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

Stabs's heuristic exchange rate mechanism (SHERM).

Discussion in 'Civ6 - Strategy & Tips' started by Callduron, May 2, 2017.

  1. Temppu

    Temppu Warlord

    Joined:
    Oct 16, 2016
    Messages:
    160
    Gender:
    Male
    I think it is safe to say that this thread has well stated that the "sherm" heuristic or is too simplistic, misleading, and even flawed. Heuristics may help in situations, where you have to make quick and simple decisions. But in complex and less hasty games such as Civ, they encourage you to stop thinking the overall situation in your game.

    Because it has already been covered that the heuristics as such work badly, I would like to comment on the decision hierarchy OP has presented.

    This hierarchy gives a somewhat distorted picture of the way decision are made in Civ. In the higher part of this hierarchy, I would point out that only few players go as far as actually evaluating numerically the potential decisions. What skilled players mostly do, is to consider their overall strategy, and then tailor their decisions to fit that strategy.

    In the lower part of the hierarchy, the gut impulse should, at least after some experience in the game, give you a hunch of your overall strategy. Hence, the gut feeling should direct you to a consistent gameplay, even though you may not always be aware of this. Then, and only then, when you don't really have a hunch of any sort, you should opt to simple heuristics that don't consider the strategy you are pursuing.

    With these arguments, I would propose the following modified decision hierarchy.

    Top 1: Quantitative evaluation that reflects your overall strategy and game situation
    Top 2: Qualitative evaluation that reflects your overall strategy and game situation
    Bottom 1: Gut feeling selection
    Bottom 2: Heuristic selection
     
    CountAccountant likes this.
  2. Victoria

    Victoria Regina Supporter

    Joined:
    Apr 11, 2011
    Messages:
    11,310
  3. Arent11

    Arent11 King

    Joined:
    Nov 18, 2016
    Messages:
    988
    I disagree :) I think the old "1 production ~ 2 culture/science/gold" conversion works reasonably well.
     
  4. Victoria

    Victoria Regina Supporter

    Joined:
    Apr 11, 2011
    Messages:
    11,310
    while gold may not truly equally 2 production it has added flexibility that pushed it toward 2.

    However on science and culture they are just waaaay to flexible to be called a CHEFTAIN let alone equate to 1/2 a production.
    In a CV before turn 100 science is of no value to me while culture perhaps equals production.
    Turn 100-180 I just cannot equate it, maybe 1 culture equals 10 prod, no it just does not equate Because you cannot substitute it. well perhaps to your detriment as the competition does hot up to some degree. It's a question of speed vs pleasure, a personal choice.
    Post turn 180 culture may equal prod again but science becomes huge and all encompassing bar defence, I even use specialists which I may not have done

    For a science culture victory it's simpler, science all the way
     
  5. MadDjinn

    MadDjinn Deity

    Joined:
    Jan 28, 2011
    Messages:
    4,554
    Civvi has actually flipped the script on gold vs production to the tune of 1 gold = 2+ production. At least, that's where my heuristical values are shifting.

    Production is great early game, but slides down to less useful later game (unless going for a tech vc, even then, its all about the great people). As soon as you can start running high gold output, you can buy things faster than produce them. The only thing you still need decent production for is the district and any wonders. Gold will both get the rest as well as allow for purchases.

    That said, the op is clearly wrong about 'no faith without religion'. I can be bad about adding faith, but you still want it for GP rushing (better conversion than gold) as well as for any number of CSs or other such things that allow for faith purchase of common items.

    Not to mention ensuring you get the 'best for the current game choice' religion that someone else founded.
     
  6. Leyrann

    Leyrann Deity

    Joined:
    Jan 11, 2015
    Messages:
    4,016
    Gender:
    Male
    Location:
    Netherlands
    Don't things you buy cost you three or four times as much gold as they'd cost you production? Sounds like 3 or 4 gold = 1 production then.
     
  7. MadDjinn

    MadDjinn Deity

    Joined:
    Jan 28, 2011
    Messages:
    4,554
    except the game also gives you more gold. And the game is awash in ways to get gold, so it's actually easier to mass gold than it is to mass production.
     
