A partially snow-buried highway sign warning of avalanches in Loveland Pass, Colorado (Image: Kimball Andrew Schmidt / Jupiter)
THE seeming impossibility of predicting the next big avalanche or earthquake has often been blamed on the inherent unpredictability of complex systems. A unique experiment suggests that this idea may be wrong. Accurate prediction may just be a matter of analysing the most relevant information.
Phenomena as diverse as earthquakes, stock market collapses and avalanches follow so-called power-law distributions. Take the intensity of earthquakes: there are very few large tremblors and many small ones.
It is now known that even a minor…
