1. We have added a Gift Upgrades feature that allows you to gift an account upgrade to another member, just in time for the holiday season. You can see the gift option when going to the Account Upgrades screen, or on any user profile screen.
    Dismiss Notice

granary or settler first?

Discussion in 'Civ3 - Strategy & Tips' started by Rustwork, Jul 7, 2010.

  1. Opies

    Opies Warlord

    Joined:
    Jun 6, 2010
    Messages:
    142
    Location:
    Canada
    Im currently playing my first emperor game (standard, all vics, iroquois) and I went granary first. I focused my first few cities towards the AI to build some borders, then filled in the inside with tight cities about CxxC. At the end of the REX I had about the same amount of land as the AI, maybe slightly more than some of them, yet I had probably double the number of cities due to the AI's wide city placement. I think that is one of the things a human player needs to take advantage of. Build the wide placement like the AI, but then fill it in with tighter cities so you can surpass them in production.
     
  2. Spoonwood

    Spoonwood Grand Philosopher

    Joined:
    Apr 30, 2008
    Messages:
    4,849
    Location:
    Ohio
    I agree that can definitely help. Though, when you move up a level, especially if you step up to Demi-God or from Deity to Sid, keep in mind that the AIs will grab territory nearer to you sooner.
     
  3. Link

    Link Scarves

    Joined:
    Sep 20, 2009
    Messages:
    1,706
    Impossible to say. It entirely depends on your situation.
     
  4. Raliuven

    Raliuven Emperor

    Joined:
    Jun 21, 2010
    Messages:
    1,428
    Location:
    Minnesota
    Regarding the early land grab - when I ran my test one thing surprised me, though it could just be an RNG thing. When I ran the first test, I didn't produce any warriors until something like the 30th turn (I don't have my notes at work, but that is a close approximate). When I ran the second test, I built warriors while waiting for my population to grow to start a settler build (timing it so that I would gain the settle the turn my city hit pop of 3). The end results were that the second build, obviously, had more warriors (7 vs. 3). But I also had them earlier.

    In the first test, my nearest neighbor (America) was already growing towards me and we were fighting for land by the 4th city (I could even have forced the issue at the 3rd city). In my second test, by the 5th city America hadn't grown in my direct at all - the closest city to me was still his capital.

    So, for the question - does early military strength deter the AI from settling in your direction if they have another choice? I haven't tested this enough yet so I thought I would ask for antidotal evidence - or maybe someone has done the work and figured this out. I know that the AI will eventually file in any space that hasn't been claimed, even that annoying two tile tundra on the backside of your empire. Or did the RNG just play a trick on me? Perhaps the AI sent his scout in a different direction?

    At higher levels this probably doesn't matter since the AI can outproduce you. Still I am curious.
     
  5. ThinkTank

    ThinkTank RL Addict

    Joined:
    Feb 2, 2007
    Messages:
    1,085
    Location:
    An inaccessible possible world
    In early game the AIs often send out settlers without escort if they can. But my impression is they will not do that if there are foreign miltary forces around. I use this sometimes by fortifying a warrior in a productive area that may be contested by an AI, so that they will either move their settlers somewhere else or at least have to build an escort for the settler first, buying me some time to get there first. This gets less effective at higher levels as they have more start units, and units are cheaper for them anyway.
     
  6. Raliuven

    Raliuven Emperor

    Joined:
    Jun 21, 2010
    Messages:
    1,428
    Location:
    Minnesota
    That makes sense. Since I ran the test on Cheiften to avoid an annoying AI influencing the results, I set up the conditions for this to happen. Because I was cranking out early warriors and a spearman cost them what, 40 shields (?) and the settler cost even more, they didn't have the resources to build towards me in an 'endangered' area. Wereas with my zero army count in model #1 at turn 30, the AI wasn't as concerned - I wasn't a threat.

    I'm not sure how (or if) that can be applied to the higher levels, but it is interesting. I am running other models and I will keep this under observation. Thanks ThinkTank!
     