  8. Browd

    Browd Dilettante Administrator

    Joined:
    Aug 5, 2012
    Messages:
    11,797
    Gender:
    Male
    Location:
    Rural Vermont
    And gold can be amassed empire wide and spent in any city -- not so much with production.
     
  9. MyOtherName

    MyOtherName Emperor

    Joined:
    Dec 7, 2004
    Messages:
    1,526
    That doesn't change the fact the base conversion rate is 4 gold to 1 cog. (2 gold to 1 cog for upgrades)

    (Better if you have bonuses to buying/upgrading. Worse if you have production bonuses)

    Setting 1 cog = 4 gold doesn't mean "ignore gold in favor of cogs", it means "value 4 gold the same as 1 cog".
     
    Leyrann likes this.
  10. MadDjinn

    MadDjinn Deity

    Joined:
    Jan 28, 2011
    Messages:
    4,554
    but it's not really. CSs give gold at a 2:1 rate vs any other yield, same for specialists. Trade routes alone (along with requisite policies) give far more gold vs production. When you tie the trade route income into commerce district output, you find that they also give far more gold than an IZ does wrt production.

    Once you start rolling into Exploration and then Democracy, gold becomes more valuable than straight production for most items.

    So while a few items might have a 4x cost in gold, the game more than makes up for it by providing that gold in the first place.
     
  11. TheMeInTeam

    TheMeInTeam Top Logic

    Joined:
    Jan 26, 2008
    Messages:
    25,713
    It's not a fair value translation because cogs are less transferrable. You can't target-focus cogs to rush buy units out of a border city every turn or transfer them from capital to satellite city to get infrastructure up there (aside shifting trade routes).

    As such you'd need to rate the value of gold relative to cogs somewhere higher than 4 gold = 1 cog, likely substantially so if pursuing VC where you don't have major production bottlenecks.
     
  12. MyOtherName

    MyOtherName Emperor

    Joined:
    Dec 7, 2004
    Messages:
    1,526
    Great. That has nothing to do with how much you value gold relative to cogs.

    Great. That has nothing to do with how much you value gold relative to cogs.

    Great. That has nothing to do with how much you value gold relative to cogs.

    Great. That has nothing to do with how much you value gold relative to cogs.

    You're completely missing the point here. Picking a heuristic like N gold = 1 cog has absolutely nothing to do with the availability of either resource: it's an expression about how useful each resource is to you. It means that if you're given the choice between getting 1 cog or less than N gold, you'd prefer the cog-giving option, and conversely if you're given the choice between getting 1 cog or more than N gold, you'd prefer the gold-giving option.

    It was a hypothetical, so that I could make a concrete statement. It was not meant as an assertion of what you should do.
     
  13. MadDjinn

    MadDjinn Deity

    Joined:
    Jan 28, 2011
    Messages:
    4,554
    and you seemed to miss mine. I said that my heuristics are shifting towards 1g = 2 cogs.

    So either you want to discuss heuristial values in a given system, and then can discuss with me where that heuristic came from, or you want to make definitive statements that your 4g = 1 cog is accurate for everything, which it's not, and everything I posted has everything to do with it.
     
  14. MyOtherName

    MyOtherName Emperor

    Joined:
    Dec 7, 2004
    Messages:
    1,526
    I missed that, because you kept harping on the availability of gold relative to cogs, and said absolutely nothing about the actual worth of gold relative to cogs.

    I don't want to make definitive statements like that. And I didn't make a definitive statement like that. I only mentioned that setting as a hypothetical example.

    Everything you posted was about the availability of gold, not the value of gold.
     
  15. MyOtherName

    MyOtherName Emperor

    Joined:
    Dec 7, 2004
    Messages:
    1,526
    Ah, I think I see what's going on; you've adjusted your own values to match availability; so that you value 1 cog = 2 gold more or less simply because you often see gold sources giving twice as much as cog sources. So, for your own values talking about availability is talking about the value.

    And maybe you haven't grokked that there are other ways of going about it, such as valuing gold for its purchasing power, or tweaking the value depending on how much you want to pay for flexibility, or tweaking the value to adjust how much gold you tend to have available for purposes unrelated to buying buildings/upgrading units.
     
