Size DOES Matter!

[Velociryx]

Probably one of the coolest things about writing strategy notes is coming up with eye-catching titles for them here on the forums....

Anyway, I'm in the thick of a mini-tourney game that illustrates a couple of key points about 4x games in general, and I thought I'd jot them down so I can keep them from rattling around in my head so much!

The game sets up an interesting dilemma, right off the bat.

Ideal city spacing sees your cities five tiles apart, eventually leading to 21 tiles of productivity, with eventually being the key word.

Hand in hand with this is the corruption monster. Two flavors of corruption.....that caused by distance from your capitol, and that caused by total NUMBER of cities.

Imagine a spectrum of players, along an axis looking something like this:

Purists<-----0----->Borgs

To the Purist mindset, the notion of planting cities as close together as blades of grass sends shivers up one side of the spine and down the other. They want to make MASTERPIECES of cities.....ultimately, getting the full 21 tiles of production out of each one they build. They want their cities to be proud showcases of art, industry, and culture.

Contrast that to the guy residing at the other end of the spectrum who looks at his cities with a colder, less discerning, but perhaps more calculating eye. All he wants is a production center, on the thinking that more production centers = more culture, more places to build troops....more stuff in general. Hell, build the cities every other tile for all it matters....They won't grow very big, it's true, but OMG will you ever have the capacity for culture and troop production!

Most of us, fall somewhere in the middle of these two extremes, I suspect, and in fact, it would not be surprising in the least to discover that most players tend to mix and match, at least to a degree.

The main advantage that the Purist sees is one of Economies of Scale. There are advantages to being big. A big city with lossa production can crank out tanks really, really fast, for example.

The main disadvantage though, is that you have fewer places to crank all those yummy tanks out FROM, and therein lies the constant struggle. More production centers, or BIGGER production centers, cos that's what it comes down to, in the end. Cities are nice, but they are, at the end of the day, merely places for you to build troopages to smash your enemies with.

Temples and other happiness enhancers exist solely to enable you to put more workers to work building troops.

Scientific enhancers exist solely to give your workers BETTER (and increasingly expensive) troops to build.

Culture producing things exist to bump your borders, and insulate you from the dreaded flip disease when fighting.

More vs. Bigger

A good, constant debate, and when you get right down to it, they're both really debating the merits of size.

The Purist sees his size gains in having relatively fewer cities to manage, but having each of those cities being wampum big production centers.

The Borg doesn't really care how productive each of his cities are, cos while the Purist is maybe working on twenty tanks, the Borg is working on a hundred or so, and yeah, they might take a while to finish individually, but with so many in production, he'll likely be getting at least some every single turn.

In the end, does it matter?

You bet it matters.

I would contend that neither approach, if played to its extreme, gets the absolute most mileage out of the game.

If you want to excel, you need both. You need "killing fields" of lossa little towns that exist for four reasons: Temple, Barracks, Library and Troops. The two cultural builds will ensure that they're not easy flip targets, the barracks and troops are for obvious reasons.

And you also need proud centers of commerce, science, and industry.

Have either one or the other exclusively, and you'll do well, sure....for a while.

Mix it up, listening to your instincts and according to the signals the game sends your way via the lay of the land you find yourself in, and you almost can't HELP but do better, and hey....once you've got the world in your pocket, if you feel like tidying up a bit and nixing some of your too-close towns....by all means, thin the herd.

Cities are, essentially factories.

They have employees (population points), jobs to do (building stuff), and even vending machines (food production?). Each city = a lil' factory out doing its thing. Now, sometimes you want a big, mammoth factory to handle really big jobs, but, as our own modern economy continues to illustrate for us, the notion of cellular manufacturing is strong, and sometimes, for some jobs, a little factory is better and more effective than a big one. Nowhere is this more true than in the ancient age.

Why?

Simply put, troops are cheap! Small cities without much production can build ancient age troops pretty quickly, and as we all know, getting a horde of swordsmen, archers, and horsemen together in the ancient age and opening up a big ol' can of whoop a$$ on all your nearby opponents is a very good way to get yourself off to a strong start.

