# Ring City Placement

Discussion in 'Civ3 Strategy Articles' started by DaviddesJ, Jun 26, 2003.

1. ### DaviddesJDeity

Joined:
May 5, 2003
Messages:
4,349
Location:
Burlingame CA
Ring City Placement (RCP) is a new approach to city placement in your empire. Other placement strategies, like OCP, are based on land use patterns. RCP is based on a different principle: reducing corruption.

Thanks to alexman and this thread Do you think you understand corruption?, we know that corruption has two sources: corruption due to the distance from the capital or forbidden palace, and corruption due to the number of cities.

There is a natural tension between these two: you can achieve low corruption of the first kind by placing many cities very close to your capital (a la ICS), but this means that if you want to exploit all of your land, you need a lot of cities, and the corruption of the second kind is severe. On the other hand, if you try to work all of your land with relatively few cities, to minimize corruption of the second kind, then the cities are spread out and far away.

Alexman discovered that corruption of the second kind depends on the number of cities that are closer to your capital or forbidden palace than a given city. When I read this, after a while, a question occurred to me: what happens if two cities are at the same distance from the capital or FP? This called for some tests, but the answer is that only cities that are closer count for corruption purposes; cities at the same distance don't increase corruption in other cities at that distance.

There's a way to take advantage of this: build several cities at exactly the same distance from the capital. I call this Ring City Placement, because the cities at the same distance form a "ring" on the map.

It's important to understand how the rings work, which has been explained in detail in the alexman thread.

This method of calculation is unintuitive at first, but with practice, it becomes easy to "see" the rings at a given distance. The Firaxis distance is always an integer, or half an integer, so there are many more city sites at the "same distance" than you might first expect. For example, there are 8 city sites at distance 3 from the capital (but some of these might be unbuildable mountains or water, of course).

[TODO: insert illustration of rings and distance calculation]

Here's a chart for various distances, of how many sites are at that distance, and how many cities can potentially be built at that distance. (Sometimes two sites at the same distance are adjacent, so only one of the two can be built.)

Code:
```Distance    Sites        Cities
2.0         4             4
2.5         8             4
3.0         8             8
3.5         8             8
4.0         12            8
4.5         12            12
5.0         12            12
5.5         16            12
6.0         16            16
6.5         16            16
7.0         20            16
7.5         20            20
8.0         20            20
8.5         24            20
9.0         24            24
9.5         24            24
10.0        28            24
etc.
```
As should be clear, RCP is not a single fixed rule governing the exact locations of cities. The player can choose the distance of the rings, and also to decide whether to build on some or all of the potential city sites in each ring.

One natural approach is to build an "inner ring" at distance 3-4, and another "outer ring" at distance 6-9. I call this 2RCP, because there are two rings. You get several inner cities with extremely low corruption (often around 10% or less), and a whole lot of outer cities with manageable corruption (well under 50%).

But there are other alternatives. You might build on some but not all of the sites in the inner ring, in order to reduce corruption in the outer ring (because each inner ring city increases corruption in every outer ring city). (Call this 1.5RCP.) Or you might decide not to build an inner ring at all, and just build a single outer ring of cities at distance 6-9, which could have as many as 20 cities with as little as 20% corruption in all of them. (Call this alternative 1RCP.) You might also make different choices at different points in the game: you could start the game with an inner ring, but once the outer ring is fully developed, abandon some or all of the cities in the inner ring.

There are also a lot of open questions, like how to combine RCP with Forbidden Palace placement, or with a Palace Jump. Clearly you can build one set of rings around your Palace and another set of rings around your Forbidden Palace, but they have to be far apart, or else the rings will interfere with each other. Are there other ways to use the Palace and FP? I'm not sure. I'm definitely still learning, myself. I also know that some people have already been experimenting with RCP, since I mentioned it in the GOTM20 Spoiler #1 thread, and I hope they will post their observations here, for a larger audience. I'm personally applying the RCP idea in GOTM20, and I hope to post detailed illustrations and examples from that game, once the GOTM period is over. (But first I have to finish!)

I do want to especially congratulate alexman for all of his work on corruption, because if he hadn't worked out the basic mechanisms, I never would have had the idea for RCP. I also really look forward to other players finding ways to use the RCP idea, which I may not have thought of, and I hope they will share those in this thread.

2. ### MoonsingerSettlerRetired Moderator

Joined:
Apr 22, 2002
Messages:
4,374
Location:
Iowa
Very good research! Thank you for sharing them. Do you have a sample screenshot of your RCP?