  16. Victoria

    Victoria Regina Supporter

    Joined:
    Apr 11, 2011
    Messages:
    11,310
    Before this post and your explanation @MadDjinn I was in @Browd camp. Now I think the answer is not so simp,e either way. It just depends so much on situation and victory.
     
  17. isau

    isau Deity

    Joined:
    Jan 15, 2007
    Messages:
    3,068
    I appreciate the argument, but I think the calculation is a bit more tricky than this.

    The main thing is Gold (and to some extent Faith) is global where Production is almost always local. Huge Production is helpful, surely, but to get buildings up quickly in a new city, it is Gold that is most useful. It is true that you need Production for the Districts built in the new city as well, so Production is useful toward that.

    Gold also allows you to put a unit in a specific spot more or less instantly. That by itself makes it extremely useful in a way that Production isn't. I would generally prefer to sap an enemy's Gold supply instead of their Production. It's Gold that lets them pop obstacles into my way instantly. It's also Gold that allows a fairly new city to immediately pop a Builder or Settler.

    It is generally true that you can get Gold without sacrificing too much on tiles. So on that point I somewhat agree, in general its better to work Production tiles. But I dont adhere to a strict 1 Production:3 Gold relationship. They're both really important.
     
    Victoria likes this.
  18. Victoria

    Victoria Regina Supporter

    Joined:
    Apr 11, 2011
    Messages:
    11,310
    Hmmm I just wrote this in another post but think it sums it up nicely in my head

    ####

    To me production is king but gold is god. They both have their place...

    Production is what maintains our cities, speeds them up and truly gives our advantage over the AI at higher levels (and combat ofc). It also is required for building districts.

    Gold is more globally usable, you can buy in a city with low production and you can buy things like Great people. Once you have built the district you can buy the buildings in there... You can even buy the builder to chop the wood to build the district to buy the buildings.

    This makes gold more godlike but it is a limited resource from a global pool and production just keeps churning away in each city. Gold is the magic that answers the prayers when production cannot.

    I am not sure you can equate one to the other value wise. they have their places in different parts of the pyramid.

    I guess ... in a production rich game gold is less godlike and in a production starved game gold is more deified. Therefore their value is not x gold = y production but instead are on a sliding scale based on terrain, like so many other things in the game seem reliant on.
     
  19. isau

    isau Deity

    Joined:
    Jan 15, 2007
    Messages:
    3,068
    I forgot to mention upgrading units. The way the game is designed right now, its generally best to build an early army and just keep upgrading it, rather than rebuilding it. That can only be done with Gold.

    I will admit I would prefer a high Production start to a high Gold start. That's mainly because there are ways to acquire Gold besides working tiles. But having a lot of Gold, especially early in the game, definitely has its perks.

    Faith is interesting. The difference between having 1 Faith and 0 Faith in the early game is huge. But having 2 Faith isn't much better than having 1 (generally).
     
  20. MadDjinn

    MadDjinn Deity

    Joined:
    Jan 28, 2011
    Messages:
    4,554
    No, I'm not using supply side heuristics. Nor am I using a cost side heuristic like you are, ie, what can I do with it.
    I'm using a holistic approach. That is, how much can I get AND what can I spend it on.

    Easy example-

    Cog to gold spending wise is 1:4, or less depending on what it is specifically. But, gold to cog income is 2:1. Which means when you factor in both income and expenses, you get 1:2.

    So while you spend more gold that production on some things where those overlap, you also get twice or more gold per cog on the income side. A cost side equation only holds up when the income is equivalent, and in the case of gold vs cogs, it's not. It also fails when taking into consideration what can be purchased with each. Some things gold can buy are more valuable than others. Gold can also buy things that production can't.

    The only time gold has a zero relative value vs production is on base district and wonders. Even on the production side, the value of adding one more cog to a pile of cogs is a diminishing return equation when compared to most anything else because it is local only and has exactly 2 items that can't be done by something else.


    @Victoria - yes, the equations are not as simple as what has been discussed in the thread so far. You really need a model that takes everything into consideration for the ability to set up actual values in a specific instance because they change over time.
     

Share This Page