The more cities you have to work with then, the faster you can accomplish this goal.

Yikes! And it's after six and time for me to head home, so more on this later, but the crux of the post is this:

Either way you look at it, and no matter how you slice it....size *does* matter!



-=Vel=-

[dexters]

Well, on the discussion of placing cities close together vs. maximizing city spacing, it is a matter of constraint.

Applying the theory of constraint -- that is one must manage the constraint(s) hindering our ability to reach our goals. For example, a constraint facing an airline is that the don't have limitless access to planes. To deal with high passenger seasons, they must manage their constraint -- their limited fleet of planes -- by reducing turnover time, speed up passenger movement and bottlenecks to squeeze more flights per day.

In Civ III, constraints come down to real-estate, threats and production.

Small island civs should utilize all available tiles to maximize production. If three are 3 land tiles outside city range and sea tiles beyond that, drop a city and utilize those tiles.

Large continental civs have room. Cities should ideally not be clumped together as space is no longer a constraint. The constraint for continental civs is defenses. As opposed to an island civ with the sea as a natural barrier, continental civs will be faced with fierce competition from other civs around as well as troops moving into your territory to attack. Hence the whole REX strategy was developed to deal with this. Therefore, maximixing production, and maximizing the amount of resources and luxuries a civ can obtain on a continent is to expand far and wide. land tiles outside city range should be seen as an acceptable trade off that can be backfilled later on once borders are established and perhaps more cities are needed to crank out units.

Now, there are exceptions. Some islands may have 2 or 3 civs together. In this case, players may want to maximizie its civ's borders with REX strategy, then deal with the constrains facing it -- small landmass and crowded island. For more of us, there is one solution. Wipe out the opposition, and we've seen many players play the "my continent" strategy.

On the other hand, one may find themselves in a large expansive continent and not meet anyone. In that case, REX may no longer be as urgent, and the constraint of having aggressive rivals threaten one's civ is lessened. Thus, the constraint in these cases will be something more practical -- such as working with limited shield productions of cities to maximize output.

[Aeson]

I don't see any reason why a player can't have the best of both worlds. Until the Industrial era, just pack the cities together, mark some as permanent, others as just temp. No need to waste tiles waiting for Hospitals (or Aquaducts for that matter). By the end of the Middle Ages, the temps start packing up to move. As you can transfer the population points in the form of Workers and Settlers to the new big permanent cities, there is nothing lost. The flood of Workers also comes just about at the right time for laying down rails.

[dexters]

You can't have the best of both worlds when you're running a small island civ and have 5 cities instead of the 10 you could have squeeze out. Because the player insist on having cities having full 21 tile productions.

At the same time, you can't be having the best of both worlds in a large pangea continent when you have a tiny little empire of 20 cities clumped together in a corner.

You not only deny yourself increased chances of having strategic resources and luxuries inside your small borders, you also deny your cities from reaching their full potential. There's plenty of land, anyone should build 20 cities spread as wide apart as possible as opposed to 20 cities clumped together, as these cities will grow into industrial powerhouses in the late game. In contrast, clumping the 20 cities together will essentially cut off your growth in the late game, especially if AI civs move in and box in the the player in a corner.

There are some exceptions, including building 2 cities close together for strategic reasons. I've built a city next to an oil resource because I knew the mandatory 1 tile border will make the oil resource fall into my territory. I didn't mind the fact that i had a pop 15 city 2 tiles away. It was a strategic decision.

But I reject the notion that REXing or having your city all clumped together is the definitive strategy. Who is more predictable, the player or the AI? Dictates to the effect give little thought to adaptation to one's environment and maximization of production under different circumstances.

[FrustratedPoet]

Originally posted by Velociryx

Imagine a spectrum of players, along an axis looking something like this:

Purists<-----0----->Borgs

-=Vel=-

ohhh .... the center tags work then ... cool

anyway ...

