In my earlier work on Chinese Poker, I analyzed a hand by comparing it with a sample set of already-organized hands. What I learned was that the vast majority of time was spent identifying which hands from the sample set were valid--- i.e., did not contain any cards from the hand being analyzed. Even with a very fast comparison (xor on bitvectors
(
Read more... )