During therapy sessions, feedback is often provided concurrently by the physical therapist as the patient attempts to perform a movement and after the movement attempt. This feedback is provided to enhance the patient's balance abilities. However, recent studies in nondisabled populations have suggested that frequent feedback may be detrimental to retention or learning of motor skills. This study compared the effects of 100% relative frequency of knowledge of performance (KP) with 66% relative frequency of KP for motor learning on balance retraining in patients with hemiplegia. Twenty patients with hemiplegic were randomly assigned to one of two experimental groups. The acquisition phase consisted of 16 blocks of 5 trials for 2 days (80 total practice trials). The retention phase consisted of 2 blocks of a short-term retention test, one day after the end of the acquisition phase and a long-term retention test, one week after the end of the short-term retention test. In the 100% feedback condition, participants received feedback after every practice trial. A faded KP schedule was used in the 66% condition. No significant differences were found between the two groups during all experimental phases (acquisition and retention phases), (p>.05). However, there were significant decreases in balance index for both groups of acquisition phase (p<.05). These results suggest that 66% relative frequency of KP is not more effective than 100% relative frequency of KP with respect to retention over time when hemiparetic patients attempt to learn balance.