I realise this isn't a programming Q&A site, so bear with me. I'm interested in finding out the kind of information someone who is good at a game naturally filters out of the vast information from the screen. A very basic example would be ignoring the background textures and not ignoring the textures of other payers characters.
It seems like it would be interesting to attempt to write a program that might highlight what is/isn't important and see how that compares to what players perceive as important. Does anyone know of any reference that might get me started on this task (don't worry about the code side of things, I'm talking about machine learning techniques etc.). I've completed the Stanford online machine learning course so I have a reasonable understanding of the principals, but I wonder if there are other, better suited techniques out there.