Abstract. A continuous pursuit-tracking task is the typical experimental paradigm to investigate implicit motor learning. Implicit motor learning is proven by a greater improvement in tracking of a repeated segment of a target path compared to random segments ( Pew, 1974). Recently, doubts about the validity of results obtained with this paradigm have been raised. Improved tracking of a repeated segment might simply be due to the characteristics of that particular segment. In response to these doubts, we seek to improve the continuous tracking task. Therefore, we computed a pool of 37 distinct target segments. Participants ( N = 36) practiced the tracking task, each one with a unique repeated segment in the middle and varying outer segments, all taken from the pool of segments. After five practice blocks of 36 trials each, a test block was performed where the repeated middle segment was replaced with a random segment. The tracking performance on the repeated segment was better than on random segments. Furthermore, we assume that learning was implicit, because participants’ answers to a posttest interview showed they were largely unaware of a repeated segment within the curves.