I'm a Borg, at least, I'm more Borg than anything else.

Is it because those shining red-eye thingies look pretty cool?

Is it because big cuboid spaceships are just f**king weird, and therefore interesting?

Is it because I am so damn rubbish at Civ3 that I am unable to visualise the full 20-something city radius?

Is it because I am so inept that I don't even know (despite just reading the first post) what the maximum city radius even is, regardless of whether or not I chose to utilise it efficiently?

(Is it because I'm more than slightly drunk?)

The answer to all of these is: YES! ... except the last question ... the answer to that is S!YE ...

The main (and real) reason why I am a Borg is because I like to play Industrial Civs .... faster workers = more tile improvements = more settlers = more workers = more tile improvements = more city improvements = more city production = more culture and more troops = more cities = etc. etc. etc. = etc. etc. etc. = MORE WINNING!!! YEAH!!!!!!

[monkeyman]

I WISH you could get away with the minimum number of cities and stand a chance of winning...

All I know about size is that if I am to win I am required to build more cities than I want to on a standard size map !micro sucks!

Does the fact that shields can't be carried over from turn to turn impact the large city/small cities dynamic? I think yes...

--mm

[Txurce]

I think size matters for most victory conditions, but absolutely not for domination. I just played a second game employing a very tight playing style:

•four cities and rush
•build barracks and temples only, maybe an aqueduct
•research only the first tech, chivalry, and cavalry
•use a GL for the FP, the rest for armies
•oscillate between opponents so no one gets too far ahead
•never stop warring - think ahead so you never stop to rebuild

My cities varied in size from 2 to 12, and were spaced more loosely than is ideal for this approach. The majority had no aqueducts.

Using this approach on Emperor with standard everything for the first time, I won a domination victory with the Chinese in 1325. This was almost one hundred years better than my prior best. (This game is detailed in the "Archer Rush" thread.)

In my second game with this approach, I was lucky enough to start as Japan on a pangaea-like continent with six other civs (the Americans were off-shore). As a result of being able to fight everyone in a rotating manner, I played worse than with the Chinese - about a century behind in research - yet won a domination victory (129k sq. mi.) in 1020. If I had played as well as in my prior game, I would have easily won before 1000 AD.

Because you can win via domination while still in the middle ages (even if the AI is in the industrial era), city size isn't as important as having the gold to upgrade all those horsemen you built early, as long as your ever-increasing number of cities never build anything but units (or wealth).

[Velociryx]

GREAT comments here!

IMO, the key to the whole size equation is in listening to, and acting on the signals that your current game is sending you.

As has been pointed out by others, sometimes you find yourself in a situation where, in order to compete (cramped with 1-2 other civs on a smallish island), you'd better cram those cities close--especially if you're playing at higher levels of difficulty where the AI gets heinous production bonuses....close spacing = the human player's means of countering AI production bonuses. So what if the AI gets 50% production bonuses? If he's got 6 cities and I have fifteen....I'm not even gonna bat an eye at his "bonus" cos I'll eat him for breakfast in short order.

Conversely, if you have plenty of room to spread out, then you are shooting yourself in the foot by confining yourself to a too-small area, and letting the AI gobble up all the juicy terrain you could have had.

In most of my games, I do a bit of both. I *never* opt for the pure Purist approach of five apart, compromising at four most of the time, and three when I'm interested in ramping up my productive capacity to obscene levels.

The four-apart spacing works best for me, because of a city's starting borders. I can create a continuous wall of borders with no cultural improvements by placing cities four apart, and they grow pretty big in the late game (tho not as big as they could....not as big as a Purist spacing approach).

The three-apart scheme is ideal along the border with a rival civ....why? A multitude of reasons, actually! But the three most important are:

1) Interlocking defense - slow moving troops can always end their turns inside the safety of a city (pending rivers and smart road construction). Fast troops can skate in a 2-3 city radius....that makes executing defense a very easy matter.

