Debunking the Arms Industry Myth

    This site uses cookies. By continuing to browse this site, you are agreeing to our Cookie Policy.

    • Rage1177 wrote:

      I was wondering has anyone ever had a game without going to war Just upgrading economy. And and simciting it out til the end of the season? That would be kinda dif and fun(ny). Anyways just sharing my experience thus far.


      I intended to do that once, as NZ, but upon seeing the overall quality of the map, I went ahead and expanded, ending the map in 4th place. However, inspired by your and @WalterChang's post, I decided to try it again. NZ, on a world war z map. Thanks!
      Overkill is an Awesome Map! :D

    • @playbabe

      I'm looking for a how-much-does-morale-change each turn formula/graph/table?

      In this earlier post in this thread: Debunking the Arms Industry Myth you showed us a nice graph that showed resource-production increases.

      The resource-production increases are the result of both population increases and morale increases.

      I'm learning about resource-production and I would like to learn how fast morale increases, and separately, how fast population increases.

      Right now I'm focused on the morale-changes part of the topic.

      I think I can maybe subtract population changes from the data in your earlier post so that the result shows only morale-change effects.

      However, if you have already done that, and especially if you have a formula for how quickly morale rises (or falls) toward a city's target-morale, I would love to get a copy of the result.

      Do you (or anyone else in the world of CoN players/teachers) have that how-much-does-morale-change formula?

      And, if you do have it, can you share it with me/us?

      The post was edited 1 time, last by KFGauss ().

    • KFGauss wrote:

      I'm looking for a how-much-does-morale-change each turn formula/graph/table?
      The morale growth only depends on the difference from its target,
      and is updated according to the following formula:

      [T-M(Day{i+1})] = [T-M(Day{i})] -C*(T-M(Day{i})) = (1-C)*[T-M(Day{i})]

      with the growth coefficient C=1/6.
      This can be expressed as an exponential function, with [T-M(Day{n})]≡D.n, as:

      D.n = D.1*(1-C)^(n-1)
      Commander Zozo001 :thumbsup:
      humble player

    • Zozo001 wrote:

      The morale growth only depends on the difference from its target,and is updated according to the following formula:

      [T-M(Day{i+1})] = [T-M(Day{i})] -C*(T-M(Day{i})) = (1-C)*[T-M(Day{i})]

      with the growth coefficient C=1/6.
      This can be expressed as an exponential function, with [T-M(Day{n})]≡D.n, as:

      D.n = D.1*(1-C)^(n-1)

      Zozo001 wrote:

      Its difference from the asymptote of 10 decays exponentially:
      Pop(t) = 10-A*exp(-kt)


      with k~=0.022 [1/day]
      Many thanks and many thanks.

      I was getting nowhere trying to puzzle out the morale changes from what I collected
      And this saves me the trouble of sifting through the several earlier population growth posts.

      I'll now try plugging my last game's home-city data into these formulae (I recorded the city morale, Civ Cas, and At-war values near the start of each game-day).

      PS: I'm now kicking myself for not recording more captured-city morale values.

      The post was edited 1 time, last by KFGauss ().

    • KFGauss wrote:

      PS: I'm now kicking myself for not recording more captured-city morale values.
      Collecting data on this is a real PITA, as the target values keep shifting. This makes accurate evaluation of parameters difficult, particularly for the population (which is more complicated due to the large role of rounding).
      Anyways, I found that may morale formula gives excellent fit. For the population, my morale effect parameter calculation is rather ill parametrized, so the overall picture is not as clear as I'd like when the morale is low (and particularly as it changes). Regardless, the growth coefficient I provided describes the curves fairly well - and it confirms the empirical finding that growth gets very slow above 8ish population (and completely stops before 10).
      Commander Zozo001 :thumbsup:
      humble player

      The post was edited 3 times, last by Zozo001 ().

    • Zozo001 wrote:

      KFGauss wrote:

      I'm looking for a how-much-does-morale-change each turn formula/graph/table?
      The morale growth only depends on the difference from its target,and is updated according to the following formula:

      [T-M(Day{i+1})] = [T-M(Day{i})] -C*(T-M(Day{i})) = (1-C)*[T-M(Day{i})]

      with the growth coefficient C=1/6.
      This can be expressed as an exponential function, with [T-M(Day{n})]≡D.n, as:

      D.n = D.1*(1-C)^(n-1)
      I can't understand anything,lol

    • Roomi Royale wrote:

      Zozo001 wrote:

      The morale growth only depends on the difference from its target,and is updated according to the following formula:
      [T-M(Day{i+1})] = [T-M(Day{i})] -C*(T-M(Day{i})) = (1-C)*[T-M(Day{i})]

      with the growth coefficient C=1/6.
      This can be expressed as an exponential function, with [T-M(Day{n})]≡D.n, as:

