Researchers have identified kinematic differences between deceptive and non-deceptive rugby reorientation actions. However, the honest and deceptive signals corresponded to ‘deception detection’ (accuracy increasing) rather than signals that caused deception (accuracy decreasing). In this study, statistical parametric mapping and multilevel modelling were applied to examine the kinematic differences between sidestep and non-deceptive actions during the time window of deception. The analysis compared three-dimensional motion capture data from 144 deceptive actions and 144 genuine actions performed by six high-skilled rugby players. Results indicated that the kinematics of deceptive actions were characterized by a combination of exaggerated head roll, outside foot and centre-of-mass displacement, and attenuated thorax roll and yaw relative to genuine actions. These are candidate sources for the cause of deception, either individually or in combination with other sources. Furthermore, the results indicate that previously identified ‘honest’ signals may not be reliable sources of information earlier in the action sequence. •Identified differences between genuine and deceptive actions in the time window in which players are susceptible to deception.•Analysis of discrete and continuous data.•Deceptive actions are characterised by a combination of exaggeration and attenuation.•Previously identified ‘honest’ signals may not be a reliable source of information for perceptual judgements.