Comment by jonnycowboy

the real question of course would be how to train the AI to have varying levels of "expertise" (ie: penalize if score is too high)