anarres said:
Just so I know, can you tell me how 1 line of code could make the AI use sensible FP placement?
I doubt one line can do that. (Can't be sure without seeing that part of the code.) But I do think one line can do what I said, i.e. "would have made the AI better in this regard without dumbing us down"
It was the suggestion of adding the test:
if (pLoopCity->pClosestCapital != pClosestCapital) continue;
The result of adding that would be "distinct" rank lists for the Palace and FP. In the Civ3 and PTW model any given city could affect the rank of other cities from both the Palace and FP. In a distinct list model each city would affect the rank of only other cities ranked from the same place.
I think that many people just assumed it already worked in the distinct list way. I.e. that there were two rank lists, one from the capital and one from the FP, and that every city appeared only on one list or the other, not on both.
Changing it to work that way would have greatly increased the value of an FP placed near the Palace, and therefore would have helped the AI because it tends to place FP that way. And would help casual players in the same way. And would make a builder approach to the FP worth considering. But it would leave the direction and distance of the FP from the Palace as significant factors for players squeezing all they can from the game. And would leave the previous behavior as it was if someone wanted entirely separate Palace and FP regions distant from each other.
It would have required some play testing of course to confirm the predicted behavior and whether it was a good thing. But that should of course have been assumed for any change to the FP model, not that it was done for what they implemented