Why granaries don't work

[Bamspeedy]

Well, yeah it would take 12 turns for a city going from size 1-2, but as you grow faster when you get to pop 3+, you'll be picking up more shields (unless the 3rd or 4th citizen doesn't have any shield tiles to use-which is not likely). So you will be slowly growing in size and probably eventually fluctuate in population size from 4-2 instead of 3-1 all the time.

I think if you only had 1 cattle nearby, then mining the cattle and then squeezing in a granary between the first and second settlers would be the best. Let's you build a few warriors/scouts, get the first settler out right away, then build the granary and it should be done before you get to size 3, so the second settler would only be delayed a few turns. And that is only 2 turns slower than irrigating it and building the granary, but allows you more of that early production (when you are at size 1).

[Ribannah]

Originally posted by vulture
If we add a granary to the mix, then it takes 8 turns to get the food, but still 12 turns to generate enough shields, so it takes 12 turns to generate a settler.

No, because your city will be working the higher-shield squares more often (and/or grow larger).

Obviously I'm ignoring the effects of mining the grasslands to generate more shields here.

There is the option of building a second Worker first, before anything else. You will catch up because thanks to the extra mines and irrigation the Granary (or the Settler, as the case may be) will be finished in less turns, and at the same time you will have roads finished to the next city location where you are also ready to start improving the land.

It all depends on how much work the original city location requires.

[DuncanK]

Originally posted by Ribannah
There is the option of building a second Worker first, before anything else. You will catch up because thanks to the extra mines and irrigation the Granary (or the Settler, as the case may be) will be finished in less turns, and at the same time you will have roads finished to the next city location where you are also ready to start improving the land.

It all depends on how much work the original city location requires.

Ribannah,

This will take you quite awhile to get your first settler out. Have you tried this, and are you sure you are getting your moneys worth out of that worker and the granary. It seems like the other nations will already have 3-4 cities when you get around to your second.

[Ribannah]

Yes, I have tried this. In almost all cases (crowded maps being the main exception) the Worker is definitely worth it. I typically catch up in empire size around 2550 BC and at that time I still have the extra Worker and double the tile improvements.

Whether the Granary is worth it, too, depends on the situation much like without the second Worker.

[DuncanK]

Ribannah,

How crowded is the exception? Four\Five players on a standard map? I'm not trying to be sarcastic. I'm just trying to get a feel for when I might want to try this strategy.

[punkbass2000]

Originally posted by DuncanK
Ribannah,

How crowded is the exception? Four\Five players on a standard map?

Uh, there are typically 8 players on a standard map. A scenario of your description would be quite empty indeed.

[Ribannah]

Originally posted by DuncanK
How crowded is the exception?

When there is the chance that by the time you want to found your third city, there is no more room.

[nbarclay]

It seems like a lot of the analysis and arguing here is in terms of one city with a granary or two without. But in reality, the more common situation is that the granary staggers things so part of the time you're one city short with a granary but part of the time you have two cities either way. That's what tends to tip the balance and make granaries worthwhile (especially for industrious civs, since they can get more production faster): you end up only part of a settler behind, not a full settler behind, but you get a population effect comparable to a full city's extra food production.

I tend to go after granaries in my earliest cities (especially my capital and any high-food city nearby) but then skip them until around the time I build aqueducts. The granary cities do a large percentage of my settler building, and the others focus their efforts elsewhere. With a good starting position, it's not rare for me to out-REX AIs on Emperor level (although my experience base with PtW is relatively limited).

By the way, as long as they have decent production, high-food cities are the ones that need granaries most in the early game. The basic effect of granaries is roughly equivalent to doubling a city's food surplus per turn. A city with a surplus of two food per turn gets the equivalent of an extra two food per turn with a granary. But a city with a surplus of five food per turn gets the equivalent of an extra five food per turn with a granary. Which is a better investment, 60 shields for the equivalent of two extra food per turn or 60 shields for the equivalent of five extra food per turn? Of course once a high-food city gets its granary, it may not get a chance to build anything but settlers and/or workers until after you're finished REXing, but there are worse fates in life than having lots of settlers and workers.

Nathan

[DuncanK]

Nathan,

It you have a square that produces 4 food and no sheilds you can produce nothing but settlers without a granary.

[ducki]

Unless that's floodplain, I'd rather have a square produce both shields and food, partly just to have balance, as shields are harder to come by in the early game.

[DaveMcW]

DuncanK, I already went over that in my big math post.

