This paper describes the architecture and progress in the development of the UvA soccer simulation team. The UvA-team uses a three layer model to separate simulator details; basic soccer skills and the overall control from each other. In the previous year we focused on getting the system up-and-running. The current eort is in the use of learning mechanisms for optimizing the parameters of the soccer skills like dribbling and shielding. We are also applying the capabilities of the coach to characterize an opponent on the basis of logles to select suited strategies against that opponent as well as to adapt parameters of the control and sensor systems during the play.