2) Having relatively more cities on the enemy's border = more places to build troops from and launch an attack.

3) Cultural Weaponry - making enemy cities flip to you is easy when you have 3 of your cities pressing their borders against the enemy one. Poof....you get bigger without firing a shot.

City spacing schemes are not really a strategy in and of themselves, it's true, but they *do* make a great building block....the key, again though, is flexibility. Listen to what the game you're currently involved in is telling you and go with the flow.

As with any game, working WITH what the game itself gives you is much easier, and much more effective than working against the grain of the game itself.

-=Vel=-

[dexters]

Good stuff vel. What I'm saying is that it is wrong to even construct a view of "purity" That is, that such a scale of purists exists. Obviously, anyone who is advocating one extreme or the other is simply misguided on the basis that no single strategy can account for the many variations and factors in each game.

Managing the growth of an empire is all about knowing the constraints that keeps you from getting what you want. If territory is spare, maximize ALL tiles. If there's lots of room, spread yer wings and fly.

Now, I know we apply this subconciously, but I'm simply spelling it out for all to see. And I believe your compromise solution is essentially this idea in practice.

I just get a little annoyed when people give you a magic forumla. There is none. There are strategies for each situation and people have to string these disparate strats to make a their game strat.

[dawidge]

Another advantage of the 3-apart spacing is that from the early game, once roads are in place, you can easily shift troops back and forth as extra defenses are needed. Want to get that brand new swordsman you built to the border cities for defense? Move it one city over, send the sword from that city over to the next, lather, rinse, repeat. Back in Civ2 and again in SMAC/X, it was fairly easy to have a standard city defender loadout of two high powered defenders and a high-powered offensive troop. In Civ3, it isn't until I've already guaranteed myself the win that I have that comfort.

With a city spacing of four, you get the 9 core tiles plus 12/2 tiles shared with cities in each other direction, for an average of 15 useable tiles per city. You need to be *very* late in the game to have more than 15 people available for work (as opposed to entertaining) in each of a cluster of cities. Heck, you have to have built a hospital in each of that city cluster for it to even be an issue.

Due to the odd "plus" shape of the useable city area and offset positioning, three-spacing in one dimension often becomes four-spacing in another, yielding an average usable number of tiles per city closer to 11 or 12. Which, for most of the game *is* optimal. When you hit the late-game, simply select every other city as a metropolis, go to the intervening "suburbs" and engineer its food production/consumption and specialist usage to limit its tile usage to the core 9. *Bing* you now have several core metropoli for cranking out the biggest and baddest units, and several suburbs for cranking out utility stuff (workers, settlers, defenders, etc.) or as a slow producer of the big, bad units. With three-space, you get more suburbs than metropoli.


"Optimal" spacing depends entirely on the game parameters, the civ you are playing, and the terrain. On arch maps, build dense. On continental and pangaea maps, build loose and take territory from your neighbors. Most important of all, is city placement for exploiting resources. In most cases, a string of 3-apart cities along the only river in the area will give you more net production through the course of the game than 5-apart cities that have to build aqueducts. I'm fairly sure that the need for denser city placement goes up with the difficulty level.

That said, I haven't used anything other than the puritan 5 tile spacing in the last few games, but that was mostly due to terrain and/or early victories that granted me lebensraum. I also just upped the level I play at (was Regent, now Monarch).

[nbarclay]

In practice, I've never been fond of spacing cities to use all 21 tiles in their radius because the shape of cities is such that doing so always, inevitably wastes land. As long as I don't get cramped in too badly, my preference is to space cities so I make about as much use of the productive land available to me as is reasonably practical but my cities can still grow to be an average of somewhere around size 17. (That's about as big as can be kept happy without gold or entertainers even with Sistine and Bach's anyhow, so I don't mind having the cities not grow all the way out to max, especially if they're very far from wherever the capital or FP is at the time.) By the way, that's not something I caculate with any precision; I just eyeball it and hope things work out about right.

