1. AI burns his cards and laughs at player one.
2. AI reboots the machine.
Sorry, I couldn't resist. Yu-gi-oh and other other collectable card games are extrememly complex due to the thousands if not millions of possible card combos, decks and plays.
Wow, I'm not even sure where to point you at to begin with. Try this, grab a deck and start playing out cards. As you play, right down the 'rules' that led to that decision. By rules, I don't mean rules of the game (also required*) but the logic that drove you make that play. Repeat until you can play like a robot, always following those rules. After several hundred (thousand?) plays, you'll have enough data to
begin building a Yu-gi-oh AI.
Amit's has a bunch of articles on AI, perhaps you can find something there to help. NOTE: this is not a Dark Basic site. You will have to take the ideas learned there and apply them to your program.
http://www-cs-students.stanford.edu/~amitp/gameprog.html
--
TAZ
* My son tried to teach me to play. Fizbin was easier to learn.
