Yes, the problems to solve IMO are anonymity for players (so that sharing player history notes doesn't give an advantage), truly random shuffling(?), and collusion. However, the way I've been thinkin...
To be real, these games were never "fair" they were just unknowns back then. The calculations are not difficult, certainly not as deep as chess bots we are talking percentage risk tolerance and psychology dynamics. AI only seems to be better than those who play too loose. Knowing that you are drawing dead is a much bigger skill gap than people think.