Rexing, build settler at high pop or asap?

[civtilidy]

Hi!
I have a question to emperor+ players. During the early rexxing phase on emperor difficulty. Lets say u start with 2 lux so city can have 5 pop with MP (desp). What is most optimal for total output of commerce, shields and food:
1. Build settlers asap when each city reach 3 pop so they fall down to 1 pop? Then they will always be at pop 1-3 and produce less but settler will come faster so it might compensate.
2. Let city grow and build settler at 5 pop or if possible same turn as 6 pop is reached. The city will always be 4-6pop but settlers will come later. What is usually best?

I often let settler build at highest pop possible if there are ready improvements available to grow into. What do you think?

Also what is your build orders for new cities? I do:
warrior-worker- granary- settler/warrior

 

[justanick]

Build settlers asap when each city reach 3 pop so they fall down to 1 pop?

I very much favour this.

Maximize net food production. Minimize food going into waste. Minimize cities being bigger than necassary. Minimize the amount of turns settler do exist by building roads. Minimize the amount of workers. All together should maximize the total net food production.

An alternative is to build granaries. This increases the time till the first settler leaves a city and often it increases the need for workers. The smaller your cities are the less workers you need. Not using up food by early workers and or any other kind of population is in general a good thing. But than again granaries can increase your effective net production of food by more than a new city. If you gain only +2 food by a new city but can double +5 food with a granary the later will be better. Early on net food production is everthing and that as i understand it is the idea of rapid expansion. Later you need workers as growth will no longer be eliminated by building settlers. Also you need military for protection and military police. And certainly you will want to leave despotism in favour of a better government ASAP. For that you need a high net commerce. So sooner or later you will want to slow down your expansion via settlers. That is either because new cities are rather corrupt or because there is no more space that can be gained without military means. The later is the more likely the higher the difficulty setting is. If you play at Sid corruption will be very bad but AI settling down everywhere is worse. At emperor things are much more mildly and actually have a chance to settle faster than AI.

Also what is your build orders for new cities? I do:
warrior-worker- granary- settler/warrior

I have no general build order. Avoiding to build anything but settlers works best in terms of rapid expansion. But a few workers in total are indispensable, so a bit of food is used up that way. If production exceeds the need for settler that excess can be used for warriors. If the excess is smaller than 10 shields wealth is an option. Cities ought not to truely reach size 3, they must finish settlers during growth to size 3. Sometimes granaries do less harm than good, but at only +2 food settlers are better than granaries.

 

[Lanzelot]

Then they will always be at pop 1-3 and produce less but settler will come faster so it might compensate.
....
The city will always be 4-6pop but settlers will come later.

Why do you think, settlers can be produced faster at 1-3 than at 4-6?! The opposite is true: At size 1-3 the town will on average have something like 3spt, so a settler needs 10 turns to build. At size 4-6 you should have like 5-6spt, so a settler can be built in 5-6 turns. The delay of getting the first settler out, will be outweighed later on by more and faster settlers. Plus you have higher commerce when running 4-6 (or even 5-7, when on a river), so you get out of Despotism faster, which means additional food (even with no food bonuses around: grasslands can be irrigated, when you get out of Despotism!)

So I usually do the opposite of what justanick said: let capital grow to 5 (while building a granary) and then worry about the settlers. By that time I often already have roads in place, so the settlers can reach their destination in 1-2 turns instead of 3-4 when they get built "too early"... Think of it that way: you keep the population inside the capital, where they are highly productive, instead of walking them through the country-side for several turns, where they do nothing and only cost maintenance.

Of course there are exceptions. It all depends on the land you got. If there are food bonus tiles around for a couple extra towns, then I skip the granary and spit out 1-2 ultra-early settlers in order to get this extra food asap. But just in order to found another +2fpt town which will not contribute much to your empire for the next 20-30 turns, it is usually not worth to cripple your capital by shrinking it back to size 1! As a general rule, size 5-7 is the best to operate the capital.

Another argument: the AI is building settlers at size 3 (as you can see in F11, when they often have pathetic little capitals even quite late in the game...) So it must be bad!

If you want to see an example of a 5-7 build order, check out the story I just started:
The Republicans go to War

Here the early strategy was kind of forced, as this is an "Always War" game and I got contact with the first neighbor already at turn 10. So the capital was forced to produce nothing but military in the early turns, in order to fend off an early Dutch attack. But it did not throw my expansion that far back: during the time I was forced to build military, the city quickly grew to size 7, where it was able to make 10spt. With a well-timed forest chop, I was then able to build the granary in a mere 5 turns! (Usually you are happy, if you can get it done in under 20 turns... ) So now that I have the necessary military to defend myself, I can spit out another settler every 4 turns and quickly make up for the stalled expansion in the first 40 turns.

