City Placement, meaningless?

[badams52]

While I was writing a reply to the creat thread about city placement started by Cumi, I got to thinking that all this discussion about city placement in the sense of winning at Civ III (SP) truly makes little difference in the final outcome as long as the rest of your strategy is sound.

To give a salient analogy, (for those of you who play/know baseball) there's a group of people who believe that a team's batting order makes little difference in the outcome of a game and amounts to only a few runs over the course of a 162-game season. More important is who bats.

Taking this idea to civ, it seems to me, and I could be wrong, but given a set amount of land area, the exact placement of your cities, including # of cities, does little to determine your overall power and production, as long as you are working all the tiles within the given area.

If my theory is correct, then city placement, though very fun to argue and analyze, has little to no effect on winning or losing a game of civ.

Maybe if there's enough interest, someone could test this theory...

[vmxa1]

I agree with your hypothesis on city placement.
You are correct about the extra runs in baseball, but the thing that needs to be added is that you will get one extra at bat for your 1-3 hitters in many games. This should payoff. I always batted clean up until I realized that I was not getting enough at bats. I toyed with movng to lead off, but I would not get enough RBI opportunities, so I settled on 3rd. Optimal AB and RBI, this is what you want for you top hitter, if they are a power hitter. Clean is reserved for those that can only swing for the fence and not be relied upon to get a high average. This maximizes their chance to payoff, as they willl have the most chance to come to the plate with runners in scoring position.
BTW, I no longer follow baseball since the strike about two strikes ago.

[Catt]

Originally posted by badams52
Taking this idea to civ, it seems to me, and I could be wrong, but given a set amount of land area, the exact placement of your cities, including # of cities, does little to determine your overall power and production, as long as you are working all the tiles within the given area.

If my theory is correct, then city placement, though very fun to argue and analyze, has little to no effect on winning or losing a game of civ.

I think your hypothesis might hold up in a scenario with a pre-made empire. But in a game that progresses from a start of only 1 settler and 1 worker (maybe a scout too), early city placement decisions seem to me to be very important. A snapshot comparison in 1500 AD does not convey the relative power, over time, of the development choices made previously.

Imagine a hypothetical empire that contains 7 cities. The capitol is on "decent" terrain. The three western cities are on fabulous terrain -- lots of rivers, cows, etc. The three eastern cities are on terrible terrain - lots of desert. The growth of the empire as a whole is radically different if the three western cities are settled before or after the three eastern cities. Even if the end result empire is an identical 7 cities, at any given point in the game one would assume that the empire which settled westerly first would be stronger and more advanced than the city that settled easterly first (cities -- i.e. production centers -- come online faster if better terrain is settled before weaker terrain).

If the above makes sense, extrapolate that to a city placing scheme that (1) ignores the fact that workable tiles will expand only with the investment of shields in cultural buildings, (2) settling on a river gives a free 100-shield growth facilitator (aqueduct), and (3) settling according to a grid, but ignoring specific terrain factors, means that some tiles may not be worked, even though an empire with the same outside boundaries but different city placement will be able to utilize the same tiles (best example that comes to mind is settling near mountain ranges -- how to make sure that the mountains get worked in the long run).

It seems to me that you're setting up a pretty tough hypothesis -- for me, taking it to its extreme, your hypothesis rests on the mistaken view that a shield of production power (or one gold piece) in 1500 AD is of equal value to a shield of productive power (or one gold) in 1500 BC. It's hard to say how much more valuable the early resource is to the later, but I think it's tough to make the case that there isn't a difference in value, biased towards early resources.

Catt

[nbarclay]

Originally posted by badams52
as long as you are working all the tiles within the given area.

That's the key, but if cities are spaced too loosely, you don't get to work all the tiles in the area until Sanitation. Similarly, if you don't take maximum advantage of coast, the area you work is smaller (at least in terms of ability to generate commerce). I follow no particular geometric pattern in placing my cities, but I like making sure few tiles (especially land tiles) are wasted in my original lands when my cities reach size 12.

