Windmills: kings of the late game?

Discussion in 'Civ4 - General Discussions' started by Iranon, Feb 20, 2009.

  1. Iranon

    Iranon Deity Whipping Boy

    Joined:
    Nov 14, 2007
    Messages:
    3,214
    Location:
    Germany
    For a long time, I regarded windmills as a niche improvement... mostly used for marginal cities that couldn't grow to a decent size without, in the rare game where Environmentalism actually looked competitive compared to State Property or corporations under Free Market, or to get the FIN bonus on riverside hills.
    Recently I've come to appreciate them very much even without the bonus, since they make almost any heavily optimised economy stronger:

    ***

    When running State Property+Caste System, relying on farms and mines is wasteful: workshops -> farms and mines -> windmills both convert hammers into food at a 2:1 ratio, but in the case of winmills we get some :commerce: thrown in to sweeten the deal.
    Hence, mines are only used in cities with excess food in which we don't wish to run (more) specialists.

    ***

    In case 1:hammers:7:commerce: towns are available, things become quite interesting for lategame optimisation. Standardising for food neutrality for 4 tiles, 2 grasslands and 2 grassland hills:

    1 farm, 3 towns gives us 5:hammers:21:commerce:: running as many of those yummy towns as possible is definitely strong!

    1 farm, 1 town, 2 mines gives us 9:hammers:7:commerce:. Giving up 3.5:commerce: per :hammers: is not good; more towns and rushbuying is almost strictly better unless we have compelling reasons for wanting straightforward production (Heroic Epic + Military academy, needing to build projects).

    2 towns, 2 windmills gives us 6:hammers:18-22:commerce:. Outright better with Environmentalism; worth considering even if we don't plan to end up there since we can grow the towns early (before Biology and when caps are restrictive) and work the windmills later - unlike cottages, they start out strong.

    *

    Now the same situation, only with plains hills - uncottageable and requiring a full Biology farm in support.

    2 towns, leaving the hills idle: 2:hammers:14:commerce:.

    2 farms + 2 mines: 10:hammers:. Is this better or worse? Not so sure myself but not as blatantly suboptimal as grassland mines.

    1 farm + 1 town + 1 mine: 6:hammers:7:commerce:... the compromise between the options above.

    1 farm + 1 cottage + 2 windmills: 7:hammers:11-15:commerce:. Generally the most attractive option, with or without Environmentalism. If the caps aren't an issue, this is probably worth it even with desert hills and without Environmentalism (5:hammers:11:commerce:).

    ***

    If I set my economy up for heavy Kremlin-assisted whipping, I can be assured of a decent return of hammers on spare food. Even at size 20 (average during regrowth so maximum in the low 20ies), 1:food: whipped is theoretically worth 1.5:hammers and getting 4:commerce: for every :hammers: sacrificed sounds good to me.
    This depends a little on the state of our growth caps and the quality of the marginal citizens to be whipped away (usually specialists). The latter isn't a problem under Representation as the average specialist output is worth more than 3:hammers: in most circumstances.
     
  2. noto2

    noto2 Emperor

    Joined:
    Jul 11, 2008
    Messages:
    1,711
    It really depends what civics you are using. For example, if running SP, windmills + watermills is much better than farms + mines. In another example, if you have 1 grassland and 2 grassland hills, 2 windmills + 1 town beats 1 farm + 2 towns, but only when running Environmentalism. You might not want to run that because it means you forgoe the benefits of other economic civics, like SP, FM, or Merc.
     
  3. pi-r8

    pi-r8 Luddite

    Joined:
    May 1, 2006
    Messages:
    2,557
    Location:
    Babylon
    Just came across this awesome post and I thought I'd bump it since it never got any attention.

    Also, I have a question. Is there ever a situation where you would want to build windmills early in the game, just after getting machinery? At that point they only generate 1 :commerce: 1 :food: so they seem pretty useless. On the other hand, a mine only generates 2 :hammers: , and I know some people use the rule of thumb 1 :food: = 2 :hammers:, so by that logic the early windmill is like a mine but with 1 free :commerce:. Perhaps it could be useful in a small city dedicated to whipping.
     
  4. Iranon

    Iranon Deity Whipping Boy

    Joined:
    Nov 14, 2007
    Messages:
    3,214
    Location:
    Germany
    Hmm, I remember being slightly disappointed with the lack of interest :)

    With a Financial leader, I actively look out for riverside windmills. Otherwise, it really depends on the local food situation and available land. Having mostly mines in my good cities (which hog the food resources most of the time) and plenty of windmills in the filler cities would be quite typical - I don't like supporting mines with vanilla farms at all.

    I've also had good successes setting up windmills + lumbermills.
    It works quite well but may not be too enjoyable unless you have a micromanagement fetish: In the early game, forests and windmills can compete only when backed up by efficient whipping cycles and sustainable chopping.
    Once you have Replacable Parts, you have health-boosting mines on flatlands and something better than farms on hills. This has some serious lategame potential with Environmentalism.
     