3. ### TxurceDeity

Joined:
Jan 4, 2002
Messages:
8,259
Location:
Venice, California
Daviddesj, congratulations for coming up with an original concept. This is a field rife for exploration and graduate research. In theory, RCP is a very powerful, although not unbalancing, approach to expansion. What I wonder at first blush is how applicable it is in actual game conditions. The following questions come to mind (including ones you raised):

1. Corruption is equal in all properly-placed ring cities. But what happens when you build a city that breaks the pattern? This is likely to happen due to geographical constraints, or the requirements of a settler factory, for example. Would such a city suffer disproportionate corruption if outside the first ring, or greatly dilute the value of that ring if inside it?

2. If the effect of breaking the pattern is serious, then the FP would need to be placed far enough away to have no effect on the outermost ring of the palace. This is probably ideal, anyway, although hard to achieve in practice.

3. Jumping the palace seemingly destroys the entire point and benefit of RCP. On the other hand, by the time you're ready to jump, you may be able to do without RCP's benefits.

I'm sure you've either considered or encountered all of these questions in GOTM20, and am very curious to hear what the various effects were.

4. ### MoonsingerSettlerRetired Moderator

Joined:
Apr 22, 2002
Messages:
4,374
Location:
Iowa
Question:

According to this picture (see attachment of this post), what is the distant of B, F, and H?

Are the following calculation correct?

A = 2
C = 3
D = 5 ???
F = 4 ???

Note: I am really interesting in the correct distance of D and F. From the look of it, D seems to be a lot closer than F.

5. ### DaveMcWDeity

Joined:
Oct 8, 2002
Messages:
6,489
Here are some RCP screenshots. You don't have to use all the locations, but each one you do use gets 2nd city corruption.

6. ### MoonsingerSettlerRetired Moderator

Joined:
Apr 22, 2002
Messages:
4,374
Location:
Iowa
Since There are 6 possible sites with the radius of 2, what I don't understand is why DaviddesJ mentioned only 4 of them. What is wrong with the fifth one and the sixth one?

7. ### Cartouche BeeAppropriations Consultant

Joined:
Jan 20, 2002
Messages:
1,731
Location:
Multiple Victory Avenue
It's about time that someone formalized this concept. I've been doing symmetrical builds for quite some time to leverage off the corruption calculations. Now I know to keep an even tighter reign on placing and positioning.

 I had a game as the Koreans lately that I posted a shot of the build, I affectionately called it the Korean Death Star build out.

8. ### jeffelammarSpace for rent

Joined:
Aug 7, 2002
Messages:
913
Location:
Unless I am Gravely Mistaken...
Distances are from Paris
A = 2.5
B=3.5
C=3.5
D=4
E=4
F=4.5
g=4.5
h=5

[EDIT] I have been working on a program to calculate the comparative advantaces of the different ring placements. I'll try to post something on this tonight.

9. ### MoonsingerSettlerRetired Moderator

Joined:
Apr 22, 2002
Messages:
4,374
Location:
Iowa
DaveMcW & DaviddesJ,

Please confirm! For some strange reason I think the distance of "A" from Paris is equal to 2 not 2.5. That's really the source of my confusion.

10. ### colOld FartRetired Moderator

Joined:
Mar 12, 2002
Messages:
5,812
Location:
Sunny Scarborough
There are only 4, Moonsinger, at a distance of 2.

If we take x and y as NE/SW and NW/SE.

There are 4 cities with coordinates of 2,0 0,2 -2,0 and 0,-2. The cities with coordinates of 2,2 are in fact 3 squares away on the algorithm, used to calculate distances. (=big number + 0,5*small number). Cities at coordiniates of 1,1 are in fact 1.5 squares away. Its weird at first but easy to figure after a while.

At a distance of 3 we can have 8 cities
4 at 3,0 etc and 4 at 2,2 etc. Unfortuantely some are adjacent so we cant build them all.

I've tried a few sample games but the price you have to pay is often building on bonus tiles and losing the bonus / mountain or sea in the wrong place. I like rings at 4 and 8 but 3 and 6 can work well too.

Utlimately you still have to decide whether that cost is worth the gain in efficiency due to lower corruption.

In your diagram A is (2,1) so the distance =2.5 = (2 + 0.5*1)

11. ### MoonsingerSettlerRetired Moderator

Joined:
Apr 22, 2002
Messages:
4,374
Location:
Iowa
Thank you col! In that case, according to my above map, Paris E, Paris N, and Paris NW are all at the distance of 2.5 instead of 2. Oh well, I need more practice.

