The after action review (AAR) is an essential component to training. To better focus AARs on critical points of team performance, Aptima previously developed a software tool called SPOTLITE that incorporates Mission Essential Competencies^SM and guides observer/instructors easily through assessment questions during live-virtual-constructive training. However, even with such tools to systematize the assessment and debrief process, performance measurement is very labor intensive. Methods are needed to automate this process, and at least approximate the final results in order to guide the AAR. One form of data that is available digitally, and thus could be a basis for such an automated process, is Internet Relay Chat. Chat allows an operator to monitor more than one channel (or chat room) at once, is persistent so operators can trace back the information and decision process, yet is as instant as radio communications. Chat is used extensively in command and control operations, and during training is itself a major point of focus for AARs. We previously presented (I/ITSEC 2009) measures of communications performance based on chat that changed over the course of a week's training at an Air Force Training Research Exercise (TREX) which models the Dynamic Targeting Cell of an Air and Space Operations Center. Since observer-based performance data also was recorded during these exercises, we investigated predicting these scores using communication measures applied to chat. Results indicate that these measures can provide an adequate (r² > 0.6) proxy for the average of the manual observer-based performance assessment of the team's performance; thus, chat analysis could provide a mechanism for missions to be ranked automatically for AAR.