Diplomacy Feedback: Opinion Mechanics

[Recursive]

Ok I'm going to start my feedback. Considering the immensity here, I am going to do my post in chunks below to keep things organized, but I will update this post frequently as I work through things. So please check back for the latest!

OPINIONS

• I think there should be some adjustment for the late game. 50 turns early on makes plenty of sense, 50 turns after Turn 300 on standard speed might as well be forever. Perhaps an Era modifier or something should be considered, I don't necessarily need the modifiers to scale continuously all game, but at some point the 50 turns base should be shorter to have any reasonable attempt at changes in the late game.

Not a bad idea, but I wonder how to implement it.

• I don't understand this statement: When a "stacked" modifier duration expires (for things like demands or stealing), the number of times the AI remembers you have done that thing is halved and then rounded down. If after doing this there is still at least 1 application of the modifier, the duration is reset back to 0 and ticks down again until all applications have expired.
  • How can the duration tick down from 0?

Resets to its original value. Will fix, thanks.

• Under the No Contested Borders area, there is this note for warmonger AIs (Warmonger AIs apply an additional bonus: (5 + Neediness) * AI Era / 2.). I would think that we would want the bonus to go down for warmongers in the late game, as warmongers are going to want to conquer far away lands as much as neighbors.

Border disputes are generally worse the closest you are, because it's easier for the tactical AI to conquer a neighbor, but you may have a good point on this with the improvements to it. The intent of the non-competition modifiers is to give a powerful boost that can enable friendship even with problematic civs if the player is staying out of their way. But considering their "diplo personality" may not match the victory condition they're going for, I've been wondering about a better alternative.

• Under We are Trade Partners, If the deal value to the AI is >= 100 or they're receiving 5x+ what they're giving up, a larger bonus is added. Personally I think 5x is a bit steep, realistically that to me isn't a deal, that is me just giving money to the AI (aka a complete bribe). Even a 2x deal is pretty steep in terms of true dealmaking, giving DOUBLE to the AI what I think a trade is worth is a major thing. I would think this bonus could be 2x or 3x, 5x just seems extremely high.

Also not a bad idea.

• You Freed our Captured Citizens: Settlers are worth 5 civilians in the Ancient Era, and 4 in the Classical Era. So does that mean returning a settler to an AI (assuming this is the first returned) is worth 10 + 5 x 4 = 30? Also I thought all AI settlers captured are automatically turned into a worker. Is it actually possible to return a settler?

Yes, if it's captured by Barbarians.

• You have Shared Intrigue with them: I feel that +10 is a bit low for coop war sharing (aka the you have gone to far). The other ones are passive, they don't hurt your diplomatically, but considered this one is me "sticking my neck out" I think I should get a little more for it.

Keep in mind the bonus is +40 when you add recent assist into it.

Responses to what you've posted so far.

 

[ridjack]

AI can pretend to have a non-hostile attitude even if Transparent Diplomacy or Show Hidden Opinion Modifiers are enabled.

Is there a technical reason this is so? Because on its own, it's kinda really uncool to ignore user choice like that.

 

[Recursive]

Is there a technical reason this is so? Because on its own, it's kinda really uncool to ignore user choice like that.

It's not a technical limitation - it's the way Putmalk designed it. As designed Transparent Diplomacy will reveal hidden modifiers and the number values, but will not allow you to tell whether the AI is being deceptive or not, because you can't tell if they're hiding anything. The setting also doesn't prevent them from using nice lines of dialogue instead of mean ones.

If you could see the AI's real approach at all times it would make diplomacy far less engaging and more gamey; it's also a huge one-way information cheat. If you absolutely want to, though, you can do so by enabling Debug Mode in DiploAIOptions.sql.

 

[azum4roll]

Everything else is fine, but They mistreated your protected City-States, and you didn't look the other way! is still a one-sided interaction that the player can't do to the AI. The player also can't tell the AI to stop bullying CS under their protection.

 

[Solic]

Will this become the new diplomacy feedback thread?

Otherwise I'd suggest to link this guide or put this guide in the OP of that thread as well.

-----------

Also have a question, if it's possible, to lessen (possibly completely negate it) religious difference modifiers when you extinguish another founder's religion, meaning no civ has units of their religion and all cities do not contain their religion?

The following scenario irks me a lot. Let's say 4 civs on a continent, 2 founders. One founder converts the whole continent. Maybe even world religion is passed and now that weak founder is going to hold a grudge indefinitely, even though they reap the benefits from better diplo relations on the continent and even world religion bonuses. It's a sunk cost fallacy for them to remain mad about it.

 

[Recursive]

Everything else is fine, but They mistreated your protected City-States, and you didn't look the other way! is still a one-sided interaction that the player can't do to the AI. The player also can't tell the AI to stop bullying CS under their protection.

Eventually.

Will this become the new diplomacy feedback thread?

Otherwise I'd suggest to link this guide or put this guide in the OP of that thread as well.

-----------

Also have a question, if it's possible, to lessen (possibly completely negate it) religious difference modifiers when you extinguish another founder's religion, meaning no civ has units of their religion and all cities do not contain their religion?

The following scenario irks me a lot. Let's say 4 civs on a continent, 2 founders. One founder converts the whole continent. Maybe even world religion is passed and now that weak founder is going to hold a grudge indefinitely, even though they reap the benefits from better diplo relations on the continent and even world religion bonuses. It's a sunk cost fallacy for them to remain mad about it.

This is a temporary thread. The guide info will be stickied eventually.

No bonus or penalty if zero cities are following their religion and the Holy City has been extinguished might be fair.

