yeah. defining the problem as hard does not even begin to describe it. I believe that an AI that plays perfectly while seeing everything (i.e. oppponent hand +

yeah. defining the problem as hard does not even begin to describe it. I believe that an AI that plays perfectly while seeing everything (i.e. oppponent hand + both libraryies, so no hidden state) would still be an impressive accomplishment and could be used to fine tune the deck-building process.