      D.n = D.1*(1-C)^(n-1)
      I can't understand anything,lol
      My apology for the terseness. M(Day{i}) is the morale on the i-th day; T is the target morale (i.e. 90% for a homeland city when there are no maluses, such as ongoing wars; note that malus and bonus contributions are listed on the city's popup info as "Morale factors", but the target itself is not shown); C is a constant (a parameter for the formula); and D.n is the difference between the morale and its target, on the n-th day.
      Commander Zozo001 :thumbsup:
      humble player

    • playbabe wrote:

      Pop 9->10 take 6,840,000 second
      As I had noted before, your model seems to be off at the high end, overestimating the growth rate by a lot when approaching the asymptote at 10. My fitted parameters (I only included >=4 Pop) agree well with yours up to 9, but then going from 8.95 to 9.95 (i.e. when 10.0 is displayed) would take 11,963,337 seconds (236 days - 100 of which would be spent on going from 9.55 to 9.95) - almost twice as long as you claim.

      Note that this can be tested empirically: e.g. my model's growth rate is 0.022/day at 9 (and 0.011/day at 9.5), yours would be much faster. All you need is observing Pop 9 cities for a month or so, to verify which model fits well.
      Commander Zozo001 :thumbsup:
      humble player

      The post was edited 2 times, last by Zozo001 ().

    • Our challenge is using an algorithm to match these home-city numbers I recorded shortly after each new game-day began. Wish me luck.

      DayMoraleCiv CasAt WarBonusesTot +/-
      70
      1700000
      2720-60-6
      373-1-80-9
      474-2-60-8
      575-3-60-9
      676-2-20-4
      777-4-40-8
      878-3-20-5
      979-2-60-8
      1080-2-40-6
      1181-1-60-7
      1279-3-60-9
      1380-2-80-10
      1481-2-80-10
      1579-2-80-10
      1680-2-60-8
      1779-3-60-9
      1876-5-100-15

    • @Zozo001 I need to confirm I'm using your formula correctly.

      Using this rearrangement of your formula:
      • NextTurnMorale = AdjustedTarget - (1 - C) * (AdjustedTarget - ThisTurnMorale)
      With
      • AdjustedTarget = UnadjustedTarget + (CivCas + AtWar + Bonuses)
      Using these constants
      • Unadjusted Target = 90
      • C = 1/6
      • (1 - C) = 0.83333


      And using these values from my table (previous post)
      • Day 5 Morale = 75
      • Start of Day 6 CivCas = -2 = End of Day 5 CivCas
      • Start of Day 6 AtWar = -2 = End of Day 5 AtWar
      I get
      • Day 6 Morale =
      • (90 -4) - (0.83333 * ((90-4) - 75) =
      • 86 - (0.83333 * (86 - 75) =
      • 86 - (0.83333 * 11) =
      • 86 - 9.1667 =
      • 76.8333 =
      • 77
      That's not what happened in the game. The actual Day 6 morale was 76.

      Also, using the method I just outlined, the calculated values don't track the actual values after Day 11 when actual values stop rising.

      Am I misinterpreting your formula?

      If what you supplied is just a curve fit, that will explain how it's different from what I'm looking for. I want to be able to look at a city at 23:59 of Day X and be able to predict what that city's morale will be at 00:01 of Day X+1. A curve fit will be good for the overall trends, but not so good for the twists and turns of each individual day's changes.

      FYI: Here's a plot of my actual values recorded during the game.
      PacificMorale.JPG

      The post was edited 1 time, last by KFGauss ().

    • KFGauss wrote:

      @Zozo001 I need to confirm I'm using your formula correctly.
      You are, and it does work. See my plot of tracking your data.

      _ KFG model.jpg

      The agreement is as good as could be expected.
      What you must understand, first of all, is that the numbers displayed by CON are rounded values of floating point quantities. I.e. when you read 76% morale, it is actually something between 75.5% and 76.5%; and the same goes for the malus values. This can easily explain a 1 unit difference (or occasionally 2), between the observed and calculated results. (In passing I mention that there might be a small random component which I tentatively found in rare cases - it is not clear whether this is a true RNG, or just a quirk in how CON internal calculations handle rounding; but this is only observable infrequently.)
      Files

      • _ KFG model.jpg

        (38.26 kB, downloaded 2 times, last: )
      Commander Zozo001 :thumbsup:
      humble player

    • playbabe wrote:

      well good news to you then, in meta data it kinda pointed at that morale is not rounded number. maybe it was rounded before updating stats to city but so far I can only see morale and building HP do this.

      the opposite is true, Unit’s HP & Damage have stupidly high digits.
      Well population is pretty clearly a higher precision floating number than displayed (at single decimal rounding), too. This has the practical effect of modifying resource production (which is observable with much higher precision), too.

      Morale not being rounded behind the scene is an obvious observation, when looking at displayed values that oscillate around their nominally reached target value.

      That displayed building HP is a rounded value of an underlying float is also clear from observing, e.g., how the functionality of airports activates when crossing the 40% readiness threshold (which only happens with a time lag after 4/10 HP appears).
      Commander Zozo001 :thumbsup:
      humble player

      The post was edited 1 time, last by Zozo001 ().