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

On Tech Diffusion and the Steamroll effect

Discussion in 'Civ4 - Caveman 2 Cosmos' started by ls612, Oct 25, 2012.

  1. wolfensoul9

    wolfensoul9 Warlord

    Joined:
    May 9, 2012
    Messages:
    255
    I agree with Necratoid. Can't make heads or tails of it.

    For those of us with little to no knowledge of computer programming, would someone mind translating?
     
  2. Thunderbrd

    Thunderbrd C2C War Dog

    Joined:
    Jan 2, 2010
    Messages:
    27,812
    Gender:
    Male
    Location:
    Las Vegas
    TBH, I'd like to look into the exact formula myself. I'm not entirely sure ls's adjustments are going to move in the direction he's thinking they will.
     
  3. pi4t

    pi4t Warlord

    Joined:
    Aug 2, 2009
    Messages:
    271
    I'm a long time lurker here, and so sorry for barging in with a post which probably is incorrect/makes no sense :p

    But I'm seeing Necratoid make an argument that 'to prevent the positive feedback effect, one should just increase the intelligence of the AI until it is at a level with the player' again and again, and not being given any form of response. The problem with the Civilisation games, though, as has been said on this thread, is that when one player gets ahead any distance they tend to stay ahead. This applies to both the player and the AI (which requires a lot of skill to catch up with and surpass on higher difficulties, becoming nigh on impossible if it's improved to the skill of a player). The important reason that improving the AI simply would not work, though is this: there are almost infinite numbers of variables existing, both in game and out of game.

    For instance, assume that Koshling programmed a perfectly balanced AI that would, with specific starting conditions, give him a challenge through the entire game without steamrolling him. Great! Now if he gives it to someone else, with a little less skill in the game (me, for instance), they'll probably find that the AI begins to get an edge on them. As the tech difference increases, and the AI begins to conquer the world, it rapidly becomes clear that they will win and the game is not worth continuing. Conversely, if the same scenario is given to a player with slightly more skill, they can get a small lead and increase it. Increasing a difficulty level, though, would likely result in a loss as the AI gains power.

    Worse still, if Koshling regenerated the map, and played on a poorer starting position, he'd find himself unable to keep up (and vice versa for better position). Even random factors (does the AI declare war when its demand is refused?) during the game could swing the balance of power and cause one civ to begin to steamroll the others.

    This issue is more prevalent in C2C than vanilla cIV, I'm guessing, simply because of the longer games. The point is that while a better AI will decrease the likelihood of the player steamrolling the AIs, but not reduce the overall chance of a steamroll significantly. This is the problem with a vicious circle, which needs some external factor to prevent it from getting out of control.

    As for what you'd get for being further ahead in research, you have all the benefits that being ahead in tech brings: improved infrastructure, the ability to crush spearmen against tanks, etc. Just as normally a research will be nigh on useless to you if you don't do anything with it, so in this circumstance the increased power will allow you to do things you otherwise couldn't: finish conquering the continent, get a head start on building factories, make a colony on the moon first, etc. Just bear in mind that this (metaphorical) 'golden age' where you're a massive way ahead of your opponents will not last forever. If I've understood the system, though, they'll never overtake you, or end up in a better position relative to you than had you researched less well through this system. So you are still gaining advantages from being ahead, but the game rewards you more for what you're doing recently than what you did 4000 years ago: as should be the case, really.

    I'll go back to lurking now :)
     
  4. Thunderbrd

    Thunderbrd C2C War Dog

    Joined:
    Jan 2, 2010
    Messages:
    27,812
    Gender:
    Male
    Location:
    Las Vegas
    WELL put and welcome to C2C! You can stop lurking now ;)
     
  5. Hydromancerx

    Hydromancerx C2C Modder

    Joined:
    Feb 27, 2008
    Messages:
    16,281
    Location:
    California, USA
    @pi4t

    You bring up some good points however I would like to remind you that resources can also play a factor. In one of my most challenging games the AI was ahead of me and I was lacking sulphur resource. Which ment no Gunpowerder units for me. My best units I could make for a long time were Balista Elephants which kept me alive until I could get steampunk. Once I got those Steamtanks online I was able to push their army back and reclaim cities that they took from me plus some of theirs. It was one of the most memorable games only because I lacked a key resource in the game and just could not upgrade my units.
     
  6. Faustmouse

    Faustmouse Deity

    Joined:
    Jan 31, 2012
    Messages:
    3,522
    Well, the lack of sulphur thing is very interesting...
    It's kind of a radical idea, but what about making Sulphur not be placed by the mapgenerator at all. Instead it would be only pop up (even if no one could see it yet) on earlier vulcano-plots. And than make vulcano-plots more likely near civs that are behind so they have Sulphur more likely than the advanced civs.