In MP, geometry can potentially be a little more important so cities can support each other militarily. But making maximum use of the land (including being able to use the best tiles early) is still king.

[Master Zen]

ln Emperor/Deity, city placement is VERY important in the early rexing phase. A poor city placement will leave you at a very large disadvantage to hopefull beat the AI after it expands.

Try some games using popular placement styles like 3-tile or Ralphing and you'll see that it is much more successful than using no placement style whatsoever.

[Artifex]

I got a question what do you do on Emperor level when you capture an empire like the Russians that have good cities but they are all placed loose with 4 tiles between each city. Just keep them or place cities in between them? Or can you just keep them as is?

[nbarclay]

The closer you get to Sanitation, the less sense it makes to add cities unless they take advantage of tiles the conquered AI wasted. When REXing or conquering territory early in the ancient era, a relatively tight spacing can make a lot of sense because you have a long time to take advantage of the tight spacing before hospitals shift the advantage to looser spacings. But as the game progresses, you have less and less time to take advantage of tighter spacing before hospitals shift the advantage in favor of a looser spacing.

Personally, I rarely build additional cities in a conquered area unless I'm trying to take advantage of tiles the previous owner wasted. You can build more than that if you want to, but you certainly shouldn't feel like you have to.

Nathan

[cumi]

Originally posted by badams52
If my theory is correct, then city placement, though very fun to argue and analyze, has little to no effect on winning or losing a game of civ.

Maybe if there's enough interest, someone could test this theory...

I think every single turn has an effect on winning or loosing the game. If there is a formula or a "perfect strategy" the game would not be interesting.

I was playing Warcraft II before a 4-5 years and there was a perfect or optimal strategy for the beginning of the game (I think, for the first 10 steps or so), like:

1.) Build a peasant, send him to mine a gold.
2.) again.
3.) with the third start to build a farm
4.) send the fourth to harvest trees
5.) start building a barrack with the 2 peasant e.t.c

If someone did not played like these steps describe, were weaker after the same number of steps.

For Civ3 I still didn't found anything like this. And I was searching a lot! Now I am beginning to think, that there is no optimal strategy on the beginning...(like warior, warior, settler, wonder building order in your first city.

The reason, why I think the beginning of the game is very important factor for winning is, that I think this game has a kind of "exponensial" game curve. So if you have a weaker start, later you will be much more weaker then others, unless some miracle happens.

So I think a city placement is one of the effects, that decides the beginning of the game. And the beginning of the game decides, what will happen later.

Let's take a look at the city placement again:

if between your cities are 5 tiles (insted of for example 3), we can easily calculate how much more turns your wrokers need to connect two cities. If you continue like this, you will have to build additional workers. If so, you will be able to start building a wonder a few turns later, in the city, that has a 1 pop smaller size. You have to wait with building a settler also couple of turns. Later, your wonder is a few turns late comparing to your opponents. You have to forget more and more wonders. The avalanche is started and you feel you are loosing a game....

This was a very simple (stupid) example, how you can loose a game because of city placement. I think the CP is not soooo important, but stil can have an effect on your game...

Have a nice day,

cheers

cumi

[vmxa1]

Working over here boss.
Da Boo

[badams52]

Originally posted by Master Zen

ln Emperor/Deity, city placement is VERY important in the early rexing phase. A poor city placement will leave you at a very large disadvantage to hopefull beat the AI after it expands.

Try some games using popular placement styles like 3-tile or Ralphing and you'll see that it is much more successful than using no placement style whatsoever.

Hmmm, I was thinking between different styles, such as OCP, 3-Tile, 4-tile, but now that you mention it, as long as all tiles within the borders are worked I could make a very asymetric or unaesthetic style of city placement and have similar production values as well.

