Actually, come to think of it, the biggest problem with simulating a "CCG-like" game is simulating credible opponents. I think game AI has REALLY lagged for some time now.
Just look at RTS games, where lazy design has usually mean that a "harder" opponent just means the computer builds faster than a human can multitask, can always see you, has unlimited resources, or other silly fallbacks.
When you look at the actual AI strategies, they seldom have even the simplest stuff - things like sneak attacks, scouting, true resource management for long or short term goals, etc.. Instead the computer will tend to spam a given unit or number of units, rarely use them in concert, and generally behave like a stereotypical Machine Menace. In the end, you're not trying to outthink an opponent, you're racing a clock or a bank account. All this does is reward the twitch-city idiots.
It's embarassing. Fifteen years later and we still have major game companies that make computer opponents attack in scheduled waves. We might as well be playing Space Invaders.
