Bamspeedy's OCP (Optimal City Placement)

[Nightfang]

While perusing the War Academy, I ran across a pictorial essay from Bamspeedy (here) about how to properly place your cities, at least your core cities, to maximize their productivity and reduce corruption. I examined the pictures, and felt that perhaps a clearer picture of how this works could be provided.

This is from a recent game where I applied Bamspeedy's city placement. I was on a standard continental map, though I was on all random, so actual type is uncertain. As you can see in the picture, space was at a premium, so I had to apply the OCP plan to make sure I took advantage of as much space as possible during my expansion. By 70AD, I had expanded to my limits, placing each city as best as possible to provide OCP. Here is a picture of how exactly it works.

Delhi, at the center of the picture, is of course the capital. I built the first ring around my capital. Placement is simple, provided you turn on gridlines (Ctrl-G) so you can get a clear picture of where your next city goes. Here is how to utilize this -

Delhi to Jaipur - three tiles north, one tile northeast.
Delhi to Calcutta - two tiles northeast, two tiles east.
Delhi to Madras - five tiles southeast.
Delhi to Bombay - three tiles south, one tile southwest
Delhi to Bangalore - two tiles southwest, two tiles west
Delhi to Karachi - five tiles northwest

This forms the first ring around your capital. These cities, once the land has been converted to optimal levels, will become your main production centers. Of these seven cities, all tiles are used, with only FOUR tile being shared with other cities.

From here, you can build out your next ring, and your next, until you are out of space.

Of course, terrain does not always permit you to place a city where you like. In my example, Punjab is one square north of where it should be because of a mountain being on the ideal spot. The whole area was surrounded by mountains, as a matter of fact, so I had to move one tile north to get placement. Same thing for Hyderabad, but in this case I was butting up against the shore, and had to pull it in one tile to get it on the coastline.

Since these cities are still close to the capital (Punjab, although improperly placed, is but nine squares from Delhi) corruption is low for a long time to come. As I expanded more to the north and west, my corruption values increased as i moved away from the capital, though a Forbidden Palace and OCP will help alleviate those issues later in the future.

Thank you Bamspeedy for this wonderful tactic, for it alone has improved my gameplay. Hopefully, my additions here will be useful to everyone.

 

[Park Ranger]

Nice demo you have a very neat and efficient city placement. One small criticism I have, however, is that if you'd moved Bengal one tile E or SE you'd have a port and still be on the river. Same for Lahore if that tile just to the SW is also on a river, I can't be sure. Especially since you are a commercial civ you shouldn't worry about being a little further from your capital.

 

[Nightfang]

Hyderabad is a port city, and off the map to the right side is where my next city was placed, and it was right on the water as well. This puts a port on each side of my controlled lands. Since I never concentrate heavily on navy, these two ports will do me just fine trade wise. On this map, in particular, Rome is on a little continent in the dark area on the left of the mini-map (which I was first to find!)

I did realize, though, after placement I could have made Bengal or Lahore a port city as well, but since the bulk of my rivals were on the same land mass as me, I did not feel it deathly important. The two ports I have are adequate.

 

[thd]

I agree, this has been very useful. I have often wonderered what the best placement of cities is, concerning wasted/used/shared tiles. Sometimes a more 'disordered' system will give you you cities that are more productive/effiecent alone, but in the long run bamspeedy's OCP works great!

 

[farting bob]

i agree, if you had moved the 2 cities to coastal places, you woud have a bigger navy chance, bt that would have ruined the setup, and could have made expansion more difficult later on.

 

[Park Ranger]

Originally posted by Nightfang
I did realize, though, after placement I could have made Bengal or Lahore a port city as well, but since the bulk of my rivals were on the same land mass as me, I did not feel it deathly important. The two ports I have are adequate.

Fair enough. Me I like two good ports on every coast but then in your case you would lose out on a couple of flood plains which are good late game tiles.

 

[anarres]

OCP is useful for high-scoring games, but of little use for earlier on.

A more dense build (like cities 3 squares from each other, so units can move between them in 1 turn) is much better for a stronger start, but as mentioned, it ruins the later-game score maximising.

