Does blue circle really good for city location?

Artosoft

Prince
Joined
Jul 23, 2006
Messages
402
Does the blue circle for city location (for settler) is the best location according to computer calculation? Or do you need to think about the best location than blue circle one? How if you have more than 1 blue circle?

Regards,
Arto.
 
Artosoft said:
Does the blue circle for city location (for settler) is the best location according to computer calculation? Or do you need to think about the best location than blue circle one? How if you have more than 1 blue circle?

Regards,
Arto.
Blue Circles tend to point out locations that would grab several resources. They are often NOT good city locations however as the computer often fails to take all the circumstances into account.

The most common example is the computer suggesting a city site that is one tile away from the coast. Which is usually a bad idea as without the ability to build a lighthouse coastal and ocean squares are merely one food and essentially worthless.
 
I don't think the blue circles take into account your neighbors.

I recently built two cities several tiles away from the 'blue circles' so I could a) prevent the persians from getting copper, and b) block the persians ability to expand.
 
As Araqiel pointed out, the blue circle points locations that capture several resources, and generally tend to be locations with transport access (rivers or coastal areas) and ones that would guarantee a more or less balanced growth. It won't help you though if you are looking for a good specialized city place, such as a unit factory or a great people farm.
 
Hey Joni said:
A question - does the AI build its cities according to the blue circles? Anybody?

It looks like they normally do it.
If, after a war, I send my settlers to the locations of razed cities,
normally there is a blue circle at the tile with the ruins of the city.

As for my strategies:
Most of the times I build my cities at the tile with the blue circle,
but there are also times (10-20 % of the cases) where I build them elsewehere,
because another spots better fits my strategy
 
The blue circle I think represents the best settling site according to whats visible at the time (in terms of a balanced F/H/C). The more fog you uncover, you'll find the blue circle will change location.
 
carl corey said:
One more thing: sometimes the blue circle seems to put you in a position with not that many resources in the fat cross. That usually seems to mean that there are some new resources that are likely to appear there, like copper, iron or horses.

To me, I think this is the only really good part about the blue circles, otherwise they are just a distraction.
 
carl corey said:
One more thing: sometimes the blue circle seems to put you in a position with not that many resources in the fat cross. That usually seems to mean that there are some new resources that are likely to appear there, like copper, iron or horses.

In the unpatched version this was true, but this was fixed in an early patch.
 
I've never paid much attention to them. I have city spots picked out long before I build Settlers for them, so I typically don't even see them most of the time.
 
Beamup, I don't know about the others, but I'm talking more about the first city placement.

Adraeianus I, I really wonder how they set up those blue circles then. If it's from what you can see at a given time, why doesn't it change when you reveal more? As in the game I talked about, where I see as clearly better to have 4 resources then only 2, but the blue circle stays in the same place. I'm really puzzled.
 
carl corey said:
Beamup, I don't know about the others, but I'm talking more about the first city placement.
For the first city, you should never, ever, ever, ever, pay any attention to them at all. Instead, you should think about it very carefully. The blue circles don't use any more information than you have - they're just a shortcut to avoid thinking about it yourself. And where to place your first city may well be the most important decision in the entire game... you certainly shouldn't even consider taking shortcuts on that choice.
 
Beamup said:
For the first city, you should never, ever, ever, ever, pay any attention to them at all. Instead, you should think about it very carefully. The blue circles don't use any more information than you have - they're just a shortcut to avoid thinking about it yourself. And where to place your first city may well be the most important decision in the entire game... you certainly shouldn't even consider taking shortcuts on that choice.

Very well put. I've only started paying attention to my first city placement in my last 5-6 games, but it can indeed make a lot of difference. By the way, is there any way to turn off those indicators? Maybe the ones for worker suggestions as well?
 
I think there is, somewhere in the options screen. I certainly recall seeing something I thought would do that, but I never bothered to push it and I haven't looked since. But IIRC it's in there.
 
Artosoft said:
Does the blue circle for city location (for settler) is the best location according to computer calculation? Or do you need to think about the best location than blue circle one? How if you have more than 1 blue circle?

Regards,
Arto.

You may want to differentiate between the 'blue circle where your first turn starts' (at least, I don't recall my settler's first turn NOT starting on a blue circle), from the other 'blue circles' you see with later settlers.

Somebody who knows the SDK would know better, but I think the code that determines starting location of turn #1, takes into account things like where Iron/Copper/Horses are -- which is why settling right where you are will almost always mean you have one of them within the culture range of your capital.

The code that determines the location of other blue circles may not take that into account.

In any case, I always settle within 1 turn of where my settler starts. I feel that ensures that I'll still get whatever goodies my capital was supposed to get.

For other cities, I generally find that where I want to place them is often a blue circle. And if not, I do reanalyze my decision to make sure I didn't make a silly mistake. But I certainly don't blindly follow the blue circle.
 
Like most things in Civ, the more you know the game the less you need the computer's help.

My first couple of games I always chose the blue circle, now I use the circle as a guide, and as I continue to play and improve I will someday completely ignore them.
 
