@Glen_b Economists would tell you that the disparity of two queues is a self-correcting problem, but, as is usual, they're premising that statement on theory rather than practice.

(i.e. the observance of a long line will cause people to join the shorter one until equilibrium is obtained)

@GeneralAbrial economic "models" like that always bothered me. they implicitly assume a very specific probability model in order to make sense. see also: the "bidding up/down" interpretation of supply and demand, "price-taking" behavior, etc

@ssdecontrol Nevertheless, such models can be useful insofar as they supply insight. For instance, queuing theory suggests that setting up one queue for multiple servers leads to shorter mean wait times than multiple dedicated queues. That can be (a) tested and (b) modified to account for additional factors (such as people moving more slowly because they aren't used to a single-queue system) in order to become realistic.

@whuber absolutely. it's just the fact that students are taught to ignore the probability models when in reality they're just as important as the decision/utility models. even pointing out that "this queuing model makes testable assumptions about arrivals that we won't cover in this course" is better than just hand-waving over it.

@ssdecontrol That is interesting. In reflecting on it I realized that I actually know little about what "students are taught" in general because I only know what I (try to) teach them!

@whuber out of curiosity, what do you teach? it seems like a lot of the examples you use here deal with natural science and engineering and not economics

@ssdecontrol I teach statistics from time to time :-). Much of my statistical career has involved environmental and geospatial applications, so that's where I tend to draw examples from.