  7. templar_x

    templar_x usually walks his talks

    Joined:
    Jun 15, 2008
    Messages:
    2,647
    Location:
    on a learning curve
    my perception is that it is not merely the presence of MILITARY that makes the AI build escorts, but the presence of HOSTILE military.

    if they are not at war, i have seen them send out settlers alone quite a bit into the game already. only if barbs are on, they will not send single settlers out from the time on when they appear.

    templar_x
     
  8. Hellfiredoom

    Hellfiredoom Warlord

    Joined:
    Dec 9, 2009
    Messages:
    222
    Location:
    Seattle
    Just to add my experience to the off-topic topic.. :lol:

    In the early game when I'm scouting the map with warriors and trudging thru enemy territory, sometimes a lone settler comes into my field of view and I see them scamper back to their city. So at least in their territory or close to it, they send out unescorted settlers. Every time I see one, just for a split second I think "ooo, slaves!", then realize the AI would demolish my meager empire. :D
     
  9. Raliuven

    Raliuven Emperor

    Joined:
    Jun 21, 2010
    Messages:
    1,428
    Location:
    Minnesota
    I've never had a moderator strike me down with off-topic wrath before (tries to scamper back into his hole . . . telegram from the military advisory to Mrs. Raliuven "we regret to inform you that your husband is dead because he was stupid"):lol:

    It's not completely off topic - I am talking about the strengths of both starting strategies!!

    That tracks. I had several warriors in and around the disputed area for most of the test. So IF that means they would not send out unescorted settlers, then no unescorted settlers = a much slower expansion - at least in that direction. So did they send their unescorted settlers in another direction? I think I need to open that test game and check out the Americans in both models to see how they developed. Maybe they ran into other trouble or had barbarian problems, thus no unescorted settlers. We were at peace the entire time.

    Thanks!
     
  10. Hellfiredoom

    Hellfiredoom Warlord

    Joined:
    Dec 9, 2009
    Messages:
    222
    Location:
    Seattle
    I should add I play without barbs/goodie huts to slow down the AI's runaway tech pace. They simply have so many starting units for scouting and trade so quickly I have to do what I can to reduce their advantage. So they have no wandering barbs to worry about for their unescorted settlers.
     
  11. Raliuven

    Raliuven Emperor

    Joined:
    Jun 21, 2010
    Messages:
    1,428
    Location:
    Minnesota
    Okay, my faith in the AI's inflexibility has been restored. I examined the opening game sequences and it is clear that America popped that 'intruding' city from a GH. The first settler created by the AI settled in the same place both times - which just happened to be directly away from me. So in one game it looked like they were advancing on the good middle land, when in fact they just got a lucky GH. Sorry - should have considered that posibility before spinning off on a no-so-off topic topic (just in case the moderator is watching).

    Of course, if there were no GH for the AI in the first place . . . seems I've heard that suggestion somewhere around here . . .
     
  12. bhavv

    bhavv Glorious World Dictator

    Joined:
    Jun 13, 2006
    Messages:
    7,358
    Wait, I thought that people usually make a worker first, then grow the city, then make settlers?

    Thats what I always do.
     
  13. Raliuven

    Raliuven Emperor

    Joined:
    Jun 21, 2010
    Messages:
    1,428
    Location:
    Minnesota
    I have continued to run tests though I have been derailed by GOTM105, so I am only part way done. I have generally found that producing a worker has not resulted in a better REX than producing a granary first. The worker vs granary approach produces about the same number of military units. Settler first results in the slowest REX but the biggest army. Also, the worker first results in the same number of workers or LESS than for the granary first approach.

    My test environment is two islands with the AI banished to a small island while I have my land mass to myself. There are no resources or luxuries on this world. My love is the industrious Civ (I only have PTW), so my test is Granary first, Settler first, Worker first - and I run the test for non-industrious Civ and then an industrious civ. I play the first 60 turns and record all production and builds. They are both expansion Civs (thus starting with pottery for granaries), and religion is assigned to the non-industrious Civ (Iroquois vs. America). Then I add Food Bonuses within a given distance and play all 6 variants to compare the results.

    I only have 10 test games done (out of 36 planned), though. Quicker worker production could help get a settler factory working earlier, which is a test I have not run yet. They could also prove useful with a quick luxury or strategic resource hookup - but that is outside the scope of my current test. Contrary to what I had expected, at least within the first 60 turns, the added workers turns did not result in a stronger food/shield production than the granary approach. The added citizens outweighed or matched the bonuses of the availability of improved tiles.

    At this point, I would surmise that without a food bonus within the radius of your start position, the worker first approach will not produce the best REX. I would even go so far as to extrapolate that unless your capital has the potential to be a settler factory, then worker first will not produce the best REX.
     
  14. Lanzelot

    Lanzelot Moderator Moderator

    Joined:
    Jun 25, 2007
    Messages:
    5,703
    Location:
    Heidelberg
    This is a very important observation indeed. It supports what I've been trying intuitively in a couple of games lately: 1-2 warriors first (or curraghs, if seafaring), then a granary, and by that time I would have hit size 6 and start with the settler/worker production, while at the same time having a size 5-6 capital with lots of income.
     
