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

Anti-Wonder Snowballing

Discussion in 'General Balance' started by Gidoza, Mar 6, 2020.

  1. Gidoza

    Gidoza Emperor

    Joined:
    Jul 26, 2013
    Messages:
    1,160
    So, much is being observed and commented on about snowballers and one player getting all the Wonders. I was thinking that a different method could be used to approach at least the Wonders part of it.

    How about a manipulation of the current "failure reward" when someone else gets the Wonder before you do? I'll explain what I mean.

    First, I'll start by saying that the failure reward is rather awful. I know that the counter-argument to bolstering it is that it's abusive somehow, but I can't say that I could see it - at least not for the underdog. And there's two ways to be an underdog here.

    A - You're the guy with 1 Wonder to someone else's 10.

    B - Someone else finished a Wonder when you had 3 turns or left to finish it.


    I'll deal with B first.

    Obvously, the risk involved with investing in a Wonder is that you don't get it. However, I would say that *not getting the Wonder is punishment enough in and of itself* - while obviously 100% returns is unreasonable, I quite frankly don't see the problem with dumping most of the rest of that production into unbuilt buildings, because you still don't have the Wonder.

    This isn't so bad if you're only like 3 turns into building it. But if you have only 3 turns *left*, the punishment increases the more you've invested in it. The point here is that this can be game-breaking in its own right: a gap that is needlessly artificial is bring created between a player who succeeded in building a Wonder and one who did not.

    A way of re-expressing the failure reward formula would be to say that the more you have invested in a Wonder, the higher proportional total returns you receive. This will naturally encourage players to aim at wonders that they *might* get - because they might actually succeed, and that success would pose a challenge to the front-runner. Upon failure, the investment still isn't a waste of time because the rewards for barely missing are still substantial.


    So the question is - what if he front-runner barely missed? This is where we go to the aforementioned A situation.

    How about we mediate failure rewards in a similar manner to how techs decrease for a player for each other player that has that tech?

    What this would mean is - for each Wonder I own, my reward for failure is determined proportionally against the player who completed it. If I have 5 and the person who completed it had 5, then the reward is moderate; if I have 10 and the player who completed it has 1, the reward is awful; if I have 1 and the player who completed it has 10, the reward is substantial.

    The setup here is such as to incentivize players to challenge the front-runner as much as possible whenever possible for those Wonders: the front-runner will either get beaten out to the Wonder and get poor payback, or will get the Wonder, but at a cost that if he leads too much, the rewards for failure will be insane.

    It's a subtle catching-up mechanic.

    Brief Synopsis: Like having Bronze/Silver rewards in WC projects (where Solve is within 3 turns of completion), but the reward is magnified or diminished based on how many Wonders the person who completed the Wonder has in relation to you.
     
    Last edited: Mar 6, 2020
  2. Moi Magnus

    Moi Magnus Emperor

    Joined:
    Mar 1, 2015
    Messages:
    1,866
    There is already an anti-wonder snowballing method: production cost increase according to your number of recent wonders.
    As a consequence, "building a wonder" is less of an investment for peoples that have only few wonders (hence less of a loss if they miss), and more of an investment for peoples who have a lot of them.

    Before thinking about adding a new mechanics, would lowering starting cost and increasing the penalty from already having wonder be a solution?
    A wonder could cost "barely more than a regular building" if you don't have any, and quickly rise up to the current cost, and then much more.
     
    Zanteogo, bigcat88, Recursive and 3 others like this.
  3. Gidoza

    Gidoza Emperor

    Joined:
    Jul 26, 2013
    Messages:
    1,160
    As has already been discussed in other threads, the mechanic of increased cost according to your recent wonders really isn't stopping anything. If you're ahead, then you'll have more techs and policies than anyone else, allowing you to start building the wonder first, and faster as well (seeing as you're ahead, you'll have more production too) - thus totally nullifying this so-called "penalty." Snowballing is happening in spite of it - so unless we make the penalty so extreme that the leader cannot build Wonders whatsoever after a few constructions (not somewhere I think any of us want to go), it seems to me that we need to have entirely a different plan. It's not about how much of an investment it is to build it - it's about the long-term benefits, and whether you manage to complete it at all.
     
    Moi Magnus likes this.
  4. CrazyG

    CrazyG Deity

    Joined:
    Oct 14, 2016
    Messages:
    4,781
    Location:
    Beijing
    I don't think the production penalty slows down the wonder spamming AI all that much. On higher difficulties the AI is drowning in extra production anytime they aren't expanding their army. Often they were just doing the research process before switching to the wonder. I do feel it as a human player though.

