The recent debate on both the existence and the cause of fractal (Lévy) patterns in animal movement resonates with much deeper and richer problems in movement ecology: (1) establishing mechanistic links between animal behavior and statistical patterns of movement, and (2) understanding what is the role of randomness (stochasticity) in animal motion. Here, the idea of behavioral intermittence is shown to be crucial to establish mechanistic connections between the behavior of organisms and the statistical properties they generate when moving. Attention is drawn to the fact that some random walk modeling procedures can impair the identification of intermittent biological mechanisms which could govern major statistical properties of movement. This fact, together with some misconceptions and prejudices regarding the role of randomness in animal motion may explain why stochastic processes have been disregarded as a potential source of adaptation in animal movement. In the near future, the advances in biotelemetry together with a more explicit consideration of behavioral intermittence, and the development of novel random walk approaches, could help us to set up the bases for a landscape-level behavioral ecology.