Ducki Does C3C at Emperor

[ducki]

with a greater production&growth capacity?

and more importantly to me lately, greater commerce and mobility from roads.

I never thought of it either - I read something by Aeson somewhere(maybe in this thread, I should reread it) that made me try it.

And yes, bonus tiles seriously affect the value of much of this stuff. By nature a 4-turn settler pump is a special case, but if you are presented with a situation that allows it and you don't know how it works, you can end up missing out on a huge advantage. Even if the circumstances don't allow a 4-turn pump you can often get "close" with a 6 turn pump that's not as efficient or as fast, but still hard to argue with considering the alternative of using every city to produce a settler at size 3.

I'm off to print this thread and reread all the good stuff the vets have said here (and all my ignorance. taking the bad with the good, heh.)

[Tolwyn]

Me too.... And maybe more important I'm going to collect all the important bits and pieces so I don't have to search a multi-page thread again

[ducki]

I need to try to do that again. I was going to try to craft a "how to build a 4-turn settler pump" article but had so much trouble extracting the juicy bits and then reconstituting them into a readable whole. Another, slightly less detailed, example of the 4-turn pump is Bamspeedy's "Babylon's Deity Settlers" article over in Civfanatics War Academy.

But I'll take another stab at a consolidation on this one unless you or someone else finishes one before I do.
Also, check out the AU501 thread(s) in the AU forum for another example of a 4-turn pump implemented by different folks on the same start if you haven't already.

[Arrian]

Well, a short version of such a 4-turn settler-pump article would list the requirements:

1) A city capable of a +5fpt surplus.
2) A granary.
3) 30 shields of production over a 4-turn span.

It's often that last bit that gets trickey. Something like this might do for an explanation:

The city needs to average 7.5 spt. Recommended size range for the city is 4-6 or 5-7 (the latter if the city has a river, for instance, since the extra citizen will almost assuredly pull in more commerce than it will eat up via the luxury slider).

Example: a city that has a grassland cow (irrigated, for 4f, 1s), a plains wheat (irrigated, for 3f, 1s), and 2 bonus grasslands (mined, for 2f, 2s) produces 7 spt at size 4. If it doesn't pick up extra shields when it grows, it will produce only 28 shields in 4 turns. That's not hard, since when the city grows to size 5 it will work another tile, which can be developed to produce +1spt, surely. A city that puts out 7 or more spt at size 4 is an idea settler pump.

The difficulty comes in with less ideal terrain. This often happens with floodplains cities. Tons of food, low on shields. In these cases, having a forest tile is a lifesaver. The game calculates things in this order: 1) commerce/tech, 2) food/growth, 3) production. Therefore, on a turn when a city grows, production is calculated INCLUDING the new citizen. Since the AI will normally autoallocate the new citizen to a high-shield tile if your city is running a high food surplus, the new citizen will be sent to a forest tile, and pull in 2 shields on that turn. Then you can reallocate the citizen to keep the city at +5fpt.

Therefore you can have a city that produces 6spt at size 4 or 5 but still manages to pump settlers every 4 turns. Starting at size 5 with 15food in the box (which I tend to call size 5.5), the city can produce: 6, 8 (6 + 2 from forest), 7, 9 (7+ 2 from forest). That's exactly 30 shields.

As Dominae once pointed out to me, forest isn't necessary if you have a plains tile and 6 worker turns to spare. Mine the plains and it's effectively a forest.

That's my attempt at a short version. Obviously one might wish to go into greater detail explaining things like what tiles to irrigate versus mine, how to best time the completion of a granary (it's best to finish it with an *almost* full food box), the benifits of forest chops, the use of the luxury slider, etc.

-Arrian

[Dominae]

Originally posted by Arrian
As Dominae once pointed out to me, forest isn't necessary if you have a plains tile and 6 worker turns to spare. Mine the plains and it's effectively a forest.

Or a Desert. Also, a Volcano or Iron Mountains fulfills the same purpose without improvement.

[Arrian]

Will the governor autoallocate a new citizen to a zero food tile like an iron mountain or volcano, though?

-Arrian

[Dominae]

Yup, as long as you end up with at least a 2 Food surplus.

[Arrian]

Ahh... gotcha. Hence my experience that it will autoallocate to forest so long as the city is at +3fpt when the city grows. The forest brings it back to +2fpt.