    Just one idea for wonder spamming, we could take a look at spies. First of all level 1 spies are garbage.
    Typically he spends like 30 turns getting my hopes up, before dieing to counterspies. I almost never successfully steal a tech with a level 1 spy.

    Secondly, the tech leader AI almost always takes banking quickly and will almost always have a constabulary built before your first tech steal, often before you even establish surveillance. It slows you down a lot.

    What I usually do is rig elections till level 2. The two elections take 15 turns each, and the traveling probably takes 6 turns. Then he travels to a capital (3 more turns), then he establishes surveillance (3 more), and then he finally begins moving towards a tech steal, which will take 25 to 30 more turns. That's around 70 turns after receiving my spy that a tech steal might happen (he could die to counterspies). It isn't uncommon for teh AI to be in industrial era at this point, which means they have 2 spies, and probably counterspy their capital.

    I've tried going to non-capital to avoid the frustrating spy deaths, but typically it increases the time to steal a tech by 10 to 20 turns.
     
  5. Gidoza

    Gidoza Emperor

    Joined:
    Jul 26, 2013
    Messages:
    1,160
    I'll throw in here that it's not just snowballing AIs. Often it's an AI, but often it's me as well. I don't care *who* is snowballing...a methodology to limit it isn't a bad thing; but in principle I never mind someone's being in first place.
     
    Rhys DeAnno likes this.
  6. amateurgamer88

    amateurgamer88 Emperor

    Joined:
    Aug 24, 2018
    Messages:
    1,112
    Gender:
    Male
    Personally, I don't like the idea where one civ or the player just grabs a lot of Wonders. Wonders are meant to be powerful and grabbing nearly all of them in an era due to being ahead just helps the first place snowball even faster. I also rarely find in my games the AI getting punished for grabbing so many Wonders. They have everything they need and still get the Wonders they have access to. I wouldn't mind see the cost of building a Wonder in the same era being noticeably higher so that it's tougher for anyone to get all the Wonders in one era. What I noticed this the most is the Renaissance Era. You have one leader who eventually takes most, if not all, of these Wonders because the civ has both the tech and policies to get these while others cannot due to lagging in tech and/or policy.
     
    Rhys DeAnno and lunker like this.
  7. SuperNoobCamper

    SuperNoobCamper Warlord

    Joined:
    Aug 27, 2017
    Messages:
    296
    Gender:
    Male
    As much as i hate dealing with wonder-spamming run-aways Looking at you,Tradition Ethiopia i don't really think any radical approach to this issue is necessary.
    Just number tweaks for both the base cost and if necessary cost modifier for other wonders would be sufficient IMO.
     
  8. tu_79

    tu_79 Deity

    Joined:
    Feb 11, 2016
    Messages:
    6,910
    Location:
    Malaga (Spain)
    If being successful at building a wonder is a reward in itself, then why is the AI receiving additional yields for doing that? It's as if the AI had built three wonders instead of just one.

    All this snowballing out of control comes from the idea of rewarding already rewarding, sum zero games, actions. It goes the same way for players that are successfully conquering cities, but here at least there's the happiness mechanic to keep things under control.

    Being the one that manages to build a wonder also means that you are not one of those who missed the wonder and lost his investment.

    If we want to keep things simple, then don't give AI handicaps for completing a wonder.
    If we want things complex, then increase the reward for not being able to complete the wonder x3.
     
  9. LifeOfBrian

    LifeOfBrian Warlord

    Joined:
    Aug 21, 2019
    Messages:
    207
    Brainstorming here, but one potential option would be to introduce an empire-wide cooldown period for building world wonders. So for example if you built a world wonder on turn 100, you could only start working on another one after xyz number of turns have passed.
     
  10. tu_79

    tu_79 Deity

    Joined:
    Feb 11, 2016
    Messages:
    6,910
    Location:
    Malaga (Spain)
    I would dislike not being able to go for a couple of wonders of the same era. Sometimes your civ is blessed with momentum, and you have to make the most of it if you want to stay competitive later.

    The problem is not in the human side, current penalty is big enough to make us reconsider building too many wonders for the same era. It is just AI in the higher difficulties who get so many yields from doing so, that they can ignore the penalties on wonders, apparently.

    If we try to solve a problem by addressing just part of the effects, we risk introducing more unexpected problems.
     
