The aim of this study was to ascertain whether the performances of implicit and explicit learners would converge over an extended period of learning. Participants practised a complex motor skill--golf putting--for 3000 trials, either with a concurrent secondary, tone-counting task (implicit learning) or without such a task (explicit learning). The cognitive demands of the secondary task were predicted to prevent the accumulation of verbalizable rules about the motor task. The implicit group reported significantly fewer rules than the explicit group on subsequent verbal protocols. The performance of the implicit group remained below that of the explicit group throughout the learning phase. However, no significant differences were found between groups during a delayed retention test. Additionally, for the participants in the explicit group only, a Reinvestment Scale score correlated positively with the number of rules accrued and negatively with overall putting performance during the learning phase. We use the results to argue against the excessive use of verbal instruction during skill acquisition, which might be unnecessary and ultimately might hamper performance under stressful conditions.