Stupid religious behavior

Cybah

Emperor
Joined
Jun 22, 2007
Messages
1,481
An AI should NEVER use (I don't know the English word that fits) a religion as state religion if:

The AI has NOT founded this religion AND no other AI is also "using" this religion as state religion except this religion is the only religion the AI can "use" at this time. The AI should switch to another one as soon as possible.

The reason should be clear: The AI has disavantages only: More potential enemies without a chance for a religious friend.
 
So always switch religion (unless/until you found your own religion) to match your strongest neighbor? Would that be a sensible check?
 
maybe not. this neighbor would probably attack you in the future even if you are a religious friend when your strength is very low (depends on AI character?). thoughts?
 
The initial post is not entirely accurate- you say "it's not the only religion it can use at this time" but technically it can switch if it has 8 cities with Hinduism and 1 with Islam. It should check all factors:
Need for happiness from the religion
Who controls the holy city/who founded
are there relationships that need improving that would be improved by switching religions

Consider- if you plan to go to war with someone, you don't want to share their faith- that makes your people get annoyed more quickly! If I'm strong enough to not care that my heathen friends might get annoyed, I may well enjoy there being the excuse that might cause them to declare on me.
 
+points if you LIKE the controller of the founder of the religion
-points if you HATE the controller of the founder of the religion
+points if you LIKE civilizations with that religion
-points if you HATE civilizations with that religion
+points based on the fraction of your civilization that follows that religion
+points if you own the holy city for that religion
+points of the religion is the new hotness
-points if the religion is old and busted

The last two are to encourage more religious dynamism, and make things more fun. And if all of the AIs agree that switching to a newer religion is a good thing (in and of itself), it means that the player has to either do that, or be considered a religious outcast...
 
Yakk, I find it very hard to agree with your last two points.

Firstly, what if the human is the one who founded the newer religion (that's pretty common)? AIs are jumping out of their shared religions to the human's one, making an easy game.

At the same time, all those AIs are putting themselves through anarchy, for no real gain.

Unless you are suggesting the last two points be very very small contributors, I don't see how it will improve anything.

IMO, there should be a factor which increaseses the likeliness of switching religion if Spiritual, since there is no anarchy anyway. This should be multipled once the sum of those points is done, rather than a separate point to be included in the sum. The points you summed in your post should determine what I'd call "Does the AI want to switch religion". Then a random factor helps them decide whether to go ahead with it.
 
Sure, the human player could have the late religion -- but the human player could have the early religion too.

And yes, there is an anarchy cost to switching religions. But those who refuse to change end up being punished (through diplomatic hits).

Once everyone wants to change religions to the 'new hotness', anyone who doesn't take the anarchy hit ends up being worse off (as their diplomatic ties fall).

Religion happens to be part of the roleplaying parts of the game. A huge chunk of it's impact is in diplomatic modifiers. And having the AIs roleplay "new world religions tend to spread over old ones" forces everyone (including the player) to change religions in order to keep diplomatic penalties from being applied.

The states that should refuse to change -- those with a shrine to their own religion -- can end up being diplomatically isolated. This isn't a bad thing either -- states with old shrines are getting a ridiculous economic benefit out of it (as each missionary produced generates 1 gold per turn forever). Undercutting them doesn't seem that silly of an idea...

Although that does bring up another idea. Have a bias towards changing to a religion that is founded by a weak civilization, in order to prevent the benefits from flowing to a civilization that is too strong, if you don't have one yourself.
 
Sure, the human player could have the late religion -- but the human player could have the early religion too.

And yes, there is an anarchy cost to switching religions. But those who refuse to change end up being punished (through diplomatic hits).

Once everyone wants to change religions to the 'new hotness', anyone who doesn't take the anarchy hit ends up being worse off (as their diplomatic ties fall).

Religion happens to be part of the roleplaying parts of the game. A huge chunk of it's impact is in diplomatic modifiers. And having the AIs roleplay "new world religions tend to spread over old ones" forces everyone (including the player) to change religions in order to keep diplomatic penalties from being applied.

The states that should refuse to change -- those with a shrine to their own religion -- can end up being diplomatically isolated. This isn't a bad thing either -- states with old shrines are getting a ridiculous economic benefit out of it (as each missionary produced generates 1 gold per turn forever). Undercutting them doesn't seem that silly of an idea...