I got curious about this once, and actually read the code in the SDK (geeks, start with the AI_foundValue() function). It's a big hairy mess to understand, but here are my notes (perhaps comprehensible, perhaps not) on the algorithm:

Spoiler :
Code:
Abbreviations:
FXT = Fat X Tile(s) (e.g. any of the 21 tiles in the fat X, including the center tile)
FOP = Factor(s)	Of Production (e.g. food/hammers/commerce produced by a tile WITHOUT improvements)
fpt = Food Per Turn (number of bread slices of the tile WITH improvements)
hpt = Hammers Per Turn
cpt = Commerce Per Turn

AI_bonusVal (ABV) is computed as follows:

0 if bonus is obsolete
ABV is the sum of
	10 for each happiness provided by the tile
	10 for each health provided by the tile
For each unit type in the game, add a unit score (US):
	 4 if the unit requires ONLY this resource
	 3 if the unit requires this resource plus something else (another resource, building, etc.)
	 1 per 10% speedup if unit builds are sped by this resource
	 Double US if this civ can currently build the unit
	 Multiply US by 3/2 if the civ can't build it, but the required tech is in the era we're currently in
	 Otherwise, divide US by the number of eras left until the required tech
For each building type in the game, add a building score (BS):	 
	 3 if the building requires ONLY this resource
	 2 if the building requires this resource plus something else (another resource, building, etc.)
	 1 per 10% speedup if the building is sped by this resource
	 Double BS if this civ can currently build the building
	 Multiply BS by 3/2 if the civ can't build it, but the required tech is in the era we're currently in
	 Otherwise, divide BS by the number of eras left until the required tech 
For each project type in the game, add a project score (PS):	 
	 1 per 10% speedup if the project is sped by this resource
	 Double PS if this civ can currently build the project
	 Multiply PS by 3/2 if the civ can't build it, but the required tech is in the era we're currently in
	 Otherwise, divide PS by the number of eras left until the required tech 

AI_foundValue (AFV) is computed as follows:

0 if any of the following are true:
	* Within 2 tiles of another city
	* More than 7 FXT are owned by another civ, or off the edge of the map
	* More than 7 Bad Tiles (see below) in fat X AND no FXT has a Good Bonus (also see below)
		Bad Tile: 	Tile is impassible or off the edge of the map -> Bad Tile
					Tile has 0 food or <= 1 FOP -> Bad Tile
					Tile is water with <= 1 food -> 1/2 Bad Tile
		Good Bonus:	Bonus of a type the civ does not yet have, or only has through trade -> Good Bonus
					Bonus has AI_bonusVal > 10 (see XML) -> Good Bonus
	* More than 7 FXT are within fat X of another city
	* More than 1 (!) FXT is within fat X of a teammate's city
	
0 if we're looking for a start spot and any of the following are true
	* Tile has a goody hut
	* Any FXT are off the edge of the map

AFV = 1000 to start
For each FXT, add a FOP score (FS):
	If the tile has 2+ food, fs = 20*fpt + 80*hpt + 40*cpt
	If the tile has 1 food, fs = 10*fpt + 40*hpt + 20*cpt
	If the tile has 0 food, fs = 10*fpt + 20*hpt + 10*cpt
	Double FS for the center tile
	Multiply FS by 3/2 for the first ring of tiles
For each FXT, add/subtract 1 for each 10% health bonus/penalty (e.g. forest tile->+5)
For each FXT with a bonus:
	If we're looking for a start spot, add 10 + 20*AI_bonusVal
	If we're not, and the civ already has a bonus of that type connected (not through trade), add 10 + 20*AI_bonusVal
	If we're not, and the civ DOESN'T have a bonus of that type connected, add 10 + 80*AI_bonusVal
For each coastal tile in the fat X:
	Add 600
	If we're looking for a start spot, add 200 more
	If we're looking for a start spot, and we don't have a city on this continent yet, add 200 more
For each hills tile in the fat X, add 200
For each river tile in the fat X, add 60
For each tile with fresh water in the fat X, add 80
If we're looking for a start spot, add 10*fpt + 10*hpt + 5*cpt for each tile within distance 5 (yes, this counts the fat X again!)
If we're looking for a start spot, add 100*(distance from starting spot for this civ) (?!?!)
For Barbarians, subtract 200*(distance to nearest city within 8 plots)
For non-barbs, subtract 500*(distance to nearest city within 5 plots)
If this continent
	Is not yet settled (by anyone), double AFV
	Is settled only by our team, multiply AFV by 3/2
	Is settled only by our team and barbs, multiply AFV by 4/3
	Is none of the above, but has at least one city from our team, multiply AFV by 5/4
Divide AFV by max(0,(# Bad Tiles)-5) + 3
If the city tile has a bonus, divide AFV by 4
Divide AFV by max(0, (# FXT off our continent)-14) + 2

Whew!

The blue circles, as far as I can tell, are placed on tiles whose AI_foundValue is higher than the value in the adjoining 8 squares. Also, from cursory examination, AI settlers will go to the highest-scored tile they can currently access.

The most important points I took away are these:
  1. The value only reflects resources that you can see with current technology
  2. The value does NOT reflect possible improvements - only natural values and improvements in place
  3. The AI does NOT count food requirements to work tiles (e.g. as far as I can tell, a site with 8 iron mines and nothing but plains tiles would be fairly highly rated, even though you can't work more than one of the mines until Biology).
  4. Food and strategic resources are heavily favored.
  5. Happiness is no more valuable than health in the calculation
  6. No value is given for chop-value of forests, just the hpt it produces
  7. No consideration is given to future locations of other cities, nor does the AI try to block other civs' access to portions of a continent
  8. The AI actively AVOIDS settling close to other civs' borders
  9. Starting locations are valued based on all tiles within radius 5

There is also a bug in the calculations - it rates the center tile (where the city will be) using its actual f/h/c value, rather than the 2/1/1 (or better) it will have with a city on it.
 
Thank you all for comments.

I will decide where I am build the city by blue circle, but not always in the blue circle.

Thanks again.

Regards,
Arto.
 
Back
Top Bottom