[NFP] Have Firaxis ever adressed whether the AI voting on diplomatic resolutions is working as intended?

[kaspergm]

This is an example from my recent game, but it's something that seems to happen more often than not:

Spoiler :

I know one *can* explain this with something like "China is persuing a diplomatic victory" (I don't think that was particularly the case, but can't rule it out), but like I said, this seems to happen so often that I can't help but wonder if there's some sort of error in the code that causes the AI to vote opposite of what it's intended to do.

 

[jasper]

Huh? People dont have the context of the last 80 turns of your one particular game.

Wha??

Did they fix diplomatic? No lol.
Do crappy things happens? Yes lol.

 

[iammaxhailme]

In my most recent game, everybody in the world downvoted 1 vote against an aid request on me (I am allied or friends with everybody) and at the same time upvoted a total of 40 votes to pass an aid request on the Netherlands (who are friends with a few, but mostly unfriendly or at war with everybody else). I don't get it.

 

[Qwan]

It seems to me that AI voting has changed somewhat with the last release, or maybe with the last few releases, although I may be seeing something that's not there. (Why do I think so? Because before virtually all AIs voted B on Patronage, but more recently I was able to help A pass. Although it could have been just that game.)

Anyways, regarding your specific inquiry about Request for Aid, there may be many factors affecting AIs supporting or voting against it. Yes, an AI pursuing a Diplomatic Victory, or an AI that likes Emergencies like Canada, are predisposed to vote for it. Score may also matter, as when I was leading the scores, AIs mostly voted against sending me Aid, while they favored Aid to me in another instance where I was near the bottom in scoring. My neighbors tend to vote against me, even if Friends or Allies. I've seen an AI vote in favor of Aid to someone with whom they are at war. Although relationships do seem to have an effect on voting for or against Aid.

But as for notifications of an AI pursing a specific victory, I generally ignore them. I've seen military-weak AIs begin to pursue a Domination victory. I've seen AIs change their mind a lot about what victory they are now pursuing or now no longer pursuing.

 

[Xur]

I honestly doubt the “AI” is much more than a random roll when it comes to resolutions. Either it’s random or the design of this feature makes it feel random. If I could disable resolutions, then I would.

 

[leandrombraz]

AFAIK relationship status doesn't have any weight in these votes, which is consistent with regular resolutions, where the AI votes based on specific parameters that establish what they "need".

What I wish Firaxis would acknowledge and change is how single minded the AI is. It's a shame that they gave resolutions two options, but there are resolutions where you don't actually have two options, because the AI will always vote on the same option and it's virtually impossible to outvote them. "Duplicates of the chosen Luxury resource grant Amenities", for example, is a really interesting bonus that would be nice to get once in a while and strategize around it, but it virtually doesn't exist in the game, since you will never, ever, be able to activate it, at least not in larger maps. "Earn double points towards Great People of this class" could be a great way to speed up the race for GP, but it doesn't exist, there's no such bonus in the game. It's there, but it isn't, because the AI is incredibly single minded. It's like you're playing with a bunch of Borgs.

 

[PiR]

Who was the resolution intended to help? You or China?

 

[Qwan]

I honestly doubt the “AI” is much more than a random roll when it comes to resolutions. Either it’s random or the design of this feature makes it feel random. If I could disable resolutions, then I would.

Definitely not random. For instance, in all my games, for Border Control the AI always places 1 (free) vote for themselves. Which is broken, because I can simply spend 10 favor to win it. Although, maybe the AI votes differently on easier levels of difficulty, or different size maps (i.e. with different numbers of players).

... the AI will always vote on the same option and it's virtually impossible to outvote them. "Earn double points towards Great People of this class" could be a great way to speed up the race for GP, but it doesn't exist, there's no such bonus in the game.

As I wrote, in my most recent game I was able to help A pass for Patronage. The first time, for double Merchant points (which I wanted). The second time, for double Scientist points (which several AIs wanted). But, interestingly, it didn't speed up getting Great Scientists as much as you might imagine. Einstein became available two (or three) eras early, so was extremely expensive. The race for Great Scientists stalled, despite double points, until after the resolution was complete. When Einstein was finally taken, so were all the other Scientists of that Era, in a single turn, because they weren't so overpriced anymore (and I didn't even get a chance to buy Nobel).

Who was the resolution intended to help? You or China?

The right side of the screen shows that the target of the Aid Request is Cleopatra. In this case, that would be the player, who is always listed first in a single-player game on the left side of the screen. Besides, in my experience, if the Aid Request is for an AI, that AI spends a lot of votes for it, not just one.

 

[leandrombraz]

As I wrote, in my most recent game I was able to help A pass for Patronage. The first time, for double Merchant points (which I wanted). The second time, for double Scientist points (which several AIs wanted).

It's possible to do it if you play on smaller maps, with less Civs, which not only reduces how many Leaders are voting, but it also means that when an AI does vote for the other option, it has considerably more impact on the outcome.