    However, this would be disliked by many players... I know that the argument "just make the AI more intelligent" isn't easy to achieve. I think a good way to start with would be to increase the AIs intelligence regarding the citiy-limit and even more important revolutions. These are already mechanics to "weaken" the leader, IF the AI won't struggle with them so much. (If it still does, it's been a while since my last game)


    And it would be REALLY nice if someone could "translate" the code so that non-coders could understand it :goodjob:
     
  7. Hydromancerx

    Hydromancerx C2C Modder

    Joined:
    Feb 27, 2008
    Messages:
    16,281
    Location:
    California, USA
    That epic game was before the volcano events were put in place. Sulphur and Obsidian are much more common now.
     
  8. wolfensoul9

    wolfensoul9 Warlord

    Joined:
    May 9, 2012
    Messages:
    255
    I really think the lot of you are exaggerating the strength of the positive feedback that comes from being technologically ahead.

    Even on deity on regular bts, the threshold for "nope, they're too far ahead, it's all over now", commonly known as "running away with the game" is farther ahead than what I think you all are making it out to be. And that's even with the massive handicaps the AI gets on deity!

    Usually, the only time where the AI will be unstoppable, and truly be in that untouchable positive feedback loop, even on deity, is if they are unreachable through distance. This is why alot of deity players (I myself am not one, but I have watched many LP's of them) fear more than anything a civ on a different continent, one that they cannot do anything to.

    The keyword here is "cannot". So long as a player has contact with, and their military can reach a far ahead civ, there's ALWAYS hope. It is that capability to fight back against those odds which marks the proper deity player.

    I realize I cannot directly prove that the positive feedback is not as insurmountable as you all are saying it is...which is frustrating...

    And you know what? Even if it was insurmountable, I'd be fine with that. Because to have reached that point means you have played well. The game is yours, revel in it. You've earned it.

    Look, the fact is, is that once you are far ahead enough, no single civ will be able to stop you. Perhaps if the others join forces, maybe, but even still, that is simply the nature of the game.

    I really feel like trying to "always make sure the weaker player has a chance, no matter how far ahead the strongest civ is" mentality will just lead to disaster.

    There comes a point when you just have to have the weak roll over and yield. Giving a smorgasborg of bonuses to those behind, just for being behind, is unfair, and simply wrong in the principle of good game design.

    There's always a chance to come back. It rests in you playing well, or even extremely well. Same goes for the AI.
     
  9. Faustmouse

    Faustmouse Deity

    Joined:
    Jan 31, 2012
    Messages:
    3,522
    Well, if that is what makes you happy, that simply don't use Tech Diffusiun. It is a game OPTION.

    Most of the people here would love to play a game where there is a challenge even in the modern era. Tech Diffusion may not be the most elegant way to achieve this, but it's ok. I don't think that if you are the techleader and the Number 2 is only ONE Tech behind you it will get much bonus (if any...) for researching it.
    But give far behind civs a huge bonus would be ok. If you have Steampower, what is the use of 1-city-civs in the classic era? You could just go in the WB and delete them, it won't make any difference.
     
  10. wolfensoul9

    wolfensoul9 Warlord

    Joined:
    May 9, 2012
    Messages:
    255
    It is the principle of the matter. Even having it as an option is wrong, in my humble opinion. I don't care if you're a one city state in salt flats. Giving bonuses to those that are behind BECAUSE THEY ARE BEHIND is unfair. It's wrong. It violates good sportsmanship. So to me, no, it is very much not okay. Even as an option.

    And yes, I do realize there are other factors in tech diffusion that the level of technological advancement, or lack thereof. Those other factors are fine, even good, in my opinion.
     
  11. AIAndy

    AIAndy Deity

    Joined:
    Jun 8, 2011
    Messages:
    3,414
    @pi4t: Very well explained.
     
  12. wolfensoul9

    wolfensoul9 Warlord

    Joined:
    May 9, 2012
    Messages:
    255
    This is not an issue. This is correct. This is good. This is proper. This is fair.

    How can you complain that having less skill, and ergo, being beaten is unfair or a problem that needs to be addressed? If you get dealt a bad hand, deal with it. Play all the better to win, or get left in the dust. This goes for the AI and player.
     
  13. 45°38'N-13°47'E

    45°38'N-13°47'E Deity

    Joined:
    Jun 7, 2008
    Messages:
    5,969
    Location:
    Just wonder...
    At the moment, I'm also trying to fix tech diffusion for civs falling behind in RoM-A New Dawn 2; but tweaking with xml had little effect. So I've found another way and it looks like it's working good enough; in techdiffusion.py; I've modified the code from

    # Extra boost if really far back
    if( teamTechCount <= curMaxTechCount - self.bonusTechsBehind ) :
    if( self.LOG_DEBUG ) : CvUtil.pyPrint( " TD : Team is way behind")
    researchPercent = researchPercent*pow((curMaxTechCount-teamTechCount-self.bonusTechsBehind + 2)/2.0, 0.5 )

    to

    # Extra boost if really far back
    if( teamTechCount <= curMaxTechCount - self.bonusTechsBehind ) :
    if( self.LOG_DEBUG ) : CvUtil.pyPrint( " TD : Team is way behind")
    researchPercent = researchPercent*pow((curMaxTechCount-teamTechCount-self.bonusTechsBehind + 2)/2.0, 0.8 )

    It seems sufficient to help civs who are really far behind (and I've set for "far behind", civs that are >6 techs behind), except those civs that only have 2-3 cities, which will always stay behind.
     