Edit: in a situation where you have both a volcano and a forest, which do you think the governor will pick?

-Arrian

[Dominae]

In Despotism, the Forest. Without the tile penalty...probably the Volcano, but I do not recall that happening in any of my games.

[ErikM]

Originally posted by ducki
In the majority of starts, I can build a Warrior in 5, then a Worker in 5(just in time to grow to 2 pop).

Not to nitpick, but in most generic starts a city will produce 25 shields by the time it grows to size 2. So the warrior/worker sequence wastes 5 shields.

It works like this (turn #/shields/food/comments)

 1  0s,0f // worker moves to bonus grassland
 2  2s,2f // worker mines (mine in 6)
 3  4s,4f // mine in 5
 4  6s,6f // mine in 4
 5  8s,8f // mine in 3
 6  10s,10f // mine in 2
 7  12s,12f // mine in 1. Worker job completes before production, so we get an extra shield
 8  15s,14f // worker road (road in 3)
 9  18s,16f // (road in 2)
10  21s,18f // (road in 1)
11  25s, Size 2 // we get an extra shield from a new worker.

This works very well for seafaring civs, since you can produce a warrior and a curragh in first 10 turns without shield waste in any sequence.

Then, assuming there is a forest nearby, on can put one citizen to mined BG and another one on forest for 5spt, or a worker in two turns. Two workers would then team up to cut some extra forests to speed up granary construction. Imo, this is an optimal sequence as it avoids any food/shield waste (although it trades two food for two shields on worker production).

If you cannot build a curragh, it is impossible to avoid some shield waste. The best sequence is probably warrior, warrior, worker:
1st warrior on turn 6
2nd warrior on turn 10
worker on turn 13
although it wastes 3 shields.

[ducki]

Great analysis, ErikM, looks like you are absolutely correct.

Except that I wouldn't limit the curragh thing to seaFARING civs, just seaside capitols. I could be wrong, but I think units(ships included) cost the same for everyone, so the seaside city of any civ is less wasteful with a Warrior-Curragh build with typical +2fpt growth.

It would seem I need to ratchet up the analyticalness of my starts just a tad.

[Theseus]

Originally posted by ducki
If desired, you can add the worker back into one of the towns after the granary build to grow faster so you can get your 3rd settler out faster.

Let us extend this thought...

Something I've noticed in, oh, the past six months, is that a number of experienced players are building a LOT of Workers, more than ever really, to dramatically decrease tile improvement time cycles, and thus increase overall production much more quickly... and then joining the Workers back to the best rivered cities to get up to 12 pop and waaaay high production in core cities very quickly.

Moral: If you've got the right terrain for it, build Workers... a LOT of Workers, more than you'd think.

[ducki]

I think the hurdle to get over - at least the hurdle I am struggling with - is the one about Workers really just being freely transferrable population.

I think of workers as Workers. I take population away from a town and spend shields so that I have a Unit capable to improving my empire visibly, physically, tangibly. That feels, to the old me, too valuable to add to a town and force myself to build that unit again, "steal" population from a town again, "waste" time and shields on it again. That doesn't feel very efficient to the old me.

The new me - the MMing, 4-turn settler pumping, cultureless military training camp building, no spearman having, fpt/spt averaging, power of population believing me - is just coming to grips with the concept of maintaining a worker pump for the express purpose of transporting cheap labor from one town to another where he is inherently two, four, six, maybe eight times more valuable(food cost, immediate commerce/production).

That version sounds very efficient to the new me.

Now to just get my head around it by doing it.
Do you have any experimental games in which you've tried specifically to explore this idea?

[ducki]

Speak of the devil

Here's a start that I think can support both a 4-turn Settler Pump and a 2-turn Worker pump.

I haven't looked closely enough to be 100% sure, and not having explored beyond the wheat, I have no idea of the true shieldcount of the site(s), but there's definitely enough food and I'm 99% sure enough shields.

Might be a good game to try out the whole "build workers specifically for pop-transfer" thing.

Game settings are:
C3C 1.15BETA
AU Mod 1.03b(flavored research), though I think you can play the save without needing the mod

Large, Random landmass, Temperate, 4bil
You are Japan.

[ducki]

And a screenie for those that just want to practice mentally, or while at work.