 

[kaspergm]

AFAIK relationship status doesn't have any weight in these votes, which is consistent with regular resolutions, where the AI votes based on specific parameters that establish what they "need"..

I don't think that's true. I have seen numerous cases as also reported above, where several aid requests have been up, and where the AI will vote against the one they are friendly with, but for the one they are hostile towards or even at war with, as also reported above:

Spoiler :

In my most recent game, everybody in the world downvoted 1 vote against an aid request on me (I am allied or friends with everybody) and at the same time upvoted a total of 40 votes to pass an aid request on the Netherlands (who are friends with a few, but mostly unfriendly or at war with everybody else). I don't get it.

I have a screenshot of this from an old game, where Peter (Russia) was at war with Frederick (Germany) and friendlywith me (China), yet Peter votes against an air request from me, but for an aid request from Germany:

Spoiler :

 

[PiR]

That would make me think they vote the opposite of what they are supposed to.
Just like they say the opposite of what they want to say.

 

[Basajaun]

AFAIK relationship status doesn't have any weight in these votes, which is consistent with regular resolutions, where the AI votes based on specific parameters that establish what they "need".

What I wish Firaxis would acknowledge and change is how single minded the AI is. It's a shame that they gave resolutions two options, but there are resolutions where you don't actually have two options, because the AI will always vote on the same option and it's virtually impossible to outvote them. "Duplicates of the chosen Luxury resource grant Amenities", for example, is a really interesting bonus that would be nice to get once in a while and strategize around it, but it virtually doesn't exist in the game, since you will never, ever, be able to activate it, at least not in larger maps. "Earn double points towards Great People of this class" could be a great way to speed up the race for GP, but it doesn't exist, there's no such bonus in the game. It's there, but it isn't, because the AI is incredibly single minded. It's like you're playing with a bunch of Borgs.

Came to comment on this one. Especially egregious, since sometimes the votes come soon enough that I didnt still discover the other half of AIs, and... know which amenity all those unknown civs half a world away vote against? Yep, the one you have most copies of, one those unknown civs shouldnt even know it exists.
I roll my eyes everytime that vote comes, I just know Im going to lose it. Doesnt matter if I got Himiko and Ive been pooling favor the whole game. There is no mathematical way to outvote the world, and allies vote against you most of the time on any relevant matter, so you are not "voting" anymore.
You are just trying to guess the outcome of a vote, and betting accordingly, regardless of the type of resolution.
Diplomacy needs to be looked over. Either completely overhauled, or the decision making process altered, and cost of votes modified. Votes costing exponentially more means that the only way to win votes on larger maps, is eliminating AIs completely, making then unable to vote

 

[Baigan]

Came to comment on this one. Especially egregious, since sometimes the votes come soon enough that I didnt still discover the other half of AIs, and... know which amenity all those unknown civs half a world away vote against? Yep, the one you have most copies of, one those unknown civs shouldnt even know it exists.
I roll my eyes everytime that vote comes, I just know Im going to lose it. Doesnt matter if I got Himiko and Ive been pooling favor the whole game. There is no mathematical way to outvote the world, and allies vote against you most of the time on any relevant matter, so you are not "voting" anymore.
You are just trying to guess the outcome of a vote, and betting accordingly, regardless of the type of resolution.
Diplomacy needs to be looked over. Either completely overhauled, or the decision making process altered, and cost of votes modified. Votes costing exponentially more means that the only way to win votes on larger maps, is eliminating AIs completely, making then unable to vote

Agreed on both the difficulty of winning luxury proposal A and the absurdity of civs voting on proposals they shouldn't have knowledge of. Emergency votes do require visibility on the target/member civs, so I'm not sure why similar visibility requirements aren't implemented in World Congress.

However, I disagree that the AI always votes against the players on the luxury vote. It's a deterministic vote that doesn't distinguish between AI vs player civs. The issue isn't the voting, it's the fact that players are better able than the AI to settle high-value cities with lots of luxuries AND improve them. The reason it feels like it targets the player more is because the AI often doesn't have nearly as many luxury tiles improved. It's still a feel-bad mechanism because there's no chance for option A to pass, but the voting algorithm isn't the problem.

 

[Qwan]

I used to be bothered about the Luxury resolution. Eventually I realized it's just a loss of 4 amenities, for 30 turns, which generally doesn't hurt me much. What astonishes me is that the AI spends so much favor to do that. Which I exploit by spending my favor on other or future resolutions.

To guess at what will win, and so turn my free vote into a diplomatic victory point, I escape out of the World Congress Screen and look at the Global Resource Report. Unfortunately, that report's incomplete: while it includes your trading, it doesn't seem to include trading between AIs (which is fair), nor luxuries AIs obtained from suzerain'd city-states. The latter you can discover by reviewing the city-states. The winning (or is it losing) luxury is usually one of the few at the bottom of that report. Basically, the AI seems to want to hurt as many other Civs as they can, without hurting themselves. What makes it uncertain in my games is that often, for a given AI, multiple luxuries equally qualify.