Roland Johansen
Deity
No, I understand that it's working as intended. I just don't see the variation being useful. Basically it seems like a variation for variation's sake, not to actually add anything useful to the game.
Maybe the variation in placement rules of tile improvements isn't that useful by itself. Although I don't think that every element of a game should necessarily lead to a choice or thinking process. Some are there just for looks or variation while they don't make the game necessarily deeper. It seems logical that a watermill is placed on the river so in this case it's probably based on some sense of realism. I know that this game is not that realistic, but the developers do try to keep some level of realism in the game.
The barbarian city placement routine is the same as the AIs.
Yes, you're right. That wasn't my smartest remark.
The barbarians of course don't know the placement of every other city build up to the moment they place a city and so their cities often don't fit nicely into the pattern of cities build by AI or humans. Ok, that one was less stupid.
What you've overlooked is that reason for city placement. A lot of the time you place a city based on resources that it'll provide, especially with food. Overlapping a city is fine if there are separate food sources. But if you are overlapping the food sources, then you've got problems. You're basically forming a city under false pretenses. Only one of the two cities can actually use the food, so one of the cities that may have been counting on the food to grow is going to end up expanding much slower than it otherwise would.
Ah, you mean that under Blake's city founding algorithm a newly placed city would count a food resource (or other valuable tile) inside its big fat cross for the full value even if it is shared with another city? Is that it?
You're right that that is not really correct. A human player can switch the valuable tile between the cities and get extra value out of it by using it by the city which can use it most. However, the AI won't do this and even if it would, the tile is still worth less when shared by two cities than when a city has it all for its own.
Halving the value of such a tile would intuitively seem the right choice. (I don't know what changes you made.) However when you do this you might get unexpected side effects where the AI places cities too far from one another because the city founding algorithm is probably balanced around this overvaluing of overlap tiles. The only way to find out if it works as you hope it does is by testing it like you asked us to. It's too hard to look at the programming code and say: yes, the AI will now place the cities like I want it to. It's a complicated issue and I admire the fact that you're trying to tackle the problem.
So I'm not trying to eliminate overlap completely (although I'm trying to discourage excessive overlap), but I'm trying to encourage each city to have it's own set of resources. Which isn't to say that a city needs resources, but I don't want it to rely on sharing them.
A city with just 2 fully developed towns can usually easily pay for its upkeep. With a courthouse, bank, market and grocer and the various buildings that improve trade routes, you probably don't even have to use tiles to break even in the late game. So a city using 2 grassland tiles (besides the founding tile) is already useful. There is also an opportunity cost of spending hammers on a settler which doesn't give a lot of return for the investment, so you probably need a little more than 2 grassland tiles. All I'm saying is that if the founding algorithm leads to areas of 3-5 decent tiles in between cities that are not being used by any city, then something is wrong. Then the AI is wasting land. So that is the thing that needs to be tested I think. Seeing if your changes work the way you hope/think.
I try to use almost every tile in my area with the lowest number of cities.
But yes, primarily I'm focused on two things - making the AI more likely to choose a river tile, and making the AI less likely to settle one tile away from the Ocean.
Bh
Why do you want to let the AI focus more on river tiles? I know they presently get some extra value in the code for founding on a river tile. The in game advantages are a health bonus and the late game bonus from the levee. The health bonus is nice, but seldom needed as happiness is usually the crucial limiting factor until the late game with factories and such. The levee bonus is so late in the game that 80% of the game has been finished already. Is the value for river founding really too low?
By the way, I modded the levee a bit so that it doesn't require that a city needs to be build on a river. A really strange requirement if you ask me. As if levees are only build in the city center and not the neighbouring lands. But of course my own modding is not the issue here, just wanted to say that I thought the levee has a strange requirement in civ4.