We have recently shown that young healthy adults (re-)learn when deliberately repeating a well-rehearsed task of mundane activities, rather than just temporally adjust to new specific task conditions (Leizerowitz et al., 2023). When “de novo” learning occurs, the attainment of stabilized performance (‘plateau’) in the first session is deemed necessary to achieve a long-lasting routine. Here, we followed the effects of practice on the Timed-Up-and-Go (TUG) task, a clinical test comprised of highly-familiar motor activities - rising from a seated position, walking, turning, and returning to a seated position, using two training protocols (Figure A). In Long-Training (18 TUG iterations) participants attained a performance plateau by mid (initial) session; Short-training (5 TUG iterations) terminated before a plateau was attained. Performance and kinematics were assessed in three consecutive sessions by video-recording and inertial measurement units; the first two sessions were 24 hours apart, the third session followed 1-week later. Both training protocols resulted in robust, continuous gains in TUG performance, evolving across the three sessions and well-retained between sessions; the overall gains in the two protocols were indistinguishable (FIGURE C). However, kinematic parameters that showed continuous learning and good retention in the Long-Training protocol showed no such effects in the Short-Training protocol(Figure D). The results suggest that although both protocols led to 'de-novo' acquisition with comparable gains in performance, the nature of the ‘how to’ knowledge acquired and retained differed; the procedural knowledge gained depended on whether or not a plateau phase was attained in the initial session.