How far cities should be from each other?

[Renata]

Hmmmm ... A strict 3-tile spacing (in my sense) would be something like this.

XXAXXBXXC
XXXXXXXXX
XXXXXXXXX
XXDXXEXXF
XXXXXXXXX
XXXXXXXXX

54 tiles/6 cities = 9 tiles per city (i.e. cities could grow to size 8 before having to hire specialists with any excess food. In retrospect this *is* significantly tighter than I prefer. (Although once you factor in coast tiles and borders with AIs and such the tile average would go up a bit, maybe to 10 or 11.)


Here's a strict 4-tile spacing (by my definition):

XXXXXXXXXXXX
XXAXXXBXXXCX
XXXXXXXXXXXX
XXXXXXXXXXXX
XXXXXXXXXXXX
XXDXXXEXXXFX
XXXXXXXXXXXX
XXXXXXXXXXXX

108 tiles for 6 cities = 18 tiles per city (cities could grow to size 17 before specialists).

This is closer to my preference for my own cities, perhaps a little looser than I would normally go. (As you can see it includes wasted tiles (tiles in no city radius), one of the things I dislike. Yes a perfect OCP gets around this problem, but I'm not at all interested in trying to play my cities to a perfect grid.) I'm usually happy when my cities max out at about size 16, on average. (With a few of the less productive ones never getting hospitals and few of the really good ones getting to work all their potential tiles for size 20.)

edit: I've gone for spaceship victory at every level through monarch so far and never had the slightest problem with spaceship parts or the Apollo program. 60 shields is plenty to build most of the spaceship parts in a timely fashion, and getting 4 or 5 good size 16 cities to that level isn't a problem, especially if you go for nuclear power early enough to build a few nuclear plants. Get even one size 20 city with double that number of shields and you're in business. In GOTM18 I built all 10 spaceship parts in the last 11 turns of my game, and I only had one city that ever made it over 100 spt.

Renata

 

[newbiebad]

I generally assumed the spacing is measured by the distance between two cities. I can't see how many of these other players could bare to live with sub-10 cities so I'm pretty sure that's the standard. Using your defination, I'd probably prefer the 4 city spacing. There becomes a point that such tight city packing creates too much much corruption by completely overwhelming the OCN. Some otherwise useful cities just outside your core might be reduced to 3rd rate cities.

 

[Sirp]

Yeah I generally count spacing by how many tiles are *between* the cities.

So CxxxxC is spaced "4 tiles apart". Anyhow this is a typical "off-by-one" problem

"Ideally" cities are placed in a perfect grid, but in reality one has to take into account numerous factors such as availability of resources, proximity of resources, the need to control land before one can build a temple or library, the shape of the coastline, the presence of mountains (which you can't settle on), the location of other civs, proximity to the capital, and so forth.

-Sirp.

 

[Strider]

It depends for me... If there is something (like a lux or resource) that I REALLY want to get, but won't get doing the "normal" tile vs tile distance... Then I'll do it... I can always use culture to close the bordors after all.

 

[theGhost]

wasnt a strategy at the beginning of the game to expand outwards in....to prevent other civs from building deep into your territory? So what is it now: build close together to join your cities (tightly packed) or spread out to claim land and attack anyone else on your turf?

 

[kaslks75]

...another benefit of building cities in a somewhat 'denser' pattern is defense. If a city gets into trouble, a tighter city layout allows for a faster 'rescue' with units from surrounding cities.


...which brings up another interesting question... Consider the following:

xxCxx
xxxxx
_xxx_
xxxxx
xxCxx

where x = tiles
C = city

How does the game decide which city controls the 3 overlapping tiles in the center?? When one city is in 'control' of a particular tile, can you change this to the other city somehow??

 

[Renata]

You mean when you have two cities sharing tiles? It's just a matter of which city gets there first. You can change tile assignments on the city screen. Open up the city screen, and click on the food/shield icons showing on the tile you don't want to work any more. That'll make an entertainer in your city. Then click on the tile you want that citizen to work instead and your entertainer will go back to work on the new tile. Sometimes you can't do it in one step: if two cities were sharing two tiles and you wanted to switch which city was working which, you'd have to make an entertainer in one city, go the city screen for the other city and move the citizen, then go back to the first and put the entertainer back to work.

Renata

 

[kaslks75]

Thanx Renata, I thought that maybe they went to the 'older' city. I could try it and see for myself, but often I'm at work when I'm on CivFanatics and I just gotta know NOW! (no - my boss doesn't care!)

 

[Procifica]

Personally, what I do is build about 3 to 4 rings around my capital city, forbidden palace (near the capital), and where I will move my capital; with the optional city placment method (or rather, my slight variation of it). What this is, is I build cities exactly 5 tiles inbetween either up and down or across. Which means: AXXXXXB. Like that.