I won't claim that that's ideal from a strategic perspective. It does very well indeed for me in the late game (if I don't win by domination first or play on Deity and give up), because more laborers per city means more laborers whose yield is relatively unaffected by corruption based on number of cities. But I'm sure I could do better earlier with a denser build.

Of course if I find myself cramped for space, I generally am willing to go with a denser build pattern. One of my favorite tricks in some situations is to have my cities initially make use of every available land tile, but plan on shifting coastal cities to using mostly water tiles later on so inland cities can expand more. That gives me the production benefits of a dense build early and the wealth benefits of a somewhat less dense build later. (Just make sure the cities using saltwater tiles are on a coast where they can build harbors if you don't want to have a hard time feeding everyone.)

Nathan

[nbarclay]

By the way, if I know in advance that I'll be going for an early domination victory, I'm far more likely to go with a build pattern based on the assumption that my cities will never be over size 12. After all, what's the use in optimizing for a phase of the game that I never intend to reach?

Nathan

[Harovan]

I'm a proponent of core cities 3-apart, and thus about 2 dashes right of the 0 on Vel's scale. Why?

First, there is corruption in the beginning of the game. On a standard map, a city 4-5 tiles apart has usually already 1 wasted shield, while a 3-apart has none. In the early game, this is a HUGE advantage to get a good start. Purists should try it, the effect is immense! Right, the corruption for # of cities will be less with 4-5 spaced cities, but in the beginning of the game this doesn't matter yet. And a won ancient age is, in most cases, a won game.

Second, the mentioned troup shifting on attacks or barbarian uprisings. Helps a lot!

Third, my cities usually don't grow far above size 12, so I don't need so much hospitals. I usually build 5 or only slightly more, for the Battlefield medicine.

Fourth, smaller cities are happier and easier to keep in WLTKD. If I have 6 or more luxuries in the late game, I often can even sell my cathedrals. Each sold cathedral is worth 2 more supported tanks without hurting research.

Fifth, smaller cities produce less pollution. The pollution for overpopulation is higher than for production.

Sixth, to produce units in the late game, size doesn't really matter. Not only for the reason Vel mentioned. Imagine, we are producing tanks. Even with a 21-tile-powerhouse you usually can't build a tank in one turn. Exceptions are rare. I can in a good size 12 city with some mountains or hills and a factory and powerplant (or Hoover) make a tank in the same 2 turns. Do I need a powerhouse? No. And I have even more 1-tank-per-2-turns cities than the purist has.

The only thing in favor to be a purist is wonder building. Here I help myself with a trick. Since the cities around my capital are dense built in the early game (given I didn't move it), I try to place my FP in a former AI capital, for two reasons: They are usually great city sites, and the AI doesn't build dense, so this city and the cities around it can be powerhouses (plus, they have low corruption).

That is my approach of the "2 centers". One (around my palace) is the center for growth and unit production, the second (around my FP) is the powerhouse for wonder building.

[Jethro83]

What a great topic, and thread.


Anyway, what I think is the downside to the Purist approach is definitely the corruption and waste. I used to be completely purist in my city-placement. It led me to always raze enemy cities, or refuse them when they flip if they were within the 21-square radius of one of mine. Just didn't work entirely.

As for the borg approach, the main downside I see is the lack of metropolises. This means giving up a hefty defence bonus.

Also, there's less free units supported under monarchy and communism. Remember that a metropolis under monarchy and communism supports 8 units, while a city only supports four. Pack your settlements too tightly, and they won't even make it to become cities, and hence only support 2 units each. For someone like me, who always has a minimum of 3 defensive units fortified in each settlement, that presents problems when I am warmongering as a religious civilization.

Another problem is early expansion. Since the AI pukes out settlers constantly, they swallow up much territory quickly. Expanding under the borg approach yields much less territory. This allows the AI to grab much of the good lands you're planning to build on yourself. The exception I've found is when you're expansionist, and build a granary before anything else in each new city, and start puking settlers out of each other city (once tried the borg approach as the Americans, and swallowed up more territory than I usually do in the early expansion phase).