 

[DaveMcW]

A few problems with perfectionist-style OCP:

1. Your cities are more corrupt for the first half of the game.
2. You miss out on short-term bonuses by moving your city 1 square over.
3. The Forbidden Palace is inevitably built late or in a bad spot.

 

[col]

Looks very pretty but by all means play that way against me if we meet at multiplay.

 

[heikeott]

Has anyone tried putting the cities really close together at the beginning of the game, and then abandoning the "in-between" cities later on when # of cities is affecting corruption and you need bigger cities to build stuff?

 

[satchel]

Originally posted by heikeott
Has anyone tried putting the cities really close together at the beginning of the game, and then abandoning the "in-between" cities later on when # of cities is affecting corruption and you need bigger cities to build stuff?

That's an interesting idea. I'm pretty certain I read about it here a long time ago, but I can't recall who posted it. It might have been in BillChin's dense build strategy article. So yes, I think someone has tried it.

It might work well if you are careful not to put early wonders in the "in between" cities. As you approach railroads and want extra workers you can shrink those cities relatively quickly while freeing up your other cities to grow and build other things besides workers.

I would like to try this. I like the aesthetic appeal of OCP but usually use a denser build, especially in the very early game, because I don't want to waste an extra turn of settler movement before getting those cities down. Heikeott's suggestion is a good compromise. I would definitely like to try it.

 

[Moonsinger]

Originally posted by heikeott
Has anyone tried putting the cities really close together at the beginning of the game, and then abandoning the "in-between" cities later on when # of cities is affecting corruption and you need bigger cities to build stuff?

Yes, I do that all the time. However, I do not disband my cities because of corruption. I usually disband them to increase the size of the surrounding cities.

 

[anarres]

Originally posted by heikeott
Has anyone tried putting the cities really close together at the beginning of the game, and then abandoning the "in-between" cities later on when # of cities is affecting corruption and you need bigger cities to build stuff?

Yes, but I finish the game long before it makes any difference to abandon them. If I used OCP tactics games would take a lot longer as well.

You really shouldn't need to abandon them anyway, unless you are milking. I think OCP is really useful for milking only, but I am happy to be proved wrong.

 

[Nightfang]

My main goal in using OCP is the maximization of my resources so as to produce super production cities. I am still playing on the India map I used in my example, and some of my cities are so embedded in jungle it is taking forever to get them to productive states. This game was experimental, so I do not care the outcome.

I think the best way to handle this will be to build maybe directly around your capital, one ring, using OCP, then get others wherever to get the empire rolling on good terrain.

I plan on experimenting some more, I will let you know what I find.

As for milking, I am not even sure I know how to do that, so you can leave me out of that group.

 

[JonathanValjean]

Thanks, Nightfang, for the further enhancement of Bamspeedy's excellent tutorial.

 

[TheNiceOne]

When I was a relative fresh civ3 player I tried to place my cities as explained above, and really itched to be able to build those hospitals that let me take advantage of it, but now I've learned better...

IMHO, this strategy is inferior to a closer city placement, and you will have serious troubles in MP against a player who has a denser build. By doing OCP you effectively leaves almost half of your territory until you get to build hospitals, and the game should be almost over before you get citizens to work on all squares.

The advantages with OCP is that you get powerhouse cities that can build wonders faster, a bit less number-of-cities-corruption than a denser city model and less upkeep due to fewer cities that need improvements.

The disadvantages outweighs this ba far though:
Early disadvantages:
You waste a few turns of movement for each settler to build the next city.
Its harder to defend cities when a warrior can't move between cities in one turn (it can move three roaded tiles).
You get more distance-corruption, and you need more time to connect cities (which also affects corruption).

Medium disadvantages
You waste almost half your territory since each city only uses 13 of its 21 squares. This results in less production and less income (and less score).
You get less culture, since fewer cities means fewer temples/libraries etc.

Late disadvantages
You need to research sanitation ASAP, but this tech can be bypassed with denser spacing, and you need to build (quite expensive) hospitals in almost every city.
You need more luxuries/happiness improvements when the cities grow above size 12, and WW will hit you faster.


