Depending on the difficulty of the task in terms of movement duration and the number of elements forming the sequence, recent research has shown that movement sequences are coded in visual-spatial coordinates or motor coordinates. An interesting question that arises is how a specific manner of performance without a change in such functional difficulties affects the representation of movement sequences. Accordingly, the present study investigated how the way in which a movement sequence is performed affects the transfer of timing properties (absolute and relative timing) from the practised to unpractised hand under mirror (same motor commands as those used in practice) and non-mirror (the same visual-spatial coordinates as those present during practice) conditions in two experiments each with segment movement time goals that were arranged differently. The study showed that after a limited amount of practice, the pattern of results obtained for relative timing differed between the two experiments. In the first experiment, there was no difference between retention and non-mirror transfer, but performance on these tasks was significantly better than that for mirror transfer, whereas in the second experiment, there was no difference between the mirror and non-mirror transfer. For total errors, no significant difference was found between the retention and transfer tests in both experiments. It was concluded that the way in which a sequence is performed could affect the representation of the task and the transfer of relative timing, while absolute timing could purposefully be maintained if necessary.