This paper reports on a qualitative study that explored the veridicality (i.e., the completeness and accuracy) of think-aloud protocols (TAPs) in English-as-a-foreign-language (EFL) writing and illustrates how retrospective verbal reports (RVRs) compensated for TAPs in understanding online revision. Forty-three Chinese sophomores, upon writing while thinking aloud, were asked to provide RVRs regarding revisions, and then to reflect on the veridicality of their TAPs. Their reflections were analyzed inductively. Various omissions of think-alouds were revealed, but they were perceived as not serious, and the accuracy of TAPs was stood by. Further evidence concerning the (in)veridicality was found in the RVRs when 516 episodes of RVRs and corresponding TAPs were compared, and the RVRs were found to offer additional information that concerned intermediate processes leading to revisions. Implications for using TAPs and RVRs in and for EFL writing research and classrooms are given.