True optimal strategies have not to my knowledge been found for any of the most popularly played poker games. Several people have made stabs at approximating them, esp the Alberta people involved in this publicity stunt, and also the CMU people (google "tuomas sandholm poker" for the Sandholm/Gilpin papers, notably their Rhode Island holdem solution).
We can easily solve a single street (one "street" is a betting round) at this point; many techniques are available to do this. However, solving multiple streets is much more complex, because the action on previous streets significantly affects the hand distributions for a particular situation. So unlike, for example, backgammon or chess, where you can specify a situation and solve from there, and the game listing to that point is irrelevant, in poker the action on previous streets is part of the specification of the current situation.
All of this, of course, has nothing to do with trying to "read" your opponent or any kind of psychology, but only with finding an equilibrium strategy that cannot be exploited. I have strong reason to believe that such a strategy would be oustandingly strong against all human opposition, such that it would be quite unlikely that the very best human player would have even a 20% chance of being ahead of the bot after 10,000 hands.
It isn't easy being the parent of a six-year-old. However, it's a pretty small price to pay for having somebody around the house who understands computers.