The eSports industry has greatly evolved within the last decade. Professional eSports players have contracts with teams, perform daily routine training activities, and participate in tournaments. At the same time, similar to typical sportsmen, the eSports athletes suffer from injuries that significantly reduce their performance or even prevent them from participating in the training process or tournaments. Also, it is not a trivial task to evaluate the eSports athlete's performance in the training process or when hiring a new team member. In this study, we address these two issues. We use EEG recording to monitor the subject's performance status and use machine learning to find the correlation between EEG recordings and eSports athletes' performance. Professional eSports players and casual players (10 players from each group) participated in gaming sessions of 20 minutes while the EEG recordings were collected before and after the game for 10 minutes every time. After filtering and removing artifacts, we collected the statistical difference before and after the game and used them as the features for two binary classification problems: player professionalism and player health state. Our results demonstrated that machine learning models trained on EEG data can detect professional players with 95% F1-score and detect player health state with 90% F1-score on 10-fold cross-validation.