So as you see, I now build my citiez closer. My private OCP is 3+2 spacing: I.e. from one city you'll find the four closest cities by going: 3up+2right, 3right+2down, 3down+2left and 3left+2up, (or mirrored of course). With this spacing, each city gets exactly 13 squares each, wich again is the number of tiles worked on by a size 12 city

The actual terrain will of course force me to not use this exact placement, but this wll matter less in this dense build than the OCP. If using the OCP and having to place a city one square off, that can actually mean 3 or so more squares that no city can work on. With my dense build though, placing one city 1 square off to the left will simply mean that more of its 13 used squares will be to the right side of the city.

 

[Globetrotter]

Originally posted by Moonsinger
Yes, I do that all the time. However, I do not disband my cities because of corruption. I usually disband them to increase the size of the surrounding cities.

Yep, same here

As for the OCP, i usually adapt to the terrain, and if i go to war early on and conquer really nearby towns, i usually don't change their positions...

 

[col]

I think the problem with OCP is the timescale that these cities take to become effective. Yes - eventually you will have some industrial powerhouses but it will take a long time before your cities need to work more than 9 tiles. You have wasted too much time and too many good tiles waiting for this.

A good rusher will take you out in the ancient ages or early middle ages before you get really get these cities going.

This approach just doesnt get you moving fast enough - too long to get settlers in position, empire too spread out to defend against a rush.

A denser build which can be thinned if necessary later in the game to allow cities to grow is much more flexible.

The OCP approach may work with a passive AI on a huge map but is unlikely to be successful against human opponents.

 

[Zachriel]

The two extremes, Optimal and Dense.

Certainly, a Dense Build is best if merely winning is your goal. Indeed, Civ3 as a puzzle has been solved. It is winnable in virtually every case with a Dense Build. The game is almost what some would call "broken," "broken" referring to the fact that the end of the game is already known at 4000BC, and the fact that advanced methods of war or peace are unnecessary for victory. The game has already been decided.

In between Optimal and Dense there is a special place, the Natural Build. The Natural Build places cities according to the natural contours of Rivers and Plains. The actual Density can be adjusted as well for game level and starting terrain. This makes for a competitive and enjoyable game.

(For an example refer to the Deity game, I, Hammurabi. The capital was more or less Optimal, but the build became denser along the Flood Plains. This led to a very competitive game and lots of fun.)

 

[Nightfang]

Further analysis -

Spent a great deal of time last night with my game, and I am fairly certain I am not happy with the way it is working out. This may be better, like Bamspeedy stated, on a Large or Huge map where you have more room for more cities. I have fewer cities than I usually do. and while they can crank out stuff pretty fast, it would be better if I could crank out more stuff, even if things took a turn or two more.

I was invaded by Jerkses (I should have seen it coming after the 25gpt deal I made with him) and lost one city (acquired through culture flip, so no biggie) before I could bring my military to bear against him. My cities, through making units quickly, could not get stuff to the front fast enough, and my exisiting units could not get in there fast enough, either. Luckily, I was ahead in techs, so I was able to field some superior units to keep him back, so no more damage was done.

Culture assimilation through this method is next to impossible. That city that I lost to the Persians has expanded to level 2 culture, and my cities beside it are barely brushing against him. I will be lucky if I can flip it back.

Steam Power made this more efficient, but that is a little late in the game to make me happy.

I think that this method will work well for producing a core of 6 or 7 cities that will be super efficient later on, while you continue to place your cities outside the first ring to get them placed well, either strategically or efficiently.

My game is in the 1200's or so, and I have yet to even build my FP. No need for it yet. Corruption is low in my cities, and since I am playing India, I don't feel right going to war. Not yet, at least. I was kinda happy Persia declared war on me, for it allowed me to field my war elephants and trigger my GA.

OCP, while useful, does not have a place in my future games, except maybe as a method of getting a roung idea where to place my next city. I will by no means be using it full tilt anylonger. I am too much a warmonger to have this in place.