Originally posted by karmina
Well, the +100% are only true for 1,2,5 and 10 food surplus, considering a size 5-6 settler factory.
Because at least in Civ1.29f your stored food will always drop to zero when the town grows. So, if you get +7 food (wheat floodplains + 2 wines), a granary will essentially only give you a bonus of 3. (nonetheless a granary normally gives you the option to get far more shields at the same growth rate)
This is true of course, food gains of other than 1,2,5 don't work out exactly to doubling the town's food. But I feel that for the purpose of deciding on granary vs. settler they work out close enough.
For example, a surplus of 4 food/turn can be used in a town with a granary to grow every 3 turns, leaving a 2 food surplus per growth cycle. By sharing the food bonus tile(s) with another town this can be micro-managed. For example, if this +4 surplus is due to an irrigated grassland wheat, then one turn out of every three I would want the town to use a mined bonus grassland instead of the wheat, to gain 2 extra shields in that turn. Another town would use the irrigated wheat for that turn. Used this way the granary makes +4 food/turn effectively become +22 every 3 turns. That's a bit less than doubling which would be +24 every 3 turns, but close enough for comparing settler first vs. granary first.
Similarly a surplus of 3 food/turn can be micromanaged to come out fairly well. In one game (GOTM18) I shared a single one food bonus tile between two towns which each had a granary, resulting in fully doubling the single bonus food. Each town used it twice per four turns, resulting in both towns growing every four turns.
My feeling is that food surpluses over 5 food/turn in a town usually suggest that the town should have a granary and use just +5, and the rest of the surplus should be used to increase growth in another town. In this game there was +9 food/turn possible for the first town, a good setup for two fast growth towns.
Originally posted by karmina
This observation and the awful lack of shields around the starting spot led me to the keen assumption that a very early 6-turn Warrior+Settler factory without a granary might be better than ...
I agree that a very low shield location can completely change the decision. If it takes a very long time to build a granary then it is probably better to build settlers instead and send them somewhere where there are shields.
But I didn't think this start position was shield-poor at all. A capital NE of the start position had the wine hills (1 shield with no work), two mined BGs (two shields each after mining), a few grasslands (1 shield each after mining), and even one plains for another shield after irrigating. It also had a forest for two bonus shields per growth. And every citizen after the first one could work a tile which produced shields, assuming that the first citizen worked the irrigated wheat. Seemed like a productive location to me
Originally posted by karmina
My decision was also based on the assumption that on a standard sized map, 13 rivals would out-settle me like hell, only allowing me to found a handful of cities. I was proved terribly wrong. Who could have known that our three neighbors wouldn't expand at all?!
To me, early exploration suggested room for at least 10 to 20 towns depending on how fast the neighbors expanded. In a situation like that I figure it is worth going for the maximum. If growth is blocked early, it is still a big enough area that extra workers will be handy and the settler factory can switch to produce them. (Who ever has enough workers aside from OCC games?

) Or the settler factory could keep pumping settlers to stockpile for later use. I'm going for domination and that means I'll want lots of settlers later. My settler factory probably won't do anything for the entire game except pump settlers
Originally posted by karmina
Another question goes out to all you QSC gurus: Do you pop-rush your granaries or the occasional settler before? I figured it would take aeons to get a non-rushed granary in Karakorum.
I agree with Justus II. It is hard to overcome the population loss. The kinds of rushes he suggests are the ones I also use most. I occasionally rush something urgent but avoid it when I can.
Originally posted by Tone
Does that mean that it is worth the risk of leaving some towns unprotected and concentrating your military against the target civ when at war?
I agree with the points Justus II and Drazek have made about this. I do things pretty much the same way. I think I use even a bit less military than Justus II does, particularly less MPs.
I think there are three possible reasons to defend towns in the early game:
1) Barbarians. I tend to not worry much about them, I just consider them an unavoidable nuisance. Because I don't worry about them a lot I get hurt by them sometimes, occasionally losing a citizen, some gold, a worker, or a settler. Overall I figure I'm ahead of things with the amount I gain by not protecting these assets vs. what I lose. If they seem bad early in the game I'll build more military to hunt and destroy camps. If I don't see many barbs I don't build up for them.
In this particular game, I saw so many rival units wandering around that I figured I was almost completely safe from barbs - if any appeared an AI would pounce on them quickly.
One more thing about barbarians: their camps will not appear within one tile of your cultural borders and they will not appear in any tile which is visible to one of your military units. (Workers and scouts don't count as military units.) In most games I find that I can be proactive early in the game - by positioning a few warriors at strategic points (hills and mountains preferred) just outside my holdings, I can eliminate much of the barbarian threat. To me this zone of visibility approach is a better use of those warriors - less warriors are needed than to defend my towns and they are accomplishing more.
2) While at war with another Civ. For this defense I prefer an offense. If I am aggressively pressing the war into enemy territory, most of the enemy's forces will be used to try to stop me. A minimal defense at home is usually all that's required to deal with the occasional enemy unit which arrives there.
3) To deter new wars. To me this is the tricky defense issue, the one which is really hard to find a balance on. I'm usually pushing the limits I think. In Medal game 5-6 I pushed too far and lost my capital early in the game.
I don't know exactly what motivates the AIs in this area. The following are my guesses and thoughts. They aren't anything more than that. But acting on these generally works for me. Warning - pushing this stuff hard can occasionally lead to undesired results

