Re blocking roads: in this example If there are 9000 cars instead of 4000, the new road doesn't make any difference, and if there are more than 9000, the new road helps. With 10000 it brings commutes from 95 minutes to 90. With 20000 it brings them from 145 to 90.
And if there are 3000 cars, the original commute is 60 minutes and it remains 60; fewer than 3000, the new road is helpful.
As I understand it, this is true in general. If you add a road, it might slow things down within a certain level of demand (here, 3000-9000 cars), but outside that it will be either helpful or indifferent. Cite, but it's not open access and I don't remember how I got hold of the full paper.
This is a new experimental weekly thread.
Guidelines: Top-level comments here should be links to things written by members of the rationalist community, preferably that would be interesting specifically to this community. Self-promotion is totally fine. Including a very brief summary or excerpt is great, but not required. Generally stick to one link per top-level comment. Recent links are preferred.
Rule: Do not link to anyone who does not want to be linked to. In particular, Scott Alexander has asked people not to link to specific posts on his tumblr. As far as I know he's never rescinded that. Do not link to posts on his tumblr.