Don't understand AI's "place city" logic. What am I missing?

jpinard

Martian
Joined
Jan 18, 2002
Messages
760
Location
Enceladus, Saturn
Look at the image below. You can see the blue circle where the computer suggests you start your city. I have my Settler where I thought I should start my first city (right next door).

Am I missing something? Is it better to have fewer tiles you can upgrade, but being so close to teh ocean? Keep in mind this is the first city of my empire. Having more hammers (to me) makes more sense, than open ocean.

Thanks to anyone who can shed light on this :)

 
Where one thing I find is blue circle has a bad tendancy of appear on top resources. If there were no stone on that hill, i would understand why Ai choose to settle there. Its a hill so give city a +25 defense, its closer to ocean so frees up more land for your other cities to come. But with a stone there, come on, AI should seriously be recoded to take into consideration of the lost of production or commerce when settle on it.
 
losss of production? A tile with 3 hammers basis is going to be able to build first workers and settlers at extra speed. I really would like to have such a tile to start :)

Also, 1st settler is placed on a """good""" spot considering all the ressources, even thoe which are hidden. Blue circles are what the AI consider to be the best considering what you see only. IT explains why the two do not coincide sometimes.
 
Well lets take a look at your posistions. You have clams and a lake which with the lighthouse give you a surplus of 4 food, and the sheep which give you a surplus of 2 food (surplus is extra food beyond that required to feed the citizen).

That means that you can work six 1 food tiles or three 0 food tiles without slowing growth ... you have five 1 food tiles and three 0 food tiles, meaning that you won't be able to work every tile to its maximum ... you should note that a farmed plains (2F 1P) is inferior to a specialist (a priest gives 1P 1G 3GPP for example) until biology.

Now lets look at the AI blue circle ...

A city tile produces 2F 1P 1C ... BUT if the tile it founds on already produces more then those numbers, you keep them (this is without improvement and the removal of any trees). So a city on the stone is 2F 3P 1C +unpillageable stone.

You still keep all your food, but this time you can cottage EVERY plain (removing the forests) and even mine that grassland hill. The total number of coast (2F 2C w/ lighthouse) doesn't change and you gain one useless sea tile (2F 1C w/ lighthouse). You also gain the +25% defense bonus. That will be a fairly balanced hammer/commerce city and perfect for the beuacracy civic.

So the AI blue dot only has one bad tile (the sea tile) while the rest are usable. The original settler location, even if you windmill the plains hills, will have several unused/worthless tiles.

I agree with the AI blue dot in this case, though it is always good to think about it (the blue dots are independent of everything u've already built ... what it recommends might not fit in well with your empire).
 
If you build a city on a plains/hill tile, the tile yields 2 :hammers: instead of one. If you settle on a stone resource, the same bonus applies. I wonder if the two stack. If so, that would make the tile yield 3 :hammers: . That is a significant boost so early in the game and would justify losing the higher hammer yield you'd get later with a quarry.

Also you'd have access to stone very early so you could build Stonehenge, GW and Pyramids very fast.
 
What you need to remember Jpinard is that if you don't have enough food to work the tile, improving it won't help you at all. That is the problem your current settler has ... too many tiles that won't be worked or have to be farmed just to function.
 
you should note that a farmed plains (2F 1P) is inferior to a specialist (a priest gives 1P 1G 3GPP for example) until biology.

Sorry newbie question, but how is it superior? Isn't food > gold earlier on in the game?

Also i'm wondering, if a city that's built right on top of a certain resource, it does save you the trouble of building the necessary infrastructure (i.e. mines, plantation) to process the resource right? So does that mean it is advisable to do so, whenever circumstances permit?

If you build a city on a plains/hill tile, the tile yields 2 :hammers: instead of one. If you settle on a stone resource, the same bonus applies. I wonder if the two stack. If so, that would make the tile yield 3 :hammers: . That is a significant boost so early in the game and would justify losing the higher hammer yield you'd get later with a quarry.

I didn't know that :goodjob:
 
Sorry newbie question, but how is it superior? Isn't food > gold earlier on in the game?



Yes, a farmed plains tile will keep growth going instead of using the citizen as a specialist (and you may as well do this when not at a happy/health cap) but it only contributes 1 hammer to the city, while a specialist will contribute a lot more and great people points.

Even with a granary, if you only have +2 food surplus the city will grow in 6-10 turns for a smallish/medium city on normal. When the city grows, you will be working another farmed plains tile (let's assume you are already working all the "better" tiles), so you invest all those turns for +1 hammer. You could be even more drastic and go full food, shutting off all your mines to grow faster ... but you still only add 1 production to the city.

Thats why farmed plains (w/o river) are bad, because they contribute less as a tile to the city then any specialist and if you need farmed plains to grow, the city is essentially topped out.

------------

The city yield question is even simpler. Cities make 2F 1P 1C ... if the tile they are building on makes more, they keep the bonus. Found a city on a 3 food tile ... it gets 3 food. Found a city on a 2C tile and the city makes 2C. If you found on a forest, subtract 1 hammer from the tile If you found on a jungle, add 1 food to the tile.

Also i'm wondering, if a city that's built right on top of a certain resource, it does save you the trouble of building the necessary infrastructure (i.e. mines, plantation) to process the resource right? So does that mean it is advisable to do so, whenever circumstances permit?