Then, the opposite direction, I build 8 tiles inbetween. This results in a nice, compact grid where maximum tiles are used with very little overlap.

Outside of this ring, I still place them down like that, to maximize territory. After most of the land is filled out in this manner, I go back and start plopping cities in culture holes (areas where your culture has like say 4 squares unfilled as an example). This is to maximize territory. Over the long-term, true ICS hurts your score because each city square only produces 2 food. True ICS though, works good in Desert and Tundra areas.

I've got two games on Chieftain skill level with projected end-game scores in the 6K to 7K range (#2 and #3 on HoF on this site currently if submitted and they qualify), using this type of strategy. It combines the best of both OCP and ICS.

 

[TheNiceOne]

It obvious that there are several factors that decides the city spacing. For some, the most important is effiency, but aestethics or role playing may also impact it.

When it comes to effiency, it is no doubt that a 4 tile spacing is a waste. You waste close to half of your tiles until after hospitals and you waste movement. Being able to move a defender from one city to the next in one turn (whihc means max 3 tiles apart) may save a city from a surprise attack.

It seems that inexperienced players are afraid of overlapping the cities, but there's no reason to be afraid of this. I think a medium overlap, where each city has the city tile plus 12 squares for itself is a very good solution. Not only for early warmongers, as Sirp says, but also for builders. A builder that packs the cities this way will typically generate 50% more income (= reasearch etc.) when all cities are size 12, than a builder who gives each city 21 tiles.

To give each city 13 squares, the patter is 3+2, i.e., you move your settler 3 spaces horizontally and 2 vertically or vice versa. Another beauty with this moderate overlap, is that it doesn't matter much if you have to move the city a tile or two from the theoretical spot, since it simply means that the citizens will work on more tiles on one side of the city.

On the lower levels, you don't lose much by giving each city 21 tiles, but one the higher levels and vs. good humans, you need a denser spacing to be competetive.

 

[Procifica]

Well, the 21 tiles (19 or 20 is actually what occurs with my spacing method) is crucial if you wish to max out score, at least for the first couple rings around your palace/forbidden palace.

Even on higher skill levels, where the AI expands fast, this method enables you to get as much territory as possible, which can be crucial.

 

[TheNiceOne]

Originally posted by Procifica
Well, the 21 tiles (19 or 20 is actually what occurs with my spacing method) is crucial if you wish to max out score, at least for the first couple rings around your palace/forbidden palace.

Even on higher skill levels, where the AI expands fast, this method enables you to get as much territory as possible, which can be crucial.

I think you're wrong with both points.

Score:
With 13 tiles to a city, I will have a much higher popoulation than you before hospitals (after the cities have maxed out) simply because I'll have more cities. On a land with 273 (13*21) squares, I will have 12*21 = 252 citizens instead of your 12*13 = 156 citizens. 61% higher population will increase my score.

After hospitals, my 13 tiles cities will still give me better scores, since the total number of citizens will be equal to yours, but more of mine will be happy. The latter is actually a big point. Happiness is much easier to maintain when you have many mid-sized cities than few large cities.

So why do you need cities with 21 tiles to max out score?


Expansion:
You will actually expand slower if you give each city 21 tiles. This is partly because each settler will have to move for a turn or three more for each city, and because bigger cities grows slower. Another huge benefit of the eraly overlap is that two cities can share the tiles than workers get to improve:
Ex: Your capital has two cows, which are the two first tiles you mine and road. When you build a settler at size 3, the population goes down to 1, leaving one cow tile unused. If the neighbour city also can use this tile, then you get much better and effective use of your best squares and early worker actions.
Also, note that early corruption will be less, both due to closer cities and less turns to build roads between the cities.

So even if each new city will cover less land, I will build cities faster to more than catch up your wider spacing.


There is one benefit of giving some cities 21 tiles, and that is easier wonder building, since those 21-tile cities will have higher production. But since there are very few wonders after hospitals, this should be easy to acquire by smart prebuilding anyway.

 

[dragon.jade]

Fewer wonders? Yes but SS parts also can be build.... And those definitely benefits from a 21 squares city output.

Just 2 cents of mine...

Simply,
Dragon.Jade ;-D

 

[anarres]

dragon.jade,

If you are needing to win by that close a margin in the Modern Age the problem is that you are using size 21 cities in the first place

The whole point of using a denser build is that comparatively you can leave the competition standing and age or two earlier.

The Nice One makes a very good point about micromanagement, it is essential to playing and winning well on the higher levels, and against good humans.


theGhost,

The 'outside in' strategy has been shown to be flawed numerous times, as has 'Optimal City Placement' (getting 21 tiles for each city). These strats work fine on lower levels where the AI won't punish you for them, but they are harder and harder to play with the higher you go.

