Traffic models can now predict large-scale jam patterns — but only in simulations

Researchers ran traffic flow equations through computer simulations and found they produce traffic jams that cluster in patterns matching real-world data — specifically, jam sizes follow a mathematical power law rather than random distribution. This suggests the equations physicists use to model traffic actually capture something real about how congestion organizes itself across a city.

For decades, traffic modelers have used continuum equations (differential equations that treat traffic flow like a fluid) to predict congestion. This paper shows those equations aren't missing something fundamental — they actually reproduce the scale-free statistics observed in real traffic data. The catch: this is a numerical confirmation on idealized grid networks, not a deployed tool. What matters is that if the equations work in simulation, they might be trustworthy for predicting where large congestion clusters will form in real cities, which could help city planners identify bottlenecks before they happen.