It's a forgone conclusion but on principle, Forbidden City still offers more flexibility without any random factors compared to Kilwa. At best you get base 3 econ cards in T1 gov't. At best 4 in T2, which I spend the bulk of my games in (and more often than not, still 3 as I am most often in Theocracy). T3 Communism gives 4.... So, at almost all times excepting Future governments, the Wildcard from City, which may as well not even have a placement requirement, offers the greatest increase to the most powerful factor of the game (how many Econ policies you can run). Timing card swaps doesn't give you that same power .. you delay builders, you delay wonder chops, you lose turns of gold/faith/science bonuses. Not so with City.
Worth mentioning since so many people are decrying the single desert tile requirement of Pyramids on Shuffle maps... if you end up with a coastal/oceans type of map you're actually at a significant disadvantage with Kilwa. Not only is there more coast for the AI to place it, competing with you, but you're limited in exploration to Cartography to find your requisite city states, which is on the opposite side of the tech tree to Machinery for Kilwa. People either aren't mentioning or are significantly downplaying the likely scenario of your 2nd city state of X type (on average across all map sizes and city state types there will be exactly two of each type) being undiscovered for a significant portion of the game, and how difficult it may end up being to wrestle control away from another heavily invested Civ, or to liberate a conquered one on the opposite side of the map.
I have no objection to the premise that Kilwa offers the greatest potential, but it does suffer from being the least reliable.
Good game!
Forbidden City [3]
Kilwa Kisiwani [10]
Pyramids [24]
Worth mentioning since so many people are decrying the single desert tile requirement of Pyramids on Shuffle maps... if you end up with a coastal/oceans type of map you're actually at a significant disadvantage with Kilwa. Not only is there more coast for the AI to place it, competing with you, but you're limited in exploration to Cartography to find your requisite city states, which is on the opposite side of the tech tree to Machinery for Kilwa. People either aren't mentioning or are significantly downplaying the likely scenario of your 2nd city state of X type (on average across all map sizes and city state types there will be exactly two of each type) being undiscovered for a significant portion of the game, and how difficult it may end up being to wrestle control away from another heavily invested Civ, or to liberate a conquered one on the opposite side of the map.
I have no objection to the premise that Kilwa offers the greatest potential, but it does suffer from being the least reliable.
Good game!
Forbidden City [3]
Kilwa Kisiwani [10]
Pyramids [24]