Yup, that's the gist. When you can easily describe to the computer what good play is (maximizing the score from the Atari game), RL is an option. If that's hard to do (say, teaching a computer to fly a helicopter) apprenticeship learning might be better.
I do some work in AI and RL. The specific approach used by DeepMind and others on the Atari games (based on deep neural networks) wouldn't scale to a game like SK, there are too many states to consider. There has been some research done applying RL to 4X games though, e.g. http://people.csail.mit.edu/camato/publications/LearningInCiv-final.pdf talks about using it with Civ IV. Th
First post. IMHO, this is the most important thread in this forum. The concepts behind FE, LH and SK are awesome, but as described above the execution has been lacking. Dr F is doing a fantastic job cataloging bugs that have existed for some time, as well as new ones that should have been be caught in regression testing. Although he's been the most vocal, I'd imagine he's expressing the frustrations of many, many customers. I've avoided becoming an early adopter of SK for