  5. pi-r8

    pi-r8 Luddite

    Joined:
    May 1, 2006
    Messages:
    2,557
    Location:
    Babylon
    sustainable chopping? you mean you actually wait for forests to regrow before you chop them down? Doesn't that take forever?
    I do like the idea of building mines on flatland and farms on hills, though :)
     
  6. fed1943

    fed1943 Emperor

    Joined:
    Nov 28, 2005
    Messages:
    1,185
    Location:
    Lisbon
    I do remember the 1st post and studiet it, just did not answer.

    IMO the esay way to think about, is not what a "cottage", mine, farm, windmill, gives but what the tile gives; and what the city/civ needs.
    Grass/hills or plain/hills? River or not?
    And time factor: "cottages" time is just turn time; all others just research time.
     
  7. Higher Game

    Higher Game National Socialist

    Joined:
    May 31, 2002
    Messages:
    1,076
    Location:
    GA, Amerikwa
    People get stuck for most of the game making specialized cities that heavily prioritize commerce or production. National wonders, the cost of infrastructure, and the distribution of resources all favor this. Balanced cities that can do everything well are expensive, rare, and tend to come when you've already won.

    There's no reason for a financial civilization not to use windmills though, especially ones that lack a health bonus and have an incentive to use environmentalism anyway. If you're bumping into the health cap, mines start to look better. Windmills are also a strong choice for post-astronomy colonies that want food.
     
  8. NihilZero

    NihilZero WHEOOHRNY

    Joined:
    Jul 20, 2009
    Messages:
    1,161
    Location:
    Dublin, Ireland
    My simple rule of thumb, post Replaceable Parts, is: windmills on pretty much all hills in commerce or food/specialist cities (maybe with one mine to round off a food surplus); in production cities, enough windmills to give me food to run more workshops and increase total net production (after growth). The commerce bonus is fairly meagre if I don't have any commerce multipliers in the city.

    In the first case (commerce and food cities) I'll usually delay getting the windmills up if I have some important infra I want sooner rather than later.

    As for whether Windmills are kings of the late game, I think the obvious answer is that SP, post electricity Watermills are. Watermills are absurdly good. Before Biology I convert as many farms as possible into them and lose no food at all! :eek:

    Occasionally. Typical scenario: if I put down a desert city somewhere to claim resources and the only way I can squeeze out one more population point is to windmill a desert hill. It's equal to an unimproved riverside plains, ie pretty terrible but still better than nothing.
     
  9. Iranon

    Iranon Deity Whipping Boy

    Joined:
    Nov 14, 2007
    Messages:
    3,214
    Location:
    Germany

    You'd chop about half of them immediately, in the form of a checkerboard... preferably in a way that you keep mostly flat grassland forests. When the tiles you originally cleared regrow, chop again - this is the sustainable part.


    The idea is that anything food-neutral, like grassland forests or coast, is a perfectly usable filler tile during a whipping cycle, especially with Representation:

    Grow to unhappiness working resources and filler, whip 2+ citizens away
    Work resources and filler until about to break the new cap
    Change filler to specialists
    Change specialists to filler once anger wears off
    Repeat.


    Naturally, filler tiles are worse than improved tiles... but since this cycle doesn't work its average tiles all the time, the quality difference isn't very important and may be overshadowed by the advantages of keeping your forests: a nice health bonus if you leave them all, a smaller health bonus and sustainable chopping if you choose a checkerboard pattern.
    Don't be dogmatic about this - if your regrow cycles are too long, by all means include a farm or two, especially in a place where you can't keep up a nice checkerboard pattern. The point is that not improving everything can be a reasonable choice... especially if you intend to run a Hippie Economy.
     
  10. karadoc

    karadoc AI programmer

    Joined:
    Oct 3, 2005
    Messages:
    1,568
    Location:
    Australia
    I suspect he didn't really mean sustainable chopping, but rather well rationed chopping. Forest growth is too slow and unreliable to be sustainable. [edit. Iranon's post came in while I was typing... apparently he really does mean sustainable!]

    I haven't really thought much about early windmill usage. About the only time I use early-game windmills is when I have a city that needs food and can't get farms. Also, I never build towns on hills... I almost always build mines on hills.

    Similarly, I typically only build watermills in cities that are desperate for productivity. Although, if I get state property then (obviously) I'll build watermills over the top of any farm that isn't required for irrigation chains. Only in rare cases will I even replace a riverside town with a watermill.

    My strategy in this area could probably use some improvement.
     
  11. vicawoo

    vicawoo Chieftain

    Joined:
    Feb 12, 2007
    Messages:
    3,226
    Much better than the current threads, although more of a strategy forums post. I don't think I read much general forums at the time.

    with 3 to 4 adjacent forests, you have a reasonable forest regrowth rate. With a lot of those, you don't exactly know where it will be, but who cares, free hammers.

    Windmills for plains hills and desert hills is strong, since even biology farms are inefficient compared to watermills/workshops/lumbermills.

