In addition to the UvA Trilearn 2003, Brainstormers, and Austin Villa Coach rcssjava code release, a few groups from the Fall 2010 version of this course completed projects that may be good starting points for future projects.

These students provided their code on an 'as-is' basis, but you may use it as you see fit as long as you acknowledge them in your reports. If you decide to use one of these code bases as your starting point, we can likely provide you with the final report associated with the code base if you ask.

• Python Agent by Jason Bradshaw - This project involved writing an agent for version 11 of the 2D Robot Soccer Simulator completely in Python. As one can imagine, this required much low level programming, and the resulting agents were very basic despite the substantial work done by Jason for this project. As such, there is plenty of room for this code base to be expanded by future students. Here is the code base, which is cleanly written and substantially documented. To run his code type './agent.py soccerpy 11 $1 $2'. Feel free to contact Jason through the contact information on github if you decide to move forward with his code and you have questions.
  
  
• Working UT Austin Villa Coachable Players by Jackie Tsay and Jonathan Black - Jackie and Jonathan worked hard to get the UT Austin Villa Coachable Players to run on the UTCS machines during the Fall 2010 semester. Although their project ended up taking a different direction, they have provided the Coachable Players that worked on the UTCS machines during the Fall 2010 semester (before the upgrade to Lucid). Jackie and Jonathan do warn you that it is difficult to alter formations using these players - so if that is something that is important to you, you may want to seriously reconsider using the UT Austin Villa Coachable Players. The players they got to work can be copied from /projects/cs344M.pstone/patched_ut_coachable_players.tar on the UTCS network.
  
  
• Addition of Roles to UvA Trilearn by Aaron Lawyer - This project involved the addition of roles and smart passing to UvA Trilearn. Aaron maintained many of the basic Trilearn skills, as his project focused more on improving actual team play. As such, most of his modifications and additions can be found in PlayerTeams.cpp. There are four different loops in PlayerTeams.cpp that correspond to the four types of heterogeneous players on the field. Each type of player acts a little differently than the others in terms of aggressiveness. The code can be copied from /projects/cs344M.pstone/UTPorto.tar.gz on the UTCS network, and can be run using the included start script (remember to run ./configure first).