Not often if you're careful though.
3a) Early in the game, while there is still good and accessible land to settle, the AIs are unlikely to attack. During this phase of the game the only thing which is worrisome is the next point. They're unlikely to attack for any other reason.
3b) When an AI has a unit positioned so that it can take an undefended town immediately (same turn) that strongly tempts the AI to attack, opportunistically.
Any military presence at all appears to eliminate this temptation. A single conscript warrior in the town changes the picture. It is just the "free" hit without any fight at all which seems to add a special temptation for the AI.
So, after the first 10 to 40 turns of a game (depending on how things look) I like to get at least one or two units hanging around close to the home area. They might explore just outside the borders but they stay close. Whenever a rival moves units toward my area, I move these home units so that they can occupy any towns which may become "threatened" before the rival can get there. I don't actually worry about defending the towns. A deity opponent in the early game can decimate me if it decides to attack. All I try to do is deter them by not offering easy bait. If there's no town they can just move into without a fight, they're less likely to declare war out of the blue.
3c) Undefended workers and settlers do not seem to add to temptation for the AI.
3d) Later in the game if possible I protect only border towns. If I can avoid rivals moving through my territory (by blocking or due to the nature of the position, e.g. being on a peninsula) then I leave no military units at all in any cities other than my borders.
3e) A military alliance is a good deterent. If an AI is allied with me against another Civ it is very unlikely that they will go to war with me. They don't even seem to be tempted when they have units passing directly past unprotected cities but I still prefer to avoid tempting them that way if I can.
3f) A ROP offers no significant protection. An AI which has a ROP with me will casually violate it if it sees a reason.
3g) Before declaring war on their own initiative (vs. being bought into an alliance against me by another Civ) an AI is likely to try extorting something. I try to always have a bit of money on hand which they can extort to ensure they have this opportunity.
3h) Later in the game I find the AIs' reasons for attacking hard to understand at times. Sometimes it seems to just happen randomly. Even then, I generally don't maintain much of a defense force. A mobile and strong offensive force can often be repositioned quickly enough to deal with the new threat. Another way to deal with a new threat is to ally one of the enemy's neighbors against them.
3i) If an AI moves a SOD (stack of doom, i.e. stack with lots of units) toward my territory, that AI has pretty much decided to attack someone. If there's no one on the other side of my territory in the direction the SOD is travelling, that someone is me for sure. And it might be me anyway.
Once the AI has decided on an attack of this type there is very little which can change its mind. It is time to prepare a defense. The only thing I've found which can change its mind in this situation is getting it to ally with me against someone else.
If you keep all of the above in mind at once

then it is possible to play the opening game with a very small military.
Originally posted by Offa
In the start, one decision that needs to be made is whether to emphasize food or shields, or take a middle course. For example early on is it better to work the hill-wine or the irrigated grass-wine? Is it better to work the irrigated grass-wine or the mined bonus grass? Does it make a significant difference? I decided to entirely emphasize food production without wasting food on growth, as the 2 bonus shields you get on growth (working the forest) will obviously be won quicker if you grow quicker. In essence really I would like to know what is the relative value of a shield compared with food.
This sure is a tough question. I think the only answer is "it depends"
But having said that, I almost always go the same way as you did this time, i.e. I emphasize food first. I only make exceptions when I have a particular reason to emphasize quick shields. If I don't know which to emphasize then I always choose the food by default. My thinking is that by emphasizing food first I'll end up getting more shields (and gold), partly due to bonuses at growth, partly due to getting more citizens and then assigning them to high production tiles while continuing to grow rapidly, and partly due to the ever present option to convert the extra citizens into workers or settlers when it seems appropriate.