That was very neat and pretty complete experience. Guessing the rules by previous results is a pretty fun concept - kinda flipping the board on humans when it comes to the way we sometimes train ai. Didn't overstay it's welcome either.
