Scully and Newell (1985) proposed that when viewing demonstrations learners perceive and imitate relative motion, and that this information is most effective early in learning when new motion patterns are acquired. Researchers who have tested these predictions have emphasized long-term changes in learning in retention rather than immediate performance effects (e.g., Al-Abood et al., 2001; Scully & Carnegie, 1997). By ignoring the immediate effects of a model on early learning, the saliency of relative motion information in the demonstration remains unclear. Three experiments are reported that examine the rate, stability, and longevity of changes in learners' relative motion when visual intrinsic knowledge of results (viKR) is readily available, removed or available but redundant. It was predicted that reduced conflict in the synergy between the coordination and control of the movement would induce greater, quicker changes in coordination when viKR is not usable. In Experiments 1 and 2, novice participants chipped a soccer ball to a target. In Experiment 1 (viKR present) groups observing a model (VIDEO or POINT-LIGHT) and a no-demonstration CONTROL group reduced radial and variable error (p < .01). However, the model did not facilitate changes in intra-limb relative motion. In Experiment 2, viKR was removed by occlusion at ball contact. Significant Group x Test period interactions showed that participants observing the models changed their knee-hip (p < .001) and knee-ankle (p < .05) relative motion to closely imitate the model by the third acquisition trial. Changes were maintained throughout practice and retention. In Experiment 3, participants practiced a novel, maximum velocity throw. After observing a video model, participants showed significantly better scores than a no model CONTROL group for elbow-shoulder relative motion (p < .05). For the knee-hip, a significant Group x Test period interaction (p < .05) was present. The CONTROL group showed no changes across acquisition, whilst the MODEL group immediately changed to more closely imitate the model. Findings support Scully and Newell’s (1985) position that learners perceive and imitate relative motion information. However, the significance of viKR is illustrated. When usable outcome information is available, the salience of relative motion is compromised as learners seek to coordinate and parameterize movement patterns. Keyword(s): coaching, performance, research