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

AI culture imbalance

Discussion in 'General Balance' started by ElliotS, Feb 4, 2018.

  1. ElliotS

    ElliotS Warmonger

    Joined:
    Jun 13, 2013
    Messages:
    2,704
    Gender:
    Male
    Location:
    Chicago
    Many high-level players have reported that the AI gets an insane amount of culture compared to everything else. This leads to them getting large policy leads and out of control.

    One reason that has been pointed out is that the AI can build everything and focus culture in many cities, because of their high production bonuses.

    Another is that unless it changed and I didn't notice, the AI gets equal amounts of all yields in it's handicap.

    I think we can all agree that gold and culture are not equally valuable.

    So the options to deal with the problem include:
    • Reduce culture from handicaps
    • Remove culture from handicaps
    • Reduce the amount of culture AI gets from culture focus
    Though I'm open to other solutions as well.

    As a note: We do not need to decrease difficulty. If fixing this issue makes games too easy, we can easily increase all bonuses slightly.
     
  2. Txurce

    Txurce Deity

    Joined:
    Jan 4, 2002
    Messages:
    8,263
    Location:
    Venice, California
    ElliottS, the only alternative I see is more catch-up mechanisms like (for example) the Travel Ban WC resolution. Is that one available early enough? Regardless, more would be needed, and it's hard to argue for additional mechanics when it would be much cleaner to reduce the culture handicap a bit, and take it from there. I lean toward this approach, which keeps the AI working as optimally as it does now. But I would suggest examining all the handicaps while we're at it, and how they balance against each other. As you said, gold doesn't equal culture. What about the other values? Doing this may help us reach the end goal more quickly.
     
  3. LukaSlovenia29

    LukaSlovenia29 Emperor

    Joined:
    Mar 13, 2016
    Messages:
    1,500
    If possible to code, I'd be in favour of option number one, reducing culture from handicaps. Like Elliot said, not all yields are the same.
     
    Grabbl likes this.
  4. Stalker0

    Stalker0 Baller Magnus

    Joined:
    Dec 31, 2005
    Messages:
    6,249
    I'm in favor of option 1. After all, AI handicap adjustment has been a major part of testing these last 2 patches so if culture sticks out as the sore thumb no reason its can't be adjusted a bit.

    I think I'd also be in favor of one more slight adjustment to ABC values for a little more mid game power and a slight reduction in late game scaling. Again, slight!
     
  5. Gothic_Empire

    Gothic_Empire AKA, Ramen Empire

    Joined:
    Aug 5, 2008
    Messages:
    1,493
    I haven't played the new patch because i've got a hotseat game with a friend going. I'm Poland, he's Japan, and we've got 8 AIs on the map @ King Difficulty.

    He's been producing a LOT of culture as Japan. I've gotten two of my free policies. We're both about even with the AIs' policies. As Poland, this disappoints me. :p
     
    IcyAngel likes this.
  6. Deadstarre

    Deadstarre Expert

    Joined:
    Nov 1, 2015
    Messages:
    960
    Location:
    New York
    On deity with Spain (I am now certain Ethiopia is obsolete) I started with culture luxury, took culture pantheon, took inspiration follower and focused culture buildings ASAP when possible to synergize with authority. Despite all this, I was behind AI by about 2 policies throughout most of the game. I can imagine that for anyone trying to play any different way it would be basically impossible to keep up on policies

    important note: it wasnt enough for them to win. I did eventually catch them on policies sometime after industrial, i'm on turn 280ish now and im certain my victory is imminent but im playing it through anyway.
     
    IcyAngel likes this.
  7. tu_79

    tu_79 Deity

    Joined:
    Feb 11, 2016
    Messages:
    6,938
    Location:
    Malaga (Spain)
    I think the early values should be lower. But that's me. I think you all are used to be in disadvantage in higher difficulties, and being able to settle as if you were two difficults lower is too much too accept.
     
  8. bigcat88

    bigcat88 King

    Joined:
    Nov 26, 2016
    Messages:
    718
    Gender:
    Male
    Like the first option.
    Don't like third option(it is a crutch!)
     
  9. Stalker0

    Stalker0 Baller Magnus

    Joined:
    Dec 31, 2005
    Messages:
    6,249
    I don't understand what you mean here. You said used to being in disadvantage, but then immediately said we are settling as i playing much lower difficulty. Not quite sure what you are saying.
     
  10. tu_79

    tu_79 Deity

    Joined:
    Feb 11, 2016
    Messages:
    6,938
    Location:
    Malaga (Spain)
    Let me put it this way. When I change ABC values so that the early game has something like 2.5 - 3 (x difficulty value), it is the right amount of difficulty for me. I feel like the AI progress as fast as I do. But when other players try this value or even higher ones, they say it's too easy to settle and claim territory, and they think they have already won. But I think we are judging with the old model in mind. We are used to be in disadvantage, so when the AI has the same strength as a player of our same level, we think it is too easy, but it's not. We only have an easier early game. There are plenty of mechanics to prevent any player to win the game just in the early game (exception in duel maps).