    Must digest 1 food numbers.
    Edit: yeah, in state property/caste, that's correct. If you don't run caste, one farm and 2 grassland mines is 8 hammers, workshop and 2 windmills is 7.
    If you don't run state property but do caste, it's 1 farm for 2 workshops, same as mines (average 8/3 hammers per grassland tile instead of 4).

    So running environmentalism over state property probably makes windmills worse.
     
  12. r_rolo1

    r_rolo1 King of myself

    Joined:
    May 19, 2006
    Messages:
    13,818
    Location:
    Lisbon, Portugal
    I do prefer watermills ;) But yes, windmills + workshops under caste + SP are really strong.
     
  13. pi-r8

    pi-r8 Luddite

    Joined:
    May 1, 2006
    Messages:
    2,557
    Location:
    Babylon
    I looked up how forest regrowth works. The formula for an empty square is:
    8 * 1.25 * [number of adjacent forests]/10000 chance of regrowth each turn

    This would imply that there's no significant difference between having your forests in a checkerboard pattern, or any random arrangements. Let's say you've got 4 forests arranged in a cross around an empty square. The center tile is adjacent to 4, the corners are adjacent to 2, and the sides are adjacent to 1. the total is (basically) 4 + 8 + 4 = 16*1.25/10000 = 0.2% chance of a regrowth each turn.

    If the 4 forests are scattered separately, they are each adjacent to 4 empty tiles, which again gives 16*1.25/10000 = 0.2% chance of a regrowth.

    This implies that the only thing which really helps is to first chop down the forests which are next to tiles that can't grow forests. Chop down the coasts and the ones next to your cities, but the exact pattern doesn't matter.
    However, Even with 20 forests in ideal conditions, you'll only have a 1% chance of regrowth each turn, so you'll be waiting an average of 100 turns for each regrowth.
     
  14. Iranon

    Iranon Deity Whipping Boy

    Joined:
    Nov 14, 2007
    Messages:
    3,214
    Location:
    Germany
    True. Note that you are assuming the same number of forests/empty tiles though; a checkerboard pattern merely allows you to cram the maximum into a given (once completely forested) area. Whether this is actually relevant will of course vary, and there are lots of other considerations... you mentioned one with non-legal tiles for forests, wanting forests on squares of overlap between cities (to get the health bonus 2-4 times) could be another.
     
  15. Derakon

    Derakon Prince

    Joined:
    Aug 2, 2010
    Messages:
    387
    Assuming you're planning to leave as many forests behind as you chop, the other reason to use a checkerboard pattern is that it disrupts travel the least, since your units should never be required to actually walk into a forest.

    This, of course, becomes irrelevant once you have roads built through those forests...
     
  16. Tatran

    Tatran Deity

    Joined:
    Aug 23, 2002
    Messages:
    4,184
    Only when I've got all of the 8 mine related resources, I'll consider to chance mines to windmills.
    But in most games I always lack one or more of those resources and have to rely on luck popping up the missing reource(s).
     
  17. vicawoo

    vicawoo Chieftain

    Joined:
    Feb 12, 2007
    Messages:
    3,226
    8 * 1.25 * [number of adjacent forests]/10000 chance of regrowth each turn
    8 * 1.25 * 16/10000 = 160/10000 = 1.6/100 chance of regrowth each turn
    for 4 forests
    for 20, that's 8%
     
  18. pi-r8

    pi-r8 Luddite

    Joined:
    May 1, 2006
    Messages:
    2,557
    Location:
    Babylon
    Ah, I forgot the second factor of 8. That's much better then.

    Back on the original topic, it occurs to me that one quick way to rank tile improvements is to count up their total bonus over base. In the best case, windmills give a total of 7, which is higher than anything except towns (8). watermills give 5, workshops give 4, lumbermils give 4 + a health bonus, and farms only give 2. Of course sometimes you prefer more food or hammers over commerce, but at least it makes sense that the windmills should be good.
     
  19. karadoc

    karadoc AI programmer

    Joined:
    Oct 3, 2005
    Messages:
    1,568
    Location:
    Australia
    I guess the point of the checkerboard pattern is that it allows you to get the full regrowth power of each remaining forest, with a minimal use of land. ie. Each forest contributes some probability of regrowth for every unforested adjacent tile — so to maximize regrowth, you want to have as many forests as you can, and have each of them surrounded by unforested tiles. Any time a forest is adjacent to another forest, that's wasted potential.

    (This has already been said... why am I posting again? :( )
     
  20. Iranon

    Iranon Deity Whipping Boy

    Joined:
    Nov 14, 2007
    Messages:
    3,214
    Location:
    Germany
    @ pi-r8: We could just use the ranking the AI does - it evaluates the respective value of food, production and commerce as 5:3:2 which I think is realistic enough.
    With all techs, this would put value over base at:

    Mines, Lumbermills: 9
    Farms: 10
    workshops: 4-12
    watermills: 10-15
    windmills: 12-16
    towns: 10-17
     

Share This Page