Week 10: When the queue is the hazard
A crowd crush is a queueing failure, and the arithmetic that predicts it is the arithmetic from week 3.
10 May 2027
This is the hardest lecture in the course, and it is also the one that has to be here.
Most of the semester treats waiting as a cost measured in minutes and general annoyance, which is fine for almost everything you will meet. There is a region of the state space, though, where that stops being true. When arrivals exceed the rate at which a space can drain and the space has a fixed volume, the queue stops behaving like a line and starts behaving like a fluid, and people die.
This is not a metaphor and it is not a separate subject. Density, flow rate and the drain rate of an exit are the same quantities you have been measuring since week 2. The failure is continuous with everything else you know.
Three failures, one shape#
- Bufferbloat. A network queue grows unboundedly because somebody added memory instead of a limit, and the system becomes unusable while looking healthy on every capacity metric.
- The convoy effect. One slow job at the front of a queue serialises everything behind it, and throughput collapses far below capacity.
- Progressive crowd collapse. Density passes a threshold, individual control of footing is lost, and force propagates through the crowd as a wave. The people at the front cannot move and the people at the back cannot know that.
The third is different in consequence and identical in structure. All three are what a queue does when it is allowed to grow without a bound.

Handled carefully#
We work from published inquiry findings and the crowd science literature, on a named historical case, and we stay with the engineering. This is a lecture about how an authority measures the safe capacity of a space and what happens when the measurement is not made.
If you would rather not be in the room for it, tell Tessa beforehand. There is a written version and it carries the same material.
After this lecture#
You know the region where a longer queue is not a worse experience but a different category of event, and you know the quantity that tells you where the boundary is.
Teaching team
- Amos Redfern, Lecturer