Newton’s cradle, a device consisting of a chain of steel balls suspended in alignment, has been used extensively in physics teaching to demonstrate the principles of conservation of momentum and kinetic energy in elastic collisions. The apparent simplicity of the device allows one to test commonly hold views regarding the intuitive understanding of physics by lay people. We present and discuss the results of two experiments wherein the extension of the chain, the height of release (experiment 1) and the material of the balls (experiment 2) were systematically varied in graphical depictions of the Newton’s cradle. Participants had to estimate the height that the last ball in the chain would reach if a collision took place. The outcomes revealed a sophisticated cognitive model wherein the magnitude of the displacement of the target ball increased in direct proportion with the acceleration of the launcher and in inverse proportion with the number of balls in the chain. The results closely mimicked the predicted behavior of a Newton’s cradle if the collisions were not perfectly elastic. This isomorphism shows that judgments of physical events are not detached from the environment, as one seldom sees a perfectly elastic collision, and it speaks unfavorably to the hypothesis that, in such tasks, people rely on simple heuristics.