It is better to make nothing but 6-turn settlers using a granary instead of nothing but 10-turn settlers without a granary. If you don't have the shields to keep up with growth just raise the luxuries a bit and let the city grow until you do.

[nbarclay]

Originally posted by DuncanK
Nathan,

It you have a square that produces 4 food and no sheilds you can produce nothing but settlers without a granary.

I hadn't analyzed it before, but if the only food bonus within a city radius is an irrigated wheat, a granary seems to be roughly a break-even proposition - or a losing one if you slow down settling nearby land that has another food bonus. That's due to a combination of three factors. (1) There's a major loss due to rounding with a granary but not without (the granary only gives a 66% effective increase in food instead of 100%). (2) With only one food bonus tile, the city can build its settler and drop back to size 1 without losing any of its growth rate. And (3) not enough production is available to get the granary finished quickly. A cow tile with no grasslands with shields around faces the same basic problems.

If your capital can get a surplus of five food and (simultaneously) at least seven production at size 5 and eight at size six, it's crazy not to build a granary there. In other cases, a lot more depends on the specific situation because happiness issues and availability of food bonuses elsewhere can both affect what choice is optimal. (Capitals get a little extra gold, so using the luxury slider to quell discontent in them is more efficient than in most other places.)

Nathan

[Bamspeedy]

Yes, use the luxury rate. As long as you'll get the 30 shields before hitting size 7, you'll be ok. At size 7, the granary gets emptied, so you don't want that. You'll be ok if you produce the settler at the same time you would hit size 7, just don't spend one full turn at size 7.

Typically for a +5 food city you'll spend 2 turns at size 5, and 2 turns at size 6. So you'll need 7 shields/turn at size 5, and 8 shields/turn at size 6. You need 2 bonus resources or wheat on floodplains for this. The problem with floodplains, though is that there is usually nothing but plains/desert around so you may have problems getting the shields (since trying to get a 2-shield tile like forests will slow down your growth).

A +4 food city just needs an average of 5 shields/turn (irrigated cattle on grassland, city center tile, 1 bonus grassland, and 1 more bonus grassland or a mined regular grassland). If you have wheat on grassland and you irrigate it, you just need 2 mined bonus grassland tiles.
Food is harder to come by in the early game, not shields, because there fewer tiles you can irrigate to actually get a benefit from.

Building a worker first: I would only think about doing this if I had a bonus resource on grassland and I irrigate it and if I'm a non-industrious civ. If you irrigate the resource on grassland you'll grow every 5 turns (without a granary), so building the worker puts you behind by 5 turns. Cutting a forest would be more viable with 2 workers, so that would help it get caught up if chopping the forest for a granary. It really does depend on terrain. On some starts the extra worker only speeds up the granary by 1 turn, so you are now 4 turns behind instead of 5. So then it really depends on what the other cities can do with the extra tile improvements. Can your first few cities get enough of a benefit from the improved tiles to make up for being built 4-5 turns later? Having more roads in place will help some cities actually be settled a little earlier, so that will help a little bit. I still don't feel comfortable doing this. I may test it further.

[nbarclay]

Originally posted by DaveMcW
DuncanK, I already went over that in my big math post.

It is better to make nothing but 6-turn settlers using a granary instead of nothing but 10-turn settlers without a granary. If you don't have the shields to keep up with growth just raise the luxuries a bit and let the city grow until you do.

With a low production rate, the extra time involved in building the granary and then the settler puts you essentially a full city behind. Even if the second city has no food bonuses, the food advantage of the granary over the second city is only about 10%, so you won't catch up in total number of cities settled until at least somewhere around tenth city. That's not much better than a breakeven proposition at best unless the map is sparsely populated (although the granary does leave a bigger, more productive city behind at that point). Worse, for whatever duration the earlier second city does offer an extra city with a food bonus, overall population growth is actually higher with the extra city than with the granary. And then there's the granary's upkeep cost and probably gold for the luxury slider along the way.

[Ribannah]

Originally posted by nbarclay
Even if the second city has no food bonuses, the food advantage of the granary over the second city is only about 10%

The way to count how much you're behind (better: how much you have invested) is not by food or cities, but by the number of turns. Let's say the first Settler you build is completed 8 turns later than without the granary, then the second one arrives 4 turns later, and on average you have caught up in empire size when your capital produces its 5th Settler 8 turns ahead, unless your second city is also a Settler factory. In that case there is a new decision tree where you might choose to invest further, and there is the (then usually better) option of building one Settler before the Granary, too.

There are variations where your capital alternates between building a Warrior and a Settler, while the other cities start with Temples or Workers and don't need to help with the police force.