The feedback from SCG competitions should not only be useful to the agent developers but also to the professor.

I think there is useful information in each competition history pointing to knowledge that the 
students still need to master. The competition that Sweet won recently is a good example where
it became apparent that the agents don't consider relation interactions yet.

Please can you compile for each competition a brief report of issues that should be 
covered in the next lecture. For example, you notice some patterns that almost all agents have.

-- Karl