The reintroduction of wolves in Yellowstone National Park has been the most-celebrated ecological experiment in history. In 1988, prior to the reintroduction, I was invited to develop a predator-prey simulation anticipating the ecological consequences of wolf recovery. Then following wolf reintroduction in 1995 my students and I documented wolves and elk for 10 years, and now I have validated the model for > 20 yrs of recovery. Features of the structured population model include multi-species functional response reflecting learning by wolves, and territorial density dependence for wolves. Simulations that include stochastic winter severity yield results that bracket observed spatial and temporal dynamics. Wolf recovery has precipitated trophic cascades in riparian vegetation, at first behavioural and then numerical as predation risk has influenced elk distribution and numbers. Yet, outside Yellowstone trophic cascades are overwhelmed by human influences of hunting and livestock.