Originally posted by Catt

Imagine a hypothetical empire that contains 7 cities. The capitol is on "decent" terrain. The three western cities are on fabulous terrain -- lots of rivers, cows, etc. The three eastern cities are on terrible terrain - lots of desert. The growth of the empire as a whole is radically different if the three western cities are settled before or after the three eastern cities. Even if the end result empire is an identical 7 cities, at any given point in the game one would assume that the empire which settled westerly first would be stronger and more advanced than the city that settled easterly first (cities -- i.e. production centers -- come online faster if better terrain is settled before weaker terrain).

Actually, when I thought up the theory, I wasn't thinking in terms of proper vs. improper REXing. I think that if you REX correctly and choose the best city places for 3-tile or OCP, then your net result after hospitals will be similar.

I emphasize after hospitals cause in OCP, you won't be working as much land at 3-tile till you reach hospitals.

Part of the reason I think of it this way is that (as Aeson proved) games are winable even with absolutely lousy starting locations. What makes this so is the AIs inability to defend it's territory well, and going to war, though more difficult with OCP, will in most cases net you with more land.

Actually that's a big assumption of my theory that I failed to mention, that defeciencies in land and city placement are easily rectified by building armies and taking cities from your hapless victims. I call it the great equalizer: war.

I've had starting places much worse than the AI, and even though the AI had more and better cities, I still conquered his empire. If the AI knew how to attack better, how to build a better army, or If I was playing a human, then I should have had no chance, but as it happened, he was dead from the beginning, even after he absorbed the closest empire to himself.

[badams52]

Originally posted by nbarclay

That's the key, but if cities are spaced too loosely, you don't get to work all the tiles in the area until Sanitation. Similarly, if you don't take maximum advantage of coast, the area you work is smaller (at least in terms of ability to generate commerce). I follow no particular geometric pattern in placing my cities, but I like making sure few tiles (especially land tiles) are wasted in my original lands when my cities reach size 12.

Hmm this is certainly true. If you had the same land, but not the same workable tiles, then the productions wouldn't be similar given the one with more workable land (more cities) would our produce the other. I first was thinking of this since I noticed that when I had similar land when I tested Ralphing v. 3-Tile (which both work all available land before hospitals) that the total shield and gold production from both empires was almost identical.

In MP, geometry can potentially be a little more important so cities can support each other militarily. But making maximum use of the land (including being able to use the best tiles early) is still king.

Actaully, I purposefully mentioned this for SP only since for many people employing the strats found here, winning an SP game is almost a foregone conclusion because of easy to win wars, tech whoring, pre-buiding wonders, and all the advantages the player has over the AI.

But in MP, I think REXing and city placement will really make a large difference in the outcome of the game cause your advantages over the AI are gone.

[badams52]

Originally posted by vmxa1
I agree with your hypothesis on city placement.
You are correct about the extra runs in baseball, but the thing that needs to be added is that you will get one extra at bat for your 1-3 hitters in many games. This should payoff. I always batted clean up until I realized that I was not getting enough at bats. I toyed with movng to lead off, but I would not get enough RBI opportunities, so I settled on 3rd. Optimal AB and RBI, this is what you want for you top hitter, if they are a power hitter. Clean is reserved for those that can only swing for the fence and not be relied upon to get a high average. This maximizes their chance to payoff, as they willl have the most chance to come to the plate with runners in scoring position.
BTW, I no longer follow baseball since the strike about two strikes ago.

There's also the theory that the optimal baseball order is to order your batters starting with the highest OBP+SLG (I forgot what it's called for the moment) and go down from there because that will get your best hitters up the most number of times during the course of a season. That being said, I still stand by my other comment that the "optimal" baseball lineup will net you 5-10 runs more a season (162 games) which is not as much of a factor as who you put in the lineup.

[Mountain Sage]

Badams52,

Sorry, I don't play Baseball, just CivIII

I think that city placement is extremely important at the very beginning of the game. Take the extreme case: your Capitol is build on a floodplain. How many more Settlers, Workers and Warriors could you crank out in, let's say, 100 turns that if it was build on a desert tile?

In fact, the city placement of the first 6-8 cities will affect the future of your game tremendously. Faster REX, more military units, improvements build more quickly etc.