  14. Koshling

    Koshling Vorlon

    Joined:
    Apr 11, 2011
    Messages:
    9,254
    I have to admit, I am sympathetic to the point Wolfensoul makes about the just-because-they-are-behind aspect of the mechanic.

    Personally I would be in favor of increasing the negative feedback effect for those mechanics that are not 'just charity' (reduced cost for techs known by other civs), and removing (or making a SEPARATE option out of) the 'just charity' part that lowers your costs simply because you are behind.

    To INCREASE the negate feedback effect from the other mechanics I would do something along the lines of the following:
    • Cost decrease which increases for each civ you are in contact with that owns the tech (not linearly, but noticeably for at least the first 4 or 5)
    • Modifier on the cost decrease based on trading status (trade route connectivity, open borders increase the effect, free trade agreement even more so)
    • Modifier on the cost decrease based on your espionage ratio with the other civ concerned
    • Modifier based on geographic closeness (at least same-landmass vs across ocean)

    We can then tune those (rationalisable, and also somewhat controllable) modifiers up to whatever degree we need to get the appropriate level of negative feedback.
     
  15. wolfensoul9

    wolfensoul9 Warlord

    Joined:
    May 9, 2012
    Messages:
    255
    Excellent! All of those bulleted ideas are grade A awesome in my opinion. I would absolutely be in favor of all of them. I hope others agree as well.

    Like I said, the only thing I dislike is the "just-because-they-are-behind" part of it. The whole "inability to keep technological spread contained" is far different, realistic, AND will help balance. So, so much better.
     
  16. crazyewok

    crazyewok Warlord

    Joined:
    Sep 12, 2012
    Messages:
    202
    Im pretty mich in favor of tec diffusion. Its pretty realistic too as tec does diffuse. And the more civ that have the tec the more likley it is to diffuse. And if i understand corrcetly its not meaning as soon as research something its gunna instantly go to every other civ. It just means if your ahead and "the" super power that instead of being 20 tec ahead you will just be 5 or 6 ahead.

    But as already said IT A OPTION! if you dont like it turn it off and quit complaining :lol:
     
  17. Koshling

    Koshling Vorlon

    Joined:
    Apr 11, 2011
    Messages:
    9,254
    The problem is that it is really two options wrapped up in one game option. The two options being mechanics for increasing tech spread based on degree of exposure (known trade partners with the tech etc.), and mechanics for making life easier because someone has a tech you don't even if you don't even know them. Wolfen (and I too for that matter) want the first but not the second.
     
  18. 45°38'N-13°47'E

    45°38'N-13°47'E Deity

    Joined:
    Jun 7, 2008
    Messages:
    5,969
    Location:
    Just wonder...
    Wait, but doesn't Tech diffusion already works like this? In some test that I've made, I've seen that increasing tech diffusion the way I've done seemed to work only if you had open borders with civs ahead of you (and vice versa). If a civ is living alone, never using open borders and so on, it looks like it's still falling behind. Isn't this what we all want? And isn't Tech diffusion already designed like this? Maybe there's something I'm not aware of...
     
  19. AIAndy

    AIAndy Deity

    Joined:
    Jun 8, 2011
    Messages:
    3,414
    Yes, it is implemented more or less that way. Currently imo there is a too weak effect for being the second Civ on a tech which allows one Civ to take the headstart.
     
  20. Thunderbrd

    Thunderbrd C2C War Dog

    Joined:
    Jan 2, 2010
    Messages:
    27,812
    Gender:
    Male
    Location:
    Las Vegas
    Here's where Wolfy, you and I are going to all agree. I love these proposed adjustments to Tech Diffusion. To me, these would make the system more rational and rooted in logical cause and effect. Surprisingly, I also agree that if Civ A doesn't know Civ B and Civ B has a tech Civ A doesn't have yet, then Civ A should get no benefit towards earning that Tech... there's no cause for it since there's no exposure there. The system should very much be more based on exposure to the civs that have the techs you don't.

    I have another one to add to the bullet list though: Add a touch of research towards a military tech you don't have when you defeat a unit that had that tech as a prerequisite. (Assumes you take some knowledge from the encounter and any leftovers from the battle.)
     

Share This Page