Many commuters want to arrive at a similar time, but a bottleneck can admit only a limited number of vehicles per minute. If everyone departs near the preferred time, a long queue forms. If departures are spread out, the queue shrinks, but some people arrive early or late instead.
The video compares three ways this tension can play out. A concentrated rush produces heavy waiting. A system-level schedule can avoid the physical queue by distributing arrivals, yet individual travelers may still have an incentive to shift their departure. In the final equilibrium, nobody can improve their own travel cost by changing time alone: queuing and schedule penalties balance across the departure window. The cumulative curves and cost plot show why a no-queue plan and a self-enforcing outcome need not coincide. This is a central lesson of bottleneck theory: congestion is partly a competition for desirable arrival times, not just a shortage of road space.
Building an extra road sounds like an unambiguous improvement. Braess's paradox shows why that intuition can fail when drivers independently choose the route that seems fastest to them. A new link changes everyone's incentives, and the resulting route changes can overload roads that previously worked well.
Watch film
Drivers consider the time they personally spend on a road, but each additional driver may slow down many others. That difference between private and collective cost can make the route pattern chosen by individuals less efficient than the pattern that minimizes total travel time.
Watch film
Route choice is learned through experience. If one road becomes slower than another, some travelers switch the next day, changing congestion on both roads. Repeating this adjustment can eventually bring the travel times of used routes close together.
Watch film