Yeah, I think poker is a great example of the limitations of game theory. I'm interested in this recent work where they got a computer to beat champion Go players. For a long time, people didn't seem to think that was possible, but then they made some sort of sophisticated machine learning algorithm that looked at tens of thousands of historical games and learned how to play extremely well. No game theory, just machine learning and tons of data. Maybe something similar could be done with poker or Steem.