Why did London's Millennium Bridge wobble dangerously when crowds walked across it?

Partly settled. Some of it is understood, some is still argued about.

When the bridge opened in 2000, engineers were surprised because pedestrian loads were assumed too weak and unsynchronized to excite a lateral sway; Strogatz, Abrams, McRobie, Eckhardt, and Ott's 2005 Nature paper proposed that once the bridge began even slightly swaying, pedestrians unconsciously adjusted their footfall to the motion for balance, which synchronized their steps and amplified the sway further -- a positive feedback loop modeled with coupled-oscillator (Kuramoto-style) mathematics. In 2021, Belykh and colleagues' Nature Communications paper re-examined the crowd data and argued synchronization of footsteps is a consequence of the wobble rather than its root cause, proposing instead that increased coherence in how each individual pedestrian's two feet load the deck (not mutual synchronization between pedestrians) is the key driver -- so the precise causal mechanism is still being actively debated among specialists even though the qualitative pedestrian-feedback picture is well accepted.

A brand-new engineering marvel was nearly shut down by an emergent synchronization effect between hundreds of strangers' individual footsteps.

Explanations that circulated

• The bridge design itself was simply structurally deficient/underdamped in an ordinary sense, unrelated to pedestrian feedback
• The wobble was caused by resonance with an external source like wind or traffic, not by the walking crowd

References

• Theoretical mechanics: crowd synchrony on the Millennium Bridge, S. H. Strogatz, D. M. Abrams, A. McRobie, B. Eckhardt, E. Ott, Nature, 2005
• Emergence of the London Millennium Bridge instability without synchronisation, I. Belykh, M. Bocian, A. J. Champneys, K. Daley, R. Jeter, J. H. G. Macdonald, A. McRobie, Nature Communications, 2021