 

[templar_x]

i support everything that Lanzelot said on this topic. under normal circumstances this will be the strongest approach up to Deity. Sid often plays differently to me, because the AS will fill in any space, so I regularly produce early settlers just to block space. this space often would not be there any more 10 or 20 turns later to be settled by me at all...

if this situation arises on DG or Deity as well, that would be the only situation when I would build settlers from a small capital. otherwise, I am convinced that the way Lanzelot describes is WAY stronger. however, the both of us once played in a team in a cross-platform game where the team from the other platform we shared our continent with favoured the opposite approach (basically the one justanick describes), which was then partly the reason for war between our nations.

thus, while I stress that I believe this way is stronger, I also want to point out that I am not willing to go to war over that issue again!

t_x

 

[justanick]

Why do you think, settlers can be produced faster at 1-3 than at 4-6?! The opposite is true: At size 1-3 the town will on average have something like 3spt, so a settler needs 10 turns to build.

20 turns is more realistic if food is limited to +2. It is not shields but food that limits settler production.

At size 4-6 you should have like 5-6spt, so a settler can be built in 5-6 turns.

For that you need extra food first. At +2 food with granary you need 10 turns per settler. But 1 settler every 10 turns or 2 settlers every 20 turns makes no difference. If however net food production is higher granary will pay of in terms of producing settlers.

Plus you have higher commerce when running 4-6 (or even 5-7, when on a river),

Usually tiles have 1 commerce after building a road and a net commerce of zero because of luxury rate. Commerce wise no granary is usually the better choice. Granaries are best for cities that will later need no aqueduct and have high net food production. They can spam out settlers early and workers laters.

By that time I often already have roads in place, so the settlers can reach their destination in 1-2 turns instead of 3-4 when they get built "too early"... Think of it that way: you keep the population inside the capital, where they are highly productive, instead of walking them through the country-side for several turns, where they do nothing and only cost maintenance.

Maintenance is not the problem in despotism. Without granaries you need less workers and can have the roads ready in time more easily. Thanks to the city tile itself having one city size 3 or having 2 cities size 1 makes no difference. But the city tile needs less worker time and no luxury rate either.

As a general rule, size 5-7 is the best to operate the capital.

As a general rule reducing net food production by delaying settlers is no good idea. It can only become good by extra food from floodplains or resources like wheat. In order to utilize higher net food production there needs to be a higher shield production which will require higher population. But it should not exceed size 6 except when finishing the settler during growth to 7.

Another argument: the AI is building settlers at size 3 (as you can see in F11, when they often have pathetic little capitals even quite late in the game...) So it must be bad!

I tend to interpret this as one of few better things about AI. Sometimes AI manages to
utilize that in a proper way, but that is just luck. AI cannot properly adapt to circumstances.

And Circumstances is what this is all about. We merely disagree about what is the rule and the exception, aren't we?

As most players will reject positions with only +2 food granary first will be the better choice for the capital. Also granary first tends to allow higher production in the short run, freeing up quite some shield for military units like warriors that are important for exploration. Once those 2 exceptions are covered settler first becomes superior until either corruption or lack of land shift balance again.

 

[Lanzelot]

And Circumstances is what this is all about. We merely disagree about what is the rule and the exception, aren't we?

Could be... I think that in the "average" start position it should not be a problem to have like 3-4 tiles with +2 commerce. Having only tiles with +1 commerce would be the "exception" for me. (Take a look at the start position in my "Republicans" game: with just a short river running through my territory, I already have 9 tiles in Berlin's BFC, which will give +2 commerce!) For the early techs that gives approx. 50% faster research when I compare 1-3 to 5-7, even if I have to increase the lux rate by two levels!

And when you take one or two closeby luxury resources into account, then the balance shifts even more into the favor of the 5-7 cycle:
assume 5 tiles with +2 commerce and two luxury resources connected. On Emperor this allows 5 citizens at lux rate 0%. 6 citizens would require 10%. This gives 3 + 5x2 = 13 commerce at size 5 and 3 + 5x2 + 1 - 1 = 13 commerce at size 6. Spitting out a settler when the town reaches 3, would yield 3 + 2 = 5 commerce at síze 1 and 3 + 2x2 = 7 commerce at size 2. So 6 commerce on average compared to 13! Research will be more than twice as fast! (And still more than 50%, even if I add +4 commerce from the early additional town. Which is already optimistic, because one commerce in the new town will probably already be corrupted, so +3 is more realistic.)

Even if the 1-3 approach enables you to get more towns earlier, becoming a Republic 50% faster should more than compensate for that, because once you are a Republic, the REX capability really "explodes"...
And now let's take this one step further and imagine that because of the slower research you missed the slingshot...! Then you'll spend ages in Despotism and will never be able to catch up compared to the 5-7 approach. Not?

I think the key to a good expansion phase is not to have the maximum number of towns, it's to have the maximum number of productive citizens. Anything else (many towns, many techs, many units and a good government) will then automatically fall into your lap later like a ripe fruit.
(The exception of course being as templar already pointed out, those SID or Deity games, where you will need to claim land early, before the AI has settled all available spots...)

 

[justanick]

So 6 commerce on average compared to 13! Research will be more than twice as fast! (And still more than 50%, even if I add +4 commerce from the early additional town. Which is already optimistic, because one commerce in the new town will probably already be corrupted, so +3 is more realistic.)

