[j2kun]

Just curious: have you watched any of these games to see what kinds of mistakes the AI makes for various settings of the parameters in your scoring functions?

Also, I'd be interested in doing offline learning. For example, if you have some formal way to record the actions taken in a game, you could ask pro players to submit their games. Of course, this would require Blizzard to allow a replay mechanism like it does for starcraft. But it would make the AI design process much more interesting than exploring game trees (without perfect information) or greedily picking the best single move.

[oyachai]

I have informally looked at what the AI does. Turns out, as far as I can see, the AI does a pretty good job... with a BIG CAVEAT. The decks that I have used so far consist solely of minions and one other special card, so there really isn't any "mistakes" the AI can make. The decks are just so simplistic at this point. As more interesting cards are implemented, I'm sure there will be situations where the AI will make interesting mistakes that we can learn from. It is something I'm really looking forward to exploring.