It depends ... you would be foolish to found on gold cause you'd trade 7-9 commerce for 1-2 (2 if on river). Resources where the improvement isnt very good but still exceed 2F 1P 1C are worth founding on (3H stone tiles, 3F sugar tiles, 2H ivory tiles, etc)
 
I would build the city on the hill a 1000 times out of 1000. You can build a workboat really quick to work the clams, then build a worker to get the sheep etc. Like stated earlier, you can then build Stonhenge, Great Wall and Pyramids really fast.

Actually, I would consider this a perfect start.
 
The city yield question is even simpler. Cities make 2F 1P 1C ... if the tile they are building on makes more, they keep the bonus. Found a city on a 3 food tile ... it gets 3 food. Found a city on a 2C tile and the city makes 2C.
That's not correct. City tiles always yield 2F 1P 1C regardless of what the tile yield was before. The only exceptions are if you directly settle on top of some resources, or a plains/hill tile. Settling on corn will add 1F, stone, horse, iron etc add 1P. But if you settle on an ordinary tile (eg flood plains, grassland hill) it's always 2F 1P 1C.
 
Actually even though the stone tile is possibly a better long term site, the tile the OP suggested actually gets a bit more production, if you intend early stone based wonders. This is based on the assumption that you can only actually work 7 tiles for quite a long time.(at size 6 no happiness resources, guessing difficulty level)

Assuming a lighthouse exists (and using best production):-

City on Stone: City tile+ Clams+Sheep(pastured)+Gra/Hill/Mine+3 x Forest/Plains= 14:food:13:hammers:4:commerce:

City on Tile of OP: City Tile+Clams+Sheep(pas)+lake+Stonequarry+2x Hills/Plains/Mines =13:food:15:hammers:6:commerce:

As that site really isn't that great wherever you build it as a longterm capital, and you may wish to move the palace elsewhere, this may also be a viable strategy.

Didn't say it is, just it might be ;)
 
That's not correct. City tiles always yield 2F 1P 1C regardless of what the tile yield was before. The only exceptions are if you directly settle on top of some resources, or a plains/hill tile. Settling on corn will add 1F, stone, horse, iron etc add 1P. But if you settle on an ordinary tile (eg flood plains, grassland hill) it's always 2F 1P 1C.

* The corn gives 1F cause the tile is 3F (greater than 2)
* The horses, iron, etc give 2P because they give 2P base (unless you are sure beyond doubt that a grassland horse gives 2P which I don't think it does)
* A grassland hill is 1F 1P ... which is not better then 2F or 1P thus no change
* A flood plain is an improvement much like a forest is (thats why it says flood plain/desert). By building the city, you remove the flood plain making the tile 0F 1C ... which again, is worse than 2F 1P 1C
* A plains hill is 2P, which is higher then 1P so a plains hill city is 2F 2P 1C.
 
If you look into the 'resources' section in the civilopedia, you'll see that it lists the bonus each resource provides. Eg stone, horse +1 :hammers:, corn +1 :food:. I don't see what this bonus could refer to other than what you gain when settling on top of the resource?

Still, this is odd because it says also sheep +1 but if you settle on a grassland/sheep tile you still only get 2 :food: 1 :hammers: 1 :commerce: Maybe you're right, but then the bonuses listed in the pedia don't make sense.
 
Those are the bonuses that the tile gets without any improvement. If that bonus is BETTER then the city's natural production, it stays.

Sugar is an easy example.

Grassland is 2F
Grassland/sugar is 3F (sugar is +1F in civlopedia)
grassland/sugar/plantation is 4F/1C/collects resource (sugar plantation is +1F +1C)

If you found a city on a grassland/sugar, 3F > 2F and the city makes 3F 1P 1C. You sacrifice the fourth food from the plantation you can no longer build by doing this.

This makes certain resources like corn (6F if irrigated) or gold (9 commerce if financial and on a river) so powerful that founding a city on them is stupid, but some of the resources are kinda "meh" and founding directly on them can be a good idea sometimes ... like sugar.
 
Yeah, but the example grassland/sheep has only 2 :food: 1 :hammers:, without a city. According to your explanations, it should yield 3 :food: (2 from grasslands and +1 for the sheep as stated in the civilopedia)

edit: Lol, forget that. When testing this, I fell victim to the 'hill displayed as flatlands' bug: The grassland/sheep tile was really a hill/sheep tile :lol:
 
A grassland/sheep has 3F ... that must be a plains sheep(1F 1P base +1F from sheep). But you have the right idea.

*EDIT* in fact, if you look at the picture the OP provided us, you will see that those sheep are on a plains tile ... with the +1F as you report.
 
Yes, don't use the blue circles. You're alot smarter than a computer program, or at least you should be. Use your own judgement.

...

Except the computer's right in this case...
I'll admit I don't usually calculate the food a city will eventually produce, but the AI does... I've found many a time, like in the OP's example, the computer wants me to settle one tile away from where my gut would place the city. I always calculate the food in these cases, and usually it turns out my chosen spot wouldn't be able to work all it's tiles anyway.

The blue circle doesn't know a lot of things- like what tiles you're reserving for future cities, or what sites blocks off a rival's expansion better, but it is good at simple calculations to maximize resources...
 
Top Bottom