Traditional approaches of researching self-regulated learning (SRL) fail to capture how learners actually employ studying tactics, how tactics are strategically adapted to specific learning contexts, and how learners adapt tactics and interweave them to form an efficient strategy. Computer traces can capture SRL "on the fly," and enable researchers to track learning events in a nonlinear environment without disrupt-ing the learner's thinking or navigation through content. More importantly, data obtained in real time allow "virtual" re-creation of learners' actions during study-ing. There were 107 Chinese university students' traces collected while they solved assigned problems through searching the web. By linking their regulatory activities during online search to their goal profiles, results showed that mastery-approach-dominant students were most strategic, whereas performance-avoidance-dominant students were least. Moderately motivated students showed a mixed pattern of deep and surface study strategies. Implications of the findings were also discussed.