Aftrerwards, you can make more 'mistakes', or let's say you can be more 'picky'. For instance, how many times did you build a city on a (almost) useless spot (like a plains tile sourronded only by mountains), just to prevent another Civ to build there, instead of on a juicy spot on a river? That city will be useless for a long time.

The size of the map is also important: on a Tiny map each city, and therefore each city placement, is a question of life and death (or so I have been yold). On a Huge map, you can 'waste' a couple of turn going for the 'best' spot.

The size of your empire is also important. On a Huge map, you'll have easily 30+ cities. Again, a 'perfect' city placement for the core cities is essential, but for the 31st?

Then, at last, the level you are playing becomes important, you can make more mistakes at Chieftain than at Deity, or Overgod.

In the long run, after 4'000 years, a 1 shield miss from your Capitol is not so important (it will still produce 60 or 70 anyway by now), specially in respect of your 50+ empire, but remember that you carried this miss for 200 turns, therefore you lost half a Wonder (more or less).

Finally, I totally agree that NOT placing the right terrain improvements AND not working them is the final sin which will make your empire go under sooner than later.

Well, have a nice week-en all of you and happy Civving

[pedrojedi]

Cumi's statement is very good. Actually, I think it is, at least for me, the chief argument for OCP. Timing in early game is very important.

if between your cities are 5 tiles (insted of for example 3), we can easily calculate how much more turns your wrokers need to connect two cities. If you continue like this, you will have to build additional workers. If so, you will be able to start building a wonder a few turns later, in the city, that has a 1 pop smaller size. You have to wait with building a settler also couple of turns. Later, your wonder is a few turns late comparing to your opponents. You have to forget more and more wonders. The avalanche is started and you feel you are loosing a game....

I usually do not follow many guiding for building cities and use every tile possible. I like to build cities on good spots with plenty of shields and lots of food, specially if there's a river about. This works for me in the first 7-8 cities. After that, the next cities will almost always be on less desirable space, because I'm not building them to necessarily be good cities, but because I want to secure a bottleneck, a good space to start a campaign against another civ or just to be a beach head. Even though I end up developing these cities, I recognize that some of them should have only the necessary improvements to survive and be protected (temple, barracks, walls, and a marketplace - people like buying and give money to me! )

I know distance is all at early game, but I just like to have all tiles for all cities later on, that's why I use 4-tile distance between cities. That's just a bad habit, hehehe.

Sometimes I don't respect these distances - a good example are tundra and desert cities, which do not have a great advantage in being with lots of space, hence the practical 3-tile distance. They are for practical means just Myciv huts.

[badams52]

Originally posted by Mountain Sage
Badams52,

I think that city placement is extremely important at the very beginning of the game. Take the extreme case: your Capitol is build on a floodplain. How many more Settlers, Workers and Warriors could you crank out in, let's say, 100 turns that if it was build on a desert tile?

I certainly agree with you that city placement in the early going is very important, just whether you use a 3-tile or OCP style is less important to winning than how much land you get, which is more of a product of wars and cultural boundaries.

The size of your empire is also important. On a Huge map, you'll have easily 30+ cities. Again, a 'perfect' city placement for the core cities is essential, but for the 31st?

This I would disagree with. 'Perfect' city placement is not important as long as you are working all your territory. The size of your core depends upon distance and amount of cities. Alexman who's studied corruption would be the best to tell us whether a 3-tile pattern near your capital outproduces an OCP pattern, but I suggest that the net result for your empire in terms of production and gold are the same regardless.

In the long run, after 4'000 years, a 1 shield miss from your Capitol is not so important (it will still produce 60 or 70 anyway by now), specially in respect of your 50+ empire, but remember that you carried this miss for 200 turns, therefore you lost half a Wonder (more or less).

Though a 1 shield miss might be important, it isn't as important as the amount of shields you lose when a improvement or unit is made.