  15. templar_x

    templar_x usually walks his talks

    Joined:
    Jun 15, 2008
    Messages:
    2,647
    Location:
    on a learning curve
    i do not think anyone would believe "worker first" could help your early development anything if there are NO RESSOURCES around. as i understand it, this is the environment you currently do your tests with.

    with no ressources, your capital will grow only every 10 turns without a granary. your starting worker, especially if you are industries, will easily be able to hold pace with the city growth then.

    so worker first is an option that logically disqualifies in that environment you describe. settler first in my eyes as well, if you read my entry on that topic that i believe that an early settler only makes sense if you have another factory-like site around.
    this is, because a settler before granary is expensive, costing exactly twice as much food than it does after the granary.

    without a food bonus, i would always build a granary first, if the environment permits it, probably after some units first - there is no hurry at 10 turns per pop point. (life is different on a crowded sid pangaea, i can tell you, that you will build your first settler asap there, otherwise you won´t have a 2nd town...)

    templar_x
     
  16. Ataxerxes

    Ataxerxes Deity

    Joined:
    Jul 29, 2009
    Messages:
    3,073
    Are you thinking of CivIV? In CivIV, that's the usual strategy because you don't start with a worker.

    In CivIII you start with a worker. Building a second worker first works horrible. The loss of pop that early hurts. It gets worse if the hammers are finished first and you wait for the second worker because you need Pop 2.
     
  17. Raliuven

    Raliuven Emperor

    Joined:
    Jun 21, 2010
    Messages:
    1,428
    Location:
    Minnesota
    I was inspired to try this based off of your observation - that is why I am trying a few different angles. Oddly, the non-industrious no food bonus, granary first has resulted in the highest available bank up to this point, with an average income of 6.57 gpt over the first 60 turns. Thus far. Obviously addition research (26 more games) is needed for a conclusion.

    At the risk of violating forum rule #137 or sounding defensive . . .:)

    This is the test environment I am using - which is why I said these were out of scope. My logic is that strategic resources are not part of the test - they help you build better units, but don't affect REX in the strictest sense. Yes, they defend your empire - no question there, which is why I qualified my earlier statement, noting that if you need military immediately for protection, then settler first is a good (best?) strategy. They can add shields and gold, true. But this is no more valid than luxuries below or having a cow or rivers nearby.

    I removed luxuries because these have indirect impact on REX. All of these effects can be compensated for with the luxury slider and it gives me a clearer picture of the bare costs each turn. You can 'figure' the added benefits of luxuries from there. Less use of the luxury slider, increased gpt from 'worked' luxuries, etc.

    By the way, for those that 'fear' the luxury slider, the past 10 test games have cured me of that problem! Bottom line, my reasoning is that all of the effects of these 'bonuses' can be layered on top of the raw, base data. My focus is on the raw REX data.

    With that said, I do not doubt what you say about the worker first strategy, however I was inspired to test this based on the idea that I could use the early worker to 'set up' new city sites. Building roads to new settlement areas can result in building a city 2-3 turns faster. That is the hypothesis.

    I don’t completely agree with your logic here. I’m not sure if you intended to imply that the presence of a luxury or strategic resource might cause you to generate a worker first?
    To address strictly food bonuses, I am adding these in increments. The no food bonuses anywhere is the ‘control group’. The overall test is to try to determine under what circumstances it pays to use a settler first strategy. How close does the nearby 'factory site' need to be to make sense? Can a worker first then settler overcome a granary first approach if the 'factory site' is outside the first 'settling' radius? How many food bonuses near the capital makes it worth considering a settler first strategy? Or a worker first strategy? Is single bonus wheat on a grassland with access to irrigation sufficient to tip the scale? Is a settler factory capital site or a factory site nearby make worker generation a higher priority? What if you are trying to maximum gpt for fastest research?

    I think that some of this is counter intuitive which is why a question like this comes up rather often. You see a city with no food bonuses and you think "well, that doesn't have potential" and discount a granary. You see a food rich city and build a granary right away. At least that was my inclination, when in fact; it was the food poor cities that could most benefit from the granary. And because each newly founded city adds a layer of complexity, the problem grows exponentially and new players have a hard time extrapolating how a decision made 60 turns ago multiplies. How much of an effect does founding a second city 10 turns earlier in the REX make a difference? This is what I am trying to figure out.