12. ### ilovetoastChieftain

Joined:
Jun 22, 2003
Messages:
7
Sure looks like 2.5 to me...

By my reading you choose the shortest movement path which....
For getting to A form Paris is either SE + S or S + SE.....
And then you calculate:
N, S, W, E = 1.0 each
NW, NE, SW, SE = 1.5 each
For totals of (1.5 + 1.0) or (1.0 + 1.5) .... 2.5 either way you cut it.

Just my take...

FWIW...I get the same numbers as jeffelammar for the A-H points when I count.

13. ### WillJCoolness Connoisseur

Joined:
Aug 9, 2002
Messages:
9,471
Location:
USA
Great stuff, DaviddesJ! But I have a feeling the other factors in city placement (f/s/c values of the surrounding terrain, being on coasts instead of one tile away, being on rivers, taking advantage of the AI borders, not bothering reposistioning captured cities, etc.) might top corruption and waste in importanace. I guess I'll have to wait and see until GOTM 20 is over. I can't wait!

14. ### ilovetoastChieftain

Joined:
Jun 22, 2003
Messages:
7
I am bored. So I made color coded rings 1-5:

File size:
63 KB
Views:
10,569
15. ### BamspeedyWe'll dig up the road!

Joined:
Dec 18, 2001
Messages:
7,780
Location:
Amish Country, Wisconsin, USA
Wow! Thanks for the screenshot with color coding, ilovetoast!. Looks like a spacing of 5 would be good for the 'builders'. It offers twice the number of cities as OCP, but would probably be about the same amount of corruption, and certainly be more powerful well before OCP is.

16. ### alexmanKing

Joined:
Feb 28, 2002
Messages:
778
Location:
Maryland, USA
DaviddesJ, your RCP idea is ingenious! It should prove to be a powerful strategy for the early game, probably becoming popular in multiplayer and higher difficulty levels where a strong start is more important.

However, as Txurce observed, the Palace move, (essential to happen ASAP for boosting your economy) would mess up the RCP strategy, so you would have to wait for a leader to build the FP instead. Just make sure there are more non-ring cities closer to the FP than the number of city-members of your Palace rings.

The main benefit of your idea for my own gameplay will probably be to adopt a hybrid approach, being concious of the opportunity to place two or more cities at equal distances from the capital so they share the lower corruption.

By the way, thanks for all the credit, but I don't deserve it. The corruption article has evolved over time, with contributions of many members of the community, including you.

17. ### TxurceDeity

Joined:
Jan 4, 2002
Messages:
8,259
Location:
Venice, California
Alexman, does a hybrid approach work? By this I mean, could you create a first ring, a partial second, and then chuck the whole thing as you move out? In this scenario, do the non-ring cities pay a disproportionate price in increased corruption, or just the standard for the number of cities and distance from the capital?

18. ### alexmanKing

Joined:
Feb 28, 2002
Messages:
778
Location:
Maryland, USA
Here's an example.

Empire A has 4 cities in addition to the capital, all at different distances from the capital. These cities all have different OCN corruption corresponding to ranks #1, #2, #3, and #4.

Empire B has the same number of cities, but the second and third cities are equidistant from the capital. The OCN corruption of the cities in this case would use ranks #1, #2, #2, and #4.

19. ### jeffelammarSpace for rent

Joined:
Aug 7, 2002
Messages:
913
Location:
RCP does not eliminate the Palace jump strategy. Since you have to plan a Palace jump anyway, you could easily build your initial cities with the ring centered on a different city than the original. You wouldn't see the corruption benefit till the palace jump happens, but you would be rocking the world when it did happen.

20. ### DaviddesJDeity

Joined:
May 5, 2003
Messages:
4,349
Location:
Burlingame CA
It certainly seems possible to plan to build a very early FP, build the initial rings centered on the FP city, and then jump the palace away. Even before the palace moves, the FP is working to generate low corruption in the rings around the FP. However, the FP city itself will probably not be the closest to your capital, thus may have some extra corruption which may delay the FP somewhat. And during the important early turns (before you can build the FP), you won't be getting the benefit of the RCP approach.

The second alternative is of course to just build the rings around the palace, and plan to build the FP far away, with a leader.

Whether there's another idea, perhaps involving overlapping rings that somehow don't interfere with each other too much (e.g., build the FP in one of the rings, and then add rings around the FP, and get lots of low-corruption cities without ever jumping the palace at all), I'm not sure.