Accumulating evidence proposes that time-pressured form-focused tasks such as grammaticality judgment tests (GJTs) can measure second language (L2) implicit knowledge. This article, however, presents that these tasks draw on automatized explicit knowledge. A battery of six grammar tests was designed to differentiate automatized explicit knowledge and implicit knowledge. While three time-pressured form-focused tasks (an auditory GJT, a visual GJT, and a fill in the blank test) were hypothesized to measure automatized explicit knowledge, three real-time understanding tasks (a visual-world task, a word-monitoring task, and a self-paced reading task) were hypothesized to measure implicit knowledge. One hundred advanced L2 Japanese learners with first language Chinese residing in Japan took all six tests. Confirmatory factor assay and multitrait-multimethod analysis supplied an array of evidence supporting that these tests evaluated two types of linguistic knowledge separately with little impact from the method effects. The results assayed separately by length of residence in Japan (a proxy for the amount of naturalistic L2 exposure) displayed that learners with longer residence in Japan can use implicit knowledge in the real-time understanding tasks with more stability than those with shorter residence. These findings point out the potential of finely tuned real-time understanding tasks as measures of implicit knowledge.