hi.
nice work you have done on the AI math
I think the "area" bit it will be borked down if one plays single continent.... or on archipelagoes....
(seen from europe, a sparsely settled america with only 2 big cities (New york and Sao Paolo) would be chosen over a compact and rich south-asia with 5 big cities with 1 in each of the 3 indonesian islands, 1 in australia, 1 in new zealand).
Maybe you could just let this part down. as it is a big approximation and you won't do it for the other GP (mostly they will stay in their own culture).
maybe the "area" bit could be managed by taking another factor into account : the civ's score, the number civ's cities bigger than 10 (or15) that have no trade route with you, the total commerce of the civ (minus the one of the city you are targetting); the size of the second biggest city of the same civ that has no trade route with you...etc
or you could do your check not by area but by civ: adding a g' of 4, and searching the 3 next-profitable actions with the same civ. That might help in case of big differences in landmasses sizes (however that would reduce the perceived importance of trade routes with CS...).
nice work you have done on the AI math
I think the "area" bit it will be borked down if one plays single continent.... or on archipelagoes....
(seen from europe, a sparsely settled america with only 2 big cities (New york and Sao Paolo) would be chosen over a compact and rich south-asia with 5 big cities with 1 in each of the 3 indonesian islands, 1 in australia, 1 in new zealand).
Maybe you could just let this part down. as it is a big approximation and you won't do it for the other GP (mostly they will stay in their own culture).
maybe the "area" bit could be managed by taking another factor into account : the civ's score, the number civ's cities bigger than 10 (or15) that have no trade route with you, the total commerce of the civ (minus the one of the city you are targetting); the size of the second biggest city of the same civ that has no trade route with you...etc
or you could do your check not by area but by civ: adding a g' of 4, and searching the 3 next-profitable actions with the same civ. That might help in case of big differences in landmasses sizes (however that would reduce the perceived importance of trade routes with CS...).