Eye movements have been recently recorded in participants watching animated triangles in short movies that normally evoke mentalizing (Frith–Happé animations). Authors have found systematic differences in oculomotor behaviour according to the degree of mental state attribution to these triangles: Participants made longer fixations and looked longer at intentional triangles than at triangles moving randomly. However, no study has yet explored kinematic characteristics of Frith–Happé animations and their influence on eye movements. In a first experiment, we have run a quantitative kinematic analysis of Frith–Happé animations and found that the time triangles spent moving and the distance between them decreased with the mentalistic complexity of their movements. In a second experiment, we have recorded eye movements in 17 participants watching Frith–Happé animations and found that some differences in fixation durations and in the proportion of gaze allocated to triangles between the different kinds of animations were entirely explained by low-level kinematic confounds. We finally present a new eye-tracking measure of visual attention, triangle pursuit duration, which does differentiate the different types of animations even after taking into account kinematic cofounds. However, some idiosyncratic kinematic properties of the Frith–Happé animations prevent an entirely satisfactory interpretation of these results. The different eye-tracking measures are interpreted as implicit and line measures of the processing of animate movements.