    You think mid game feels now easy, in comparison with the early game. It may be. And if you made a good beginning (lucky start, early bloomer civ), it might be easier than what it should. The right values are very difficult to realise now, because AI has been improved, and as this thread suggests, even handicap bonuses may not be giving the right proportions of yields (difficulty might be fine, but all AI being better at culture than any other thing does not feel right, if this is true). But I'm not talking about culture vs other handicap yields.

    I guess a good way to test handicap values would be to use two or three games, using Carthage (all bonuses in the early game), Poland (or other civ with bonuses spreaded among eras) and France (no noticeable bonuses until Medieval). This could give a better insight on how well tuned handicap values are, at least for early game.
     
  11. IcyAngel

    IcyAngel Prince

    Joined:
    Jan 15, 2016
    Messages:
    534
    By the way, the AI switching to culture production can't be it because they get a lead very early on (before production-to-culture) and actually tend to lose it very slowly over time; as deadstarre noted, in late eras the lead appears to fall off a bit (but you'll only catch up if you focused on culture like mad...).
     
    ElliotS likes this.
  12. Grabbl

    Grabbl Prince

    Joined:
    May 25, 2016
    Messages:
    429
    I agree and think that it is because of better prioritizing together with the handicap bonuses (no evidence, just a guess). In late game, you can catch up when focussing on culture because at one point, also the player will have built all culture buildings (or sufficiently of them) and "maxed" out culture production. In my observation, it is as you described, AI get a fast lead of a few policies early on which then stays more or less constant (when the player focusses on culture) and decreases a bit in the late game.

    Consequently, I would favor the first proposition of lowering culture yield from handicap and see how this works out.
     
  13. Deadstarre

    Deadstarre Expert

    Joined:
    Nov 1, 2015
    Messages:
    960
    Location:
    New York
    I wonder how many people have realized yet that Temples became core culture buildings now. If you dont tech that direction or if you delay on grand temple / temple building even a little bit, youll be in the culture dust very quickly
     
  14. Txurce

    Txurce Deity

    Joined:
    Jan 4, 2002
    Messages:
    8,263
    Location:
    Venice, California
    Good point. I always build them, but not as early as I should.
     
  15. Stalker0

    Stalker0 Baller Magnus

    Joined:
    Dec 31, 2005
    Messages:
    6,249
    For me I'm actually looking at it the opposite way. I think the end game is too hard, in that the AI's ramp up is late but too sharp. You can be neck and neck with the AI and think you are doing very good (or even dominating), but then suddenly the AI kicks in to that final gear and just gets further and further away.

    Also, this creates a greater disparity between the win conditions. Conditions that take longer to accomplish (like science) will be generally harder because you are dealing with those late AI handicaps for a longer period of time. Whereas CVs that tend to be shorter do not. A change in the curve won't solve that completely of course, but narrows the gap a bit.

    So really I was pushing for a less steep curve. Take down some of the late game ramp up in exchange for some bonuses more in the middle. I mean I am also fine just lowering the end game bonuses and going up a difficulty if I truly need to (probably wouldn't need to for me)...but that may make the game too easy for some of the diety players who win most of their games anyway.
     
  16. Txurce

    Txurce Deity

    Joined:
    Jan 4, 2002
    Messages:
    8,263
    Location:
    Venice, California
    This conversation is a good example of why changes need to be very well considered. Different playstyles and skill levels affect how difficult an era may seem. I would need a lot of convincing to want to veer from a smooth, era by era, numerical ramp-up.
     
  17. Stalker0

    Stalker0 Baller Magnus

    Joined:
    Dec 31, 2005
    Messages:
    6,249
    I wonder if you and I are actually saying the same thing:)

    I am actually really good with the model we have developed, of the AI getting weaker bonuses early on, and steadily increasing them as the game goes on. But I think the curve could use some tuning. Right now its too exponential to me. Its too weak for too long, to where a player can get a false sense of confidence...to the point where people may quit games because "they have it in the bag". But as the game goes on, those bonuses don't just increase, that exponential curve hits hard. You go from "I got this" to "Um....I still got this" to "Holy crap how do I catch up!". I would like to modify the curve slightly so that the game get more competitive earlier but doesn't ultimately cap as high.
     
  18. ElliotS

    ElliotS Warmonger

    Joined:
    Jun 13, 2013
    Messages:
    2,704
    Gender:
    Male
    Location:
    Chicago
    Well for that you 100% need to play around with the handicap scalers. Find a value that you think works well and post it for discussion, so we can try it out. We also may want different curves at different difficulties. It seems like manageable Deity difficulty curves are often too weak for king players on king because of other bonuses or lack thereof.
     
  19. Txurce

    Txurce Deity

    Joined:
    Jan 4, 2002
    Messages:
    8,263
    Location:
    Venice, California
    We are saying the same thing, in that I agree with how you want to proceed. I was pointing out that the disagreement between you and tu_79 was a good example of the need to be careful in how we proceed, and not base it on any single player's opinion.
     
  20. MorphBer

    MorphBer Warlord

    Joined:
    Nov 21, 2017
    Messages:
    180
    Maybe the fact, that humans effectively lost one policy from the oracle also has its part in the culture issue.
     

Share This Page