Take building a swordsman which costs 30 production point. A city with 8 production build in 4 turns with 2 wasted shields (unless you micromanage tremendously)....here's a chart:

  Production   Turns to build Wasted Shields
 ------------ ---------------- -------------
    15                2               0
    14                3              12
    13                3               9
    12                3               6
    11                3               3
    10                3               0
     9                4               6
     8                4               2
     7                5               5
     6                5               0
     5                6               0
     4                7               2
     3               10               0
     2               15               0
     1               30               0

So the difference in the lower production cities is more obvious, but move up from 6 to 7 and you end up with the same time but wasting 5 sheilds. And if I inclued 10-14, all those would have the same turns (3) while each progressively wastes more shields.

I think the wasted shields you get from a city is more due to the remainder of the cost divided by the production than the lack of having a few shilds missing.

Edit: formatting

[Theseus]

badams52, thus my interest in disposable towns... and, invevitably, ralphing.

Maximal use of good tiles followed by larger cities and minimization of corruption.

The exact pattern, I've come to think, is not important.

[Dominae]

Originally posted by Theseus
Maximal use of good tiles followed by larger cities and minimization of corruption.

By the time you get to Sanitation, you've either conquered quite a bit of land, or are planning to in the very near future. Therefore the reduced Corruption due to number of cities is, I guess, not really that big of a difference if you've got 4-spacing or 3-spacing. You'll still be way over the OCN because of all the crap towns you'll be conquering. This is just a hunch, so I would like to see some tests done on this. But I'm not going to do any anytime soon!

So, assuming what I've said above is correct, all you're getting with the "camp and abandon" strategy (ralphing, let's say) is size 13+ cities in the late-game. We should figure out if these are really worth it.


Dominae

[Master Zen]

Originally posted by Theseus
badams52, thus my interest in disposable towns... and, invevitably, ralphing.

Maximal use of good tiles followed by larger cities and minimization of corruption.

The exact pattern, I've come to think, is not important.

Ralphing is actually quite flexible even with hostile terrain around, the basic premise can be rather easily reproduced. You can basically make "ralphing patches" of ralphed-out cities separated by hostile terrain and even fit in a few cities in between.

How many times did I say ralph?

[badams52]

Once I finish my AU206, I'll beging to analyze ralphing more. I'm still in the industrial age and am disbanding the last 2 camps netting 6 workers each.

[Ephesos]

Okay, I've looked at your ideas on placement, and they all assume one thing: that the player is a bloodthirsty conqueror. What about peaceful games (just humor me here)?

My idea is that in a peaceful game, cities should be spread out so they don't overlap too much (say, never more than 3 squares on a side), and so they radiate out from your Palace-Forbidden Palace axis. It seems to work for me. As long as you devote time to building courthouses, you should be fine. Besides, I think the "ralphing" concept makes things too easy (I really don't use it very often).

[Nor Me]

Ephesos, you are right that close city placement will eventually hurt you as a builder (generally because waste slows improvement building meaning less tiles are affected by, say, universities.) OCP is fine if you can do well in the early game with it. In one game I replayed (AU204) a few times with different city spacing, Optimal City Placement worked best. I had similar luck (1st war gave a leader 3 times). But generally you fall behind at the beginning.

Originally posted by Ephesos
Besides, I think the "ralphing" concept makes things too easy (I really don't use it very often).

That is why an OCP+camp type placement would normally be better. Camps can buils settlers rather than military units and there is no real reason not to build libraries, marketplaces etc. early if you don't need the military. If I have a lot of cities, camps start being disbanded early. It does sometimes make sense to have 12 tile holes in your empire in the middle ages but only near the Palace and FP or between them if they're close. Of course if you want a challenge try 10 spacing.

[badams52]

Originally posted by Ephesos
My idea is that in a peaceful game, cities should be spread out so they don't overlap too much (say, never more than 3 squares on a side), and so they radiate out from your Palace-Forbidden Palace axis. It seems to work for me. As long as you devote time to building courthouses, you should be fine. Besides, I think the "ralphing" concept makes things too easy (I really don't use it very often).

