Considering that the prizes in CS:GO tournaments reach several million dollars each, any information that can potentially lead them to victory in the match and the tournament would mean a lot to the teams. A system that would indicate the shortcomings of professional teams, and give a prediction of the outcome of the match, taking into account the composition of the team and the compatibility of their playing styles, would be of great use when creating teams. The idea of this paper is to show through several experiments how certain factors affect the prediction of match's outcome and which combination of them would give the best prediction results. In addition to the success of professional players during their career, the factor of the country the player comes from as well as his style of play was also taken into account. The paper also presents an algorithm for evaluating player ratings. The KMeans clustering algorithm was used to predict the style of play, while the XGBoost Regressor, Random Forest algorithms, and a neural network were used to predict the match's outcome. In most cases, the experiments produced the expected results, given the nature of the problem. In the paper, it was shown that data on the style of play and player's country data, in most cases, have a positive effect on the prediction of the outcome of the match.