But as Vel said, striking a balance is the key. I've found it effective to go purist with cities in the immediate vicinity of the capital, and then outside of that, go borg and pop cities down left, right and centre. I've done that, and gained plenty of territory, while having enough production in a few key settlements to get the Colossus and Great Library (and sometimes the Pyramids), and a numerous amount of smaller cities to crank out temples/libraries (depends on presence/absense of Scientific trait), as well as a respectable ancient era army. Using these towns, I have also been able to fill the garrisons of the purist cities while they focus on trying to beat the AI in the ever-going wonder race.

[nbarclay]

At least on a standard or smaller map, non-expansionist civs can max their research, aim for Pottery, prebuild something else, and generally get a granary just as quickly as an expansionist civ can. I'm not sure I've tried the trick on larger maps, but I'm pretty sure the tech cost is too high on huge.

Nathan

[Harovan]

Originally posted by LordAzreal
Also, there's less free units supported under monarchy and communism. Remember that a metropolis under monarchy and communism supports 8 units, while a city only supports four. Pack your settlements too tightly, and they won't even make it to become cities, and hence only support 2 units each.

If I pack my settlements tightest (2-apart in every direction), each settlement has 4 tiles (the city tile and 3 extra) and will probably remain a town. A metropolis can control 20-21 tiles and supports 8 units. On the same 20-21 tiles I can place 5 towns that support 2 units each, making 10.

[Zachriel]

Purists<-----0----->Borgs

Somwhere in the middle of the spectrum, in a slightly different plane perhaps, is the "natural" expansion, which accounts for the local geography; some towns close, some farther away, hugging rivers, taking the best lands, contesting other lands.

[nbarclay]

Originally posted by Sir Ralph


If I pack my settlements tightest (2-apart in every direction), each settlement has 4 tiles (the city tile and 3 extra) and will probably remain a town. A metropolis can control 20-21 tiles and supports 8 units. On the same 20-21 tiles I can place 5 towns that support 2 units each, making 10.

Why 20-21? All you need for a metropolis is 13. If someone averages 15 tiles per metropolis or less, that's more unit support per tile than even the densest pack.

Nathan

[Jethro83]

Originally posted by Sir Ralph


If I pack my settlements tightest (2-apart in every direction), each settlement has 4 tiles (the city tile and 3 extra) and will probably remain a town. A metropolis can control 20-21 tiles and supports 8 units. On the same 20-21 tiles I can place 5 towns that support 2 units each, making 10.

How many units do you garrison in each of these towns? As I said before, that approach is a problem, since I tend to place an absolute minimum of 3 units in every city's garrison (since that is also the limit of units usable as police under monarchy), regardless of size. If I have a few metropolises (even if they're only size 13), each one can support the three defensive units I place in each one, as well as five more offensive units. If I were to try packing five towns into that same space, sure I'd have a total support of 10, but with three units in each town, I'd need to support 15 units, and that's just defensive units.

I don't mind a little bit of overlap, so long as there are enough food-producing tiles usable by each settlement to reach size 13.

[dawidge]

Bear in mind that closely-packed cities do not require as many defenders per city (I usually average 2-per for most of the game, including wartime). When a portion of your empire is threatened, you can ripple defenders into place in one turn in any age (rivers should be carefully crossed until Engineering).

[Harovan]

Originally posted by LordAzreal
How many units do you garrison in each of these towns? As I said before, that approach is a problem, since I tend to place an absolute minimum of 3 units in every city's garrison (since that is also the limit of units usable as police under monarchy), regardless of size.

I don't build that tight. My post was an "IF" case. Generally, I defend my cities with 1 defender, border cities may have more. If I own my own continent, interior cities are often undefended. Well... at least in single player . I avoid Monarchy and Communism (and prolongued wars), so police is not an issue.

[King of Rasslin]

This thread looks like a clone of my "proper city placement" thread!!!

