On simple expectations and observations of intelligent agents: A complexity study