If synthetic teammates are to be considered team players, then they must be better equipped to handle the subtleties of communication and coordination with their human teammates. In this study, the team communication behaviors of a human-automation team were analyzed for the identification of which are the best predictors of team performance. The LASSO (Least Absolute Shrinkage and Selection Operator) method was used to select the team communication behaviors that were the best predictors of team performance and, in the end, 16 such role related communication behaviors (at both the role and condition level) were included in the final model. Findings indicate that in general, negatively perceived communication behaviors are predictors of negative team performance. Through this study, we also learned that even when human team members follow their optimal and expected communication behaviors when communicating with a synthetic teammate, these behaviors are still predictors of negative team performance. This finding holds important future considerations: even if human team members are properly communicating with a synthetic teammate, the errors and lack of human-like behavior on the part of the latter can still result in a negative team performance.