J/K. Actually, gj Vel on explaining things a little more clearly than I did. I always make cities as close as possible to each other to reduce corruption. However, I might get rid of some cities by the industrial age so I won't have to make so many factories. I use the borg strat to get culture and troops early on but I switch to purist later on, like around the late middle ages or early industrial.

This switch from early borg to late purist is good for me because:

#1- I make banks and universities late. Because I don't want to pay a ton in maintenance later on, I disband a few cities by the time I plan to stop my middle age wars. I make workers with my cities I don't need and use them to clean up the pollution that will come soon.

#2- Production needs change. I would rather make 20 tanks quickly than 100 slowly. As units get more expensive, the purists gain more power. Although, because of shield waste and making more shields than a unit costs, the borg will always have a productive edge over the purists. I think the flexibility and speed of purists would be more useful as units get more expensive.

#3- Food. Cities can't break size 6 early in the game. It is very late in the game by the time they can break size 12. You should go borg when you are restricted in size and gradually advance as aquaducts and hospitials become available to you. Although the game is yours by the time you are making the hospitals That's why borg is so important, especially early on. The borg have the advantage of faster growth, however. Having twice as many cities amounts to having twice as much growth. It takes too long to grow a super city, even with granaries.

#4- Defence. You go borg early on to protect your cities! Sprawling empires have a hard time directing troops. However, railroads come with the industrial age. At this point, you can get away with sprawling super cities because your armies can "teleport" to a hot spot instantly! Use railroads liberally when they are available.

#5- Improvement costs. Well, this is the same as unit costs. But you get the idea. Making 10 factories in 10 super cities is easier than making 25 factories in midsize cities. This is the same as units.

So, there you have my take on the matter. However, I did succeed in staying borg in one of my games. 25 factories are very hard to make, but I had saved a lot of gold and had leader rushed the Hoover Dam. I became very, very powerful by using the Hoover Dam to make 25 fast hydro plants.

One thing you must remember is to make your border cities first, then "fill in" accordingly. A common borg mistake is to make your cities close together at the start of the game. Then they wonder why they don't have enough resources! Make border cities first to slow AI expansion, then go borg. This insures that you will have the resources you need to wage war. Borg rule.

[Theseus]

First, to dexters:

"The Goal" applied to Civ3?

[Theseus]

Second, GREAT thread. Vel, you gotta knack for getting the discussion going.

I have a thesis on this whole topic:

There are (or were... I'll get back to that) 3 types of players:

* Perfect Builders: 21 tiles per city, goddamnit.

* Borg: New concept, specific to Civ3... Aeson and Sir Ralph come to mind (no insult guys, I know you're more sophisticated than that, but Aeson's monster civ, and Sir Ralph's first mini-tourney are indelible).

* Flex Builders.

I'm the last, I think. Actually, most of us are becoming that... the game demands it.

Early cities: pack'em in. Take advantage of resources, rivers, flood plains, hills, etc.

Maybe you keep all those core cities, maybe not. As time goes on, spread out a little, taking advantage of the increase in territory you get from your early cities.

Maybe you want to abut a neighboring AI civ, and steal a resource through location. Maybe you want a chokepoint, or a forward military base.

Flexibility: As Vel says, read the game. There are times to Borg, there are times to be perfect. There are times to be both.

[Theseus]

Oops, sorry, didn't finish the thought.

I think there is still a lot of behavior left over from Civ2 and other like games.

Flames notwithstanding, Civ3 is significantly different and more complex / demanding than anything that at least I have seen before.

Vel, for example, broke apart the basic strats for standard settings... but stray from there, and all of a sudden there are factors and nuances that REQUIRE changes in approach.

I like watching what Aeson does... where does his cold game fit in the city spacing spectrum? Does the concept even apply?