    Rhys DeAnno and SuperNoobCamper like this.
  11. LifeOfBrian

    LifeOfBrian Warlord

    Joined:
    Aug 21, 2019
    Messages:
    207
    Tu_79, I understand your concerns about introducing unwanted problems and I agree there might be problems with this idea, I just wanted to throw it out there for discussion.
     
    tu_79 likes this.
  12. CrazyG

    CrazyG Deity

    Joined:
    Oct 14, 2016
    Messages:
    4,781
    Location:
    Beijing
    I think the problem is the AI's big bonus yields for building a wonder.

    If a human is building wonders like crazy, there is a solution: go to a higher difficulty.
     
    Zanteogo and tu_79 like this.
  13. Heinz_Guderian

    Heinz_Guderian Chieftain

    Joined:
    Jan 2, 2020
    Messages:
    82
    Location:
    Sweden
    They simply shouldn't get bonuses from building wonders. I find it so irrational that they should get event based yields at all since that just promotes runaways. It is frankly unbelievable that it still is in the game.
     
  14. Gazebo

    Gazebo Lord of the Community Patch Supporter

    Joined:
    Sep 26, 2010
    Messages:
    17,825
    Gender:
    Male
    Location:
    Little Rock
    The AI gets GAP and Gold for a wonder. Not a huge amount.

    G
     
  15. CrazyG

    CrazyG Deity

    Joined:
    Oct 14, 2016
    Messages:
    4,781
    Location:
    Beijing
    how many GAP?
     
  16. tu_79

    tu_79 Deity

    Joined:
    Feb 11, 2016
    Messages:
    6,910
    Location:
    Malaga (Spain)
    Standard King, Polynesia gets 86 gold from building a wonder at turn 235, that's not much actually.

    Polynesia, 235, CBP AI DIFFICULTY BONUS FROM HISTORIC EVENT: WONDER - Received Handicap Bonus (86 in Yields): GOLD, GAP.

    Here's something that explains a bit some of these differences:
    Polynesia, 250, CBP AI DIFFICULTY BONUS FROM HISTORIC EVENT: GP - Received Handicap Bonus (132 in Yields): GOLD, GAP
    The Iroquois, 252, CBP AI DIFFICULTY BONUS FROM HISTORIC EVENT: GP - Received Handicap Bonus (86 in Yields): GOLD, GAP
    Polynesia, 253, CBP AI DIFFICULTY BONUS FROM HISTORIC EVENT: DIG/TRADE - Received Handicap Bonus (132 in Yields): GOLD.
    Austria, 253, CBP AI DIFFICULTY BONUS FROM HISTORIC EVENT: GP - Received Handicap Bonus (86 in Yields): GOLD, GAP
    Songhai, 254, CBP AI DIFFICULTY BONUS FROM HISTORIC EVENT: GP - Received Handicap Bonus (52 in Yields): GOLD, GAP
    Austria, 254, CBP AI DIFFICULTY BONUS FROM HISTORIC EVENT: GP - Received Handicap Bonus (86 in Yields): GOLD, GAP

    Turn 250 Polynesia gets x2.5 yields compared to Songhai a few turns later. That's because they get the scaling on the era that the player is still on. A civ that's 2 eras in advance gets more than double the handicap yields.

    Couldn't this work as the free spy on Renaissance, that grants a free spy to every player when the first player gets to the era?
    Actually, most handicaps could scale on the highest era ever achieved, otherwise the lowest players become irrelevant.
     
    bigcat88, cerk and Rhys DeAnno like this.
  17. Gazebo

    Gazebo Lord of the Community Patch Supporter

    Joined:
    Sep 26, 2010
    Messages:
    17,825
    Gender:
    Male
    Location:
    Little Rock
    Could use Game.GetCurrentEra() which is the 'average' era of all major civs in the game.
     
  18. kenneth1221

    kenneth1221 Warlord

    Joined:
    Apr 18, 2016
    Messages:
    203
    Small question: is that rubberbandy, or would the effect be closer to ensuring a reversion to the mean slowly enough that you wouldn't notice?
     
  19. Gazebo

    Gazebo Lord of the Community Patch Supporter

    Joined:
    Sep 26, 2010
    Messages:
    17,825
    Gender:
    Male
    Location:
    Little Rock
    I don't think it is 'rubberbandy' in that it's not giving an advantage, so much as flattening outliers. Besides, it won't affect human play, as the values won't ever be 'visible' in the UI.

    G
     
    Rhys DeAnno and Heinz_Guderian like this.
  20. kenneth1221

    kenneth1221 Warlord

    Joined:
    Apr 18, 2016
    Messages:
    203
    Personally, I'm all for it. May as well limit the snowballing of snowballing.
     

Share This Page