 

[notaspambot]

May I ask for clarification on how the "We are their greatest rival in the World Congress / United Nations." modifier works? If it doesn't take alignment into account, then does the rival refer to the amount of votes you have, or something else?

 

[Recursive]

May I ask for clarification on how the "We are their greatest rival in the World Congress / United Nations." modifier works? If it doesn't take alignment into account, then does the rival refer to the amount of votes you have, or something else?

If going for diplo victory, it's the player with the highest number of votes (not counting the AI, their teammates, their vassals, or their resurrectors). Ties are broken by worst alignment. If still tied, the tie is broken by worst opinion score.

Otherwise, it's the player with the worst alignment. Ties are broken by highest vote count. If still tied, the tie is broken by worst opinion score.

Exception: If a player has 75%+ of the required votes to win a diplo victory and the greatest rival doesn't, that player becomes the greatest rival if victory competition is enabled for that AI. Ties are broken by most votes, and then by worst opinion.

 

[LarryAR]

I am impressed and humbled by this masterwork. Kudos.

 

[ridjack]

Question that I don't think was answered in the main post:

I've noticed that if an AI asks to jointly DoW another civ, you can refuse and then ask them again on your turn and they'll accept, "but I'm telling everyone it was my idea." How does this affect their opinion modifiers?

There are plenty of times I know I probably want to say yes, but I'd like to be able to glance over the map and the current situation before I do. If the above lets me do that, that's probably good enough; otherwise, I would like to put a forward a suggestion for some sort of "let me get back to you on that" option that lets you take a look, then they can come back on the next turn and ask again.

 

[Recursive]

Question that I don't think was answered in the main post:

I've noticed that if an AI asks to jointly DoW another civ, you can refuse and then ask them again on your turn and they'll accept, "but I'm telling everyone it was my idea." How does this affect their opinion modifiers?

There are plenty of times I know I probably want to say yes, but I'd like to be able to glance over the map and the current situation before I do. If the above lets me do that, that's probably good enough; otherwise, I would like to put a forward a suggestion for some sort of "let me get back to you on that" option that lets you take a look, then they can come back on the next turn and ask again.

You get a penalty to recent assist value for refusing to agree, and -1 to your "coop war score" (which is not used for opinion, but affects other things).

Once the coop war begins, you receive +1 to coop war score and recent assist is unchanged.

The dialogue option is a good idea, but will have to wait. I can in the meantime max out recent assist value when you agree to a coop war.

 

[azum4roll]

Would be nice to be able to take a glance of the map in any AI dialogue, not just this one.

 

[Zuizgond]

Would be nice to be able to take a glance of the map in any AI dialogue, not just this one.

Yes especially when AI wants to trade a luxury resource during its turn and you want to know if this specific lux would trigger a WLTKD in any of your cities or if the lux is a CS quest requirement.

If you're not sure and thus decline the trade proposal, then the AI has probably traded it with someone else when you turn starts.

 

[notaspambot]

Yes especially when AI wants to trade a luxury resource during its turn and you want to know if this specific lux would trigger a WLTKD in any of your cities or if the lux is a CS quest requirement.

If you're not sure and thus decline the trade proposal, then the AI has probably traded it with someone else when you turn starts.

CS quests and WLTKD resources can be identified by looking at the diplomacy sidebar and hovering over the relevant resources on the trade deal screen respectively, iirc.

---

I'd like to request some QoL changes to the Request Help function so that players are able to know just how much the AI is willing to spare, and the number of turns until the next request can be made. It'd be a pure QoL change since players can already just spam the button repeatedly until it works (unless the AI is lying and hates you in secret), and there are no penalties if the AI refuses either.

 

[Recursive]

CS quests and WLTKD resources can be identified by looking at the diplomacy sidebar and hovering over the relevant resources on the trade deal screen respectively, iirc.

---

I'd like to request some QoL changes to the Request Help function so that players are able to know just how much the AI is willing to spare, and the number of turns until the next request can be made. It'd be a pure QoL change since players can already just spam the button repeatedly until it works (unless the AI is lying and hates you in secret), and there are no penalties if the AI refuses either.

An extensive amount of work would be required for this.

 

[Legen]

About the "They believe we are building World Wonders too aggressively!", can it not count the wonders that are restricted by policies that the AI didn't go for? It doesn't make much sense for an AI to resent someone for building a WW that they couldn't compete for anyway, such as a Statecraft AI hating someone for "denying" the Red Fort, or an Order AI hating you for "denying" Prora.

 

[Recursive]

About the "They believe we are building World Wonders too aggressively!", can it not count the wonders that are restricted by policies that the AI didn't go for? It doesn't make much sense for an AI to resent someone for building a WW that they couldn't compete for anyway, such as a Statecraft AI hating someone for "denying" the Red Fort, or an Order AI hating you for "denying" Prora.

Not feasible to code. Right now it increments "Wonders Constructed" by 1 every time you build a Wonder (except Congress/Corporation wonders). The amount of work and memory/performance impact is not worthwhile.

Besides, the reason this modifier exists is to punish players who build lots of Wonders at the cost of other things, and thereby make themselves a juicy conquest target. AI can benefit from the Wonder even if they didn't construct it.

There's a separate competition penalty, "You are competing for World Wonders", which does take this into account.

 

[ridjack]

These two modifiers are based on the player's Recent Assist Value. Unlike Vassal Protection, however, only a single value is used for Recent Assist. Diplomatic actions that please the AI add Recent Assist Value, and diplomatic actions that disappoint the AI

Just noticed this small error. That sentence cuts off suddenly.

 

[Recursive]

Just noticed this small error. That sentence cuts off suddenly.

Fixed.