[vulture]

It seems to me that the best approach changes significantly under different circumstances. In the early game, you set down a few cities. At this stage you can forget about optimal use of your land, in my opinion. If you have 5 cities working a total of 15 tiles, it doesn't really matter how close together those tiles are, except for corruption issues. In terms of using your land efficiently, it doesn't matter so much whether you build at 3, 4, 5 or (God forbid) 6 spaces. As long as you have no strong constraints on available room, then is doesn't make any difference. The important issues are corruption (possibly significant) and using good tiles (more important in my opinion). I will quite cheerfully build my initial cities 6 or 7 tiles apart if there is a really tempting site somewhere with a few good tiles. If you only get to work 10 tiles early on in the game, you want to work the 10 best that you can (unless they're far enough away that corruption is really an issue). Sometimes that will mean packing the cities in, if (as in the German mini-tourney game) the amount of useful land is seriously limited.

When your expansion runs up against other civs (or the coastline), then is the time that you gain from building close together. Simply because you can't go around building cities wherever you want to without causing wars, so you have to pack cities together. The issue here is no longer 'I have 5 cities and am working 15 tiles', but 'my empire has 300 tiles, and I want to work as many of them as I can'. This is obviously done by packing your cities in so that as many useful squares as possible are being used.

The last stage is after you've conquered enough territory (and built your FP) such that new cities you build/conquer on the periphery are massively corrupt. Now you want to try and minimise the corruption on as many worked tiles as possible. The optimal solution is to have your city spacing such that your core cities are all size 20 and have no overlap (you can in principle use 21 out of every 23 tiles if you do this (or size 18 cities using every tile with 2 overlap is also possible on theory)). Consider a tile a certain distance away from your capital or FP. If you use a more dense than optimal build of cities, that tile is worked by a city with the same (more or less) distance corruption but worse no. of cities corruption so it has more corruption than the optimal layout. More widely spaced cities give you tiles worked with the same no. of cities corruption but worse distance corruption (and wasted space in your empire).

Rearranging the positions of all your cities is obviously not practical. The best solution as it appears to me is to try and build cities at close to the purist approach in the early game, going for good production tiles. When you can't expand freely any more, put down small cities through your core to make optimal use of the tiles their. And when you get past 1.5x the optimal number of cities (or 3x with a well placed FP) and can build hospitals then start disbanding the small cities to try and work more tiles with lower corruption.

Plans may change subject to game conditions...

[Arrian]

I'm trying. Really trying. But I always end up looking at the map and seeing 3-4 tiles (mostly 4) between my cities (as in 1-2 tiles overlap maximum). I'll have to force Aeson's concept of "temp" cities on myself. That's a good compromise, as you can later remove them.

On the other hand, I just about never raze enemy cities, and the AI is no purist. Therefore, my second core area (forbidden) is built with overlap. So I guess that pushes me a little bit toward the "borg" style.

I also tend to be less picky about city spacing when I'm building highly corrupt cities far away from my palace or fp. These cities may well become productive once I switch goverments, buy courthouses, and get WLTKD going, but for a good long while they just sit producing 1 shield/turn.

I suspect that I've been getting away with a non-efficient strategy because I've been playing on Monarch. I think Emperor is going to necessitate a more calculated look at things.

-Arrian the purist holdout.

[King of Rasslin]

Some AIs are more purist than others. I have noticed that the Babylonians have plenty of space between their cities and the Zulu have their cities packed very close to each other.

[King of Rasslin]

Respectful bump for this excellent thread.

[Cort Haus]

Nice bump, KoR - I missed this one before.

My two-shields-worth :

I tend to find that river & resource-access (if available) strongly influences my choice of city placement. Apply industrial worker-power to cities on rivers, and the gold & science quickly come pouring in. I've just researched my way to a end-of ancient era tech lead with 1000G in the kitty (monarch) with this as priority, whilst fighting a long-term defensive (at first) then offensive campaign.

For a builder play-style (< emp) there are two benefits to wide-spacing with culture to secure territory. The first is the long-term resource access (gotta get some desert/tundra/jungle) and the second is ROP's. Peaceful terrain leaders can scoop maybe 20-40G from each Civ every 20 turns whilst maintaining good relations.