At +2 food at no point in time granary first will allow more productive tiles than settler first. Any additional citizen of the capital is a productive tile less elsewhere. So it just comes down to corruption. Chances are that the second city is not corrupt. Later corruption will be relevant, but till corruption outweights the -1gtp from the granary quite some time will pass.

Delaying the first settler by 10(?) turns in favour of a granary will delay each settlers from the second city by 10 turns. It will also delay settler from those new towns by 10 turns. Corruption will sooner or later stop this but then those town will each grow 10 turns earlier in case of settler first. So unless by now i am more confused than i hope this mean that those towns have "permanantly" one citizen more than with granary first. Now guess what effect this will have on getting republic ASAP.

As republic those cities of more or less equal size harmonize well with the central luxury rate and if sufficiently small when entering anarchy no problems due to military police becoming ineffective will arise. During anarchy a capital of size 5 will have a problem with happiness if only 2 luxuries are available.

I think the key to a good expansion phase is not to have the maximum number of towns, it's to have the maximum number of productive citizens.

Usually the maximum number of towns means the maximum number of productive tiles. The city tiles are no citizens, but they must be counted in aswell.

 

[Theov]

Funny, after many years I ask myself this too.

I build it at 3 often. I never investigated what's quicker.
I feel that filling up the granary is the actual settler production, not so much the amount of shields. If that makes sense.

 

[Aabraxan]

If you haven't discovered "Cracker's Opening Plays," OP, this might be a good time to read it: http://www.civfanatics.com/content/civ3/strategy/cracker/civ3_starts/index.htm

 

[Puppeteer]

Since food is uncorruptible and there are so few improved tiles in the early game, it has always made sense to me to pop settlers as soon as possible, so size 3 until the FP message (10 cities on standard-sized map) or until out of immediately-productive land (unable to produce 30s and 40 food in 20 turns), whichever is first. But Lanzelot is a much better player than I am, so he's probably right. Also, on higher difficulty the AI start with military units and the barbs are more powerful, so REXing is riskier.

There is also a point where the core cities need to start producing commerce and military, and I never quite figured out how to do that smoothly. I always seem to either expand too slowly or be too underdeveloped.

Also, workers need to be worked in there, too, to road for settler travel and to get the core ready to produce. What I look for is a food bonus city and then 1spt edge cities to make early workers.

 

[Spoonwood]

What got missed here is the possibility of producing settlers size 3-5 instead of size 1-3. Size 1 is pretty bad, since it produces only 3 shields per turn maximum in despotism. 3+5+5+5+5+5+5 = 3 + (10 + (10 + 10)) = 33, and some calculation for the jaguar warrior, so you have 3 shields overflow on any produced ancient age unit. Waiting to produce a settler on the basis of food like the AIs loses time potentially building units. Producing a settler in 6 turns is possible with 2 bonus grasslands, but 3 shields is 6 turns on archer with growth back to size 2. A warrior-settler combo would make more sense with that approach. 1-3 seems rather awkward for producing both military and settlers/workers from a capital.

If you have three bonus grasslands, a size 3-5 capital can produce an archer in 3 turns, a second archer or spearman in 3 turns, and then a settler in 4 turns. The first archer has an overflow of 1, the second an overflow of 3 or 2, and the settler probably can get produced with no overflow. The size 3-5 city though is not so handy for producing a worker or two, but a tile swap with a nearby new city might work out well for that.

If you have +4fpt I think the answer is the higher sized city, because you'll have a capital that pretty much already gets to that size while building a granary.

For a +2 fpt capital, I think the answer here is kind of neither the 2-4 approach nor the 4-6 or 5-7 approach. The 4-6 or 5-7 approach is better once there exist some other towns to support the capital producing units. If you have 10 shield per turn potential for the capital, then the 4-6 or 5-7 approach I think better once you have a few other towns up without a doubt. Maybe 4 makes for a good rule since it also means that first army can get built. The 10 shield capital is simpler for producing workers and simpler all around. But, it takes a while to get there, and a whole bunch comes out missed out on food wise and production wise if you do that too early I think, and I say that after trying the "wait to size 5" a whole bunch of times.

But also, having a 10 shield capital site is by no means reliable. 4 bonus grasslands/sugars pushes beyond an average start, I think, and I think even 3 is just above average. To speak certainly if you have a fresh water plains and no food bonus start, not even a sugar, that's only 7 shields at size 6! In such a case, probably a 3-5 capital probably works better than anything else. Also instead of having more archers early, I think I'd rather have more cities with fewer archers, but also other cities putting out some capitals or making granaries of their own sooner rather than later.

If you're making swordsman or horseman or numidian mercenaries, then the higher production 4-6 capital I think in a better position. But still, no other cities growing than the capital stands rather weak for overall growth and production, especially if those other cities can share some tile here and there with the capital.

I didn't assume an agricultural civ in the above. If you play agricultural getting out a 10 shield capital becomes more viable in more starts with still 10 or sub 10 turn growth.