    I don’t doubt your experience or skill (see rule #137), but I think these questions have merit and I’m not convinced that there is a clear cut answer at this point. I will fully admit I am probably a little obsessive on this topic, though. Okay, that covers my posting limit for the day.:D
     
  18. templar_x

    templar_x usually walks his talks

    Joined:
    Jun 15, 2008
    Messages:
    2,647
    Location:
    on a learning curve
    @ Raliuven - i am only talking FOOD ressources here. you need only food and shields for REX.

    to be clear on that, i find it great you´re making this test and surely do not want to devaluate it by any means. i am only suggesting - more for the readers than for yourself - not to get carried away with results of the control group now.

    imo, IF you´ve got only +2food places around (and not playing with very limited land or 20k or sth like that), so you want a normal REX, then worker first or settler first hardly will be better than granary first. the limited food is just too expensive. (if you do not have a single BG, it might be different)

    the most interesting part of the test in my eyes is, how good a 2nd (ord 3rd..) city site must be to justify a settler build before the granary. regarding the worker build, it is pretty easy to answer this for your start simply with the starting spreadsheet.
    but every option really has to be played out optimally, otherwise it only shows which playing routine one player masters best. so you might even want to have 2-3 players play the most important variations.

    re your questions, my $ 0,02:

    the corruption must no make it impossible to run it as a factory with sufficient shields.

    sure it can, but it´s quite a work on a spreadsheet to prove it in any single case. it is especially possible if there are lots of improvements to install to make both factories run.

    it´s not so important how many food bonuses are near the capital, but what the 2nd city site looks like. you basically barter worse early development of the capital for earlier development of the 2nd town. so it should be worth it.

    for worker first see above, i´d say it´s a matter of how many turns of improvements your factory tiles need. if you only need to irrigate one tile and the factory is done, you won´t build another worker!
    fast early research and fastest possible REX to some extent contradict each other. if you want beakers right from the start, you cannot allow your capital to shrunk in the early game.

    i have very limited time, but i would be ready to play a couple of variations. but please for those that are "close", like comparing both approaches with 2 factory-like city sites.

    templar_x
     
  19. Raliuven

    Raliuven Emperor

    Joined:
    Jun 21, 2010
    Messages:
    1,428
    Location:
    Minnesota
    @templar x - I do appreciate the constructive criticism and just to validate your concerns - this is very preliminary data and is mostly on the test group. Also, there is something to be said for the quality of the player. I am quickly gaining confidence with Emperor (PTW) and I am looking towards Deity, so anyone reading should take that into consideration.

    Once I dispatch GOTM105, I intend to turn back towards the project, though my wife did just buy me Civ4 complete for my birthday. :D As I drill down into the more 'interesting' tests, I will keep your offer in mind. I would love to see your 'optimal' play vs. the test groups on the key tests - also recognizing that the level of detail I record does take 4+ hours per test. :sad:

    I would also like to pull out this particular gem:

    I hadn't thought about it that way. That is actually more insightful that it might seem at first glance. This will be part of the test by design, but I hadn’t actually thought of it like that. Dumb luck there, so at least no need to redesign the test! Obviously you are going to weigh the pros and cons to an approach, but I tend to froth at the mouth at a good capital site and all focus beyond its unimproved borders becomes blurry . . .

    I had wondered about that, which is why I decided to keep close track of the money. I do keep research at a minimum (40 turns) during the test - in hindsight it might have been better to turn it off for a better 'raw data' picture. 40 hours into the test, I decided to just keep it a constant. I do track gross income; expenses and net income for each turn (and for each city for each turn). It will be interesting to see if the Granary first approach always generates a higher base income - and thus more research power. Of course, I am only tracking for 60 turns, so it could be that the exponential growth could flip things at 100 turns. I might do a more extensive test based off the findings this time around. Did I mention my obsession? :lol:
     
  20. Lanzelot

    Lanzelot Moderator Moderator

    Joined:
    Jun 25, 2007
    Messages:
    5,703
    Location:
    Heidelberg
    Independent of the purpose of this test (finding out what is best for fastest REX), I would like to add that sometimes other strategic factors influence the decision.

    For example, if you play 20K and want to use the 2nd town as 20K site, you of course don't build a granary first...

    Also I remember one case, where mining a hill was necessary to get a four-turner going. I think I built a worker before the granary in this case (and used the hill for speeding up the granary).

    Thanks for doing this research! Looking forward to the outcome. :goodjob:
     

Share This Page