 

[Procifica]

If you notice though, I said its better for the first few rings of cities. It means more tiles with less corruption, which means more science and income. Larger cities aren't hard to keep happy when you have all 8 luxuries. Also, each city tile only produces 2 food, which means if you put one on grassland, you're losing 2 food (1 population point) which OCP would allow you to have.

The argument I've seen is ICS gives a higher score earlier on, but eventually OCP can catch up to ICS.

I personally think a mix of the two is the way to go, but that's my opinion. My france HoF game is a good mix of the two, and considering its my second attempt at the Chieftain #1 score, and I haven't played Civ3 much (but a Civ2 veteran), I think I've done really good and if I had a chance to redo the map over, I'd have even more points because of little things which I didn't do which could have maxed out my score better.

I agree 110% that micromanagement is EXTREMELY important to get a good score or a win, on any skill level.

 

[theGhost]

no wonder....i did it successfully once cutting them off at a relatively large chokepoint and then building in. I tried it again on the next difficulty setting and it went downhill

 

[TheNiceOne]

Originally posted by dragon.jade
Fewer wonders? Yes but SS parts also can be build.... And those definitely benefits from a 21 squares city output.

It is very easy to prebuild those SS parts by using solar plants as prebuild items. You just have to plan better and start prebuilding earlier with 13 squares for each city.

...and as anarres writes, having a denser build will probably hand the game to you much earlier

 

[TheNiceOne]

Originally posted by Procifica
If you notice though, I said its better for the first few rings of cities. It means more tiles with less corruption, which means more science and income.

I agree that it is slightly better for the first ring, from somewhere around the middle of the industrial age when the cities have grown. But I maintain that it is much worse from the start of the game, through the two first ages and the start of the third age. So at best, you will catch up a little bit of what you've lost through most of the game.

Larger cities aren't hard to keep happy when you have all 8 luxuries. Also, each city tile only produces 2 food, which means if you put one on grassland, you're losing 2 food (1 population point) which OCP would allow you to have.

If you have access to all 8 luxuries, then you're either playing on a very easy difficulty setting, or you're paying lots of money. On emperor and deity, I'm very happy if I can get my hand on 4-5 luxuries. And I can certainly not expect to have all 8 without paying 50+ gpt for it. With size 12 cities, you can stop at 6 luxuries and still have better happiness than with 21 cities/8 luxuries.

Granted, you loose two food on those additional city squares, but if you gain a few shields from the square, which you wouldn't get if you had irrigated it, so that's about even.

The argument I've seen is ICS gives a higher score earlier on, but eventually OCP can catch up to ICS.

In a previous post in this thread I explained why this isn't true - even after the OCP cities have reached size 21+, a denser model will continue to score more higher.... and I've not seen any counter arguments.

I personally think a mix of the two is the way to go, but that's my opinion. My france HoF game is a good mix of the two, and considering its my second attempt at the Chieftain #1 score, and I haven't played Civ3 much (but a Civ2 veteran), I think I've done really good and if I had a chance to redo the map over, I'd have even more points because of little things which I didn't do which could have maxed out my score better.

I agree 110% that micromanagement is EXTREMELY important to get a good score or a win, on any skill level.

I used OCP for my first 5-10 games as well, before I started to appreciate the gain by a denser build. A mix is certainly better than a pure OCP, but by giving all cities around your capital/FP 21 squares, you waste almost half of your potentially most productive tiles for most of your game. Those 9 unused squares for each city would be close to corruption free if you had used them, and would boost your income and production massively. There is no way this loss can be compensated for (except playing on a low difficulty level)

 

[Sirp]

TheNiceOne: Yeah dense builds can be good for builders too, but the advantage is most noticeable when warmongering.

I basically use a 4-tile spacing, however getting things like fresh water access is far more important than worrying about your spacing model.

I think about 16 tiles per city like Renata said is reasonable. That's generally what I end up doing. Most cities have at least a few tiles in their radius that are mediocre, and don't make that much difference if they're not worked until late.

Desert tiles are close to useless until rails come along anyway.

-Sirp.

 

[col]

I think the OCP / ICS debate runs deeper than mathematics. I dont think if you do the sums there is much question about which is more successful at utilising resources more efficiently.

Some people LIKE to see big cites thriving and producing massive numbers of shields and beakers. Players who like to build tend to almost always prefer OCP. Its almost SimCity like to SEE the power of a large city. They like to line up their cities neatly in pretty patterns.

Pragmatists - who often end up as warmongers or opportunists - tend to favour ICS - or a 'tight build' as some call it. They perhaps subconsciously look to SEE the hoards of cities spreading and dominating the map.

Freud would probably be able to explain it.

I'm a tight builder usually. I just cant stand to see all those good tiles go to waste in the first half of the game - and my experience of human to human play suggests that OCP players dont last too long.