Peaceful or warlike the reason for close spacing and the greatness of it is the same: you work more tiles given the same amount of land during the ancient and middle ages when games are won and lost. Ralphing just allows your cities to get much bigger by disbanding cities that don't fit the OCP format when your tech allows you to work all the tiles within a city's radius. So my theory only really kicks in until you've reached the industrial era and have sanitation at had to make your OCP cities grow.

And as for Ralphing, or creating rings of camps around your OCP cities, I think it is quite a powerful strategy, but by the time you get to the industrial age, you've probably already won the game if your other strategy is sound, so my feeling is that whichever city placement scheme you use, by the time the industrial era rolls around, it's moot and you should have the game well in hand.

Persoanlly, I would be more inclined to use 4-tile just out of preference. Ralphing takes too much micromanagement of cities for my tastes. I like to let the city manager decide which tiles to use.

[Master Zen]

Originally posted by Ephesos
Okay, I've looked at your ideas on placement, and they all assume one thing: that the player is a bloodthirsty conqueror. What about peaceful games (just humor me here)?

Peace? what the hell is that?

C'mon people, stop smoking and let out your stress warmongering!!!

My idea is that in a peaceful game, cities should be spread out so they don't overlap too much (say, never more than 3 squares on a side), and so they radiate out from your Palace-Forbidden Palace axis. It seems to work for me. As long as you devote time to building courthouses, you should be fine. Besides, I think the "ralphing" concept makes things too easy (I really don't use it very often).

Agreed. Just let's see how long your neighbors like your peaceful attitude...

[Master Zen]

Originally posted by badams52

Persoanlly, I would be more inclined to use 4-tile just out of preference. Ralphing takes too much micromanagement of cities for my tastes. I like to let the city manager decide which tiles to use.

Camp cities, whether using Ralphing or any other technique does not overburden you. You only build barracks and military units, usually of the same tipe for a long time.

[badams52]

Originally posted by Master Zen

Camp cities, whether using Ralphing or any other technique does not overburden you. You only build barracks and military units, usually of the same tipe for a long time.

I'm talking about management of cities in terms of deciding which tiles get worked. I usually let my city manager decide so that I don't have to worry about rioting and unrest as my cities grow. Building in those cities isn't my gripe. For me, the less micro management, the better.

[Bamspeedy]

Originally posted by Ephesos
Okay, I've looked at your ideas on placement, and they all assume one thing: that the player is a bloodthirsty conqueror. What about peaceful games (just humor me here)?

Try a culture victory.
If you have 100 cities and you pop-rush a temple, or library (or both) in every one of those cities.....

[Mountain Sage]

Originally posted by Master Zen


Peace? what the hell is that?

C'mon people, stop smoking and let out your stress warmongering!!!

Agreed. Just let's see how long your neighbors like your peaceful attitude...

Ok folks,

This time you really made me

I'll post a nice little game save this Saturday night (GMT time) about city placement and 'The Virtues of Being a Peacenik' and I hope it will teach you a lesson or two

[bongo]

Jugding by your 'emperor arcipalego' gamesave you'll easily out-peace the rest of us

[Master Zen]

Originally posted by Mountain Sage

Ok folks,

This time you really made me

I'll post a nice little game save this Saturday night (GMT time) about city placement and 'The Virtues of Being a Peacenik' and I hope it will teach you a lesson or two

Ok, I'll post "Starting a world war all by yourself and living to tell about it".

[Mountain Sage]

Here is my screenshot about my military from 950 BC to 990 AD.

Why did I keep these units?
Spearmen 1 and 2 : no. 1 for 14 July parades (odd years), no. 2 for even years.
Galley: it's Joan personal fishing vessel (sharks, marlin, kraken...). Quote from Joan ' As long as I float in my milk baths my caravel shall float on our waters'

Why no modern navy? Quote 'Steel is better around some body parts that on a floating coffin'.
Why no planes? Quote ' What goes up, must come down'. Now you know...