Although that does bring up another idea. Have a bias towards changing to a religion that is founded by a weak civilization, in order to prevent the benefits from flowing to a civilization that is too strong, if you don't have one yourself.

Yes but if you own the shrine you still get the gpt when you are under another religion.

I don't know much about the way religion switches happen at the moment, but they often seem sensible at least. My point about the human religion, I think, is that normally at the higher difficulties the human player never gets the first one or two (or even third) religion. They need to get CoL or Theo etc. for the religion - one of the later religions.

You might want more religious dynamics, but civs more likely to adopt the newest religion, more games will end up with Christianity or Islaam dominatoring, and you'd have the same problem. Apart from lack of uptake of the last maybe 3 religions, I've seen the earlier 4 become dominant in different games.

And I still think you can have this idea of changing religions without the need for choosing the newer one. The newer ones are less likely to spread further because for most cities it's hard to get more than two religions. I don't think AIs even try to spread religions to cities that have two religions already, because of the lower odds.

As much as religion is part of the roleplaying in the game, I kind of thought the extent religion already played in the game was sufficient.

It's just that having the world want to adopt the newest religion, even though it is the religion most likely to spread the least, seems odd. I say they keep their old religion (you can call it their tradition). If anything, I'd be inclined to say civs should have a tendency to keep old religions, because they have a richer history in that religion than the newer one, and we all know how stubborn populations are to change. :)
 
In the roleplaying sense, that isn't how it worked in the "real world".

With the exception of areas with a "founding religion", or areas cut off, the later "great religions" became dominate over most of the world.

The bias doesn't have to be ridiculously large. People with the shrine should tend to keep their religion... but those without it might change, then convince their friends to change. Then a block of friendly civilizations form.

Once a block of civilizations that are friends form, civilizations that don't like them resist changing to that religion.

And yes, the fact that the AI 'cheats better' in the early than the late/mid game means that humans tend to be further behind in the early game. One of the points of an improved AI is to level that out -- to make the AI better at the mid/late game. With a good enough AI, you'd have a player who starts ahead in the early game, and then has issues keeping up in the mid/late game. :p
 
I think religion is too big of a thing when it comes to strategy in CivIV to let too much "roleplaying" into it. It should be decisions based solely on strategic reasons IMHO.
 
Civilization is a zero-sum game in a sense.

If everyone changes to a new religion, then everyone is hurt. If you refuse to follow the herd, then you get diplomatically isolated and punished.

There is some common harm from having to spread the new religion over more of your cities -- a region that doesn't switch to the new religion is slightly better off, if it doesn't cause diplomatic harm.
 
Yakk,

I really value all of your contributions and advice toward this mod, and I usually agree with almost every point you make, but the push for the adopting of newer religions for what seems to be only role-playing reasons is really baffling me, and I really disagree with it.

You say the civs will need to follow the herd at risk of falling out of favour with everyone, but someone has to go first right? If 8 civs are buddhist and 1 suddenly changes to Islaam, that opens a window where they are likely to get attacked, and for what reason? Liking a new flavour?

Unless the AIs are aware in advance that other AIs are going to follow suit, and not cancel deals/declare war in the meantime, I really don't see how switching religions is a beneficial move, strategically.

I understand you are trying to make this one factor out of several... my view is that this should easily be the smallest factor, and really I don't see why it needs to be a factor at all.

I think a good way to approach AI improvements is for the AI to consider every other player a human. Playing against 8 Buddhist humans, an AI is hardly going to have an easy game if it switches religion to something new. Ok with 8 humans you could argue the humans won't care about religious differences either, but because they are AIs they are in fact programmed to hate heathens.

Maybe a better way to put it is develop the AI in a way where it is not aware of the way other AIs will react. If we developed 8 unique AIs (8 different better AI mods maybe...) which one would win more games? The one that switched religion away from other AIs to the newer one?
 
I'm not in favour of the religion change because in my view it is more of a gameplay change than an AI change and thus not really what this mod is about.

I do really understand that there is some attractiveness to creating a mod that changes the religion element in the game to one where newer religions have a chance. In this sense a mod where more than one religion can naturally spread to a city combined with a higher natural spread rate for the later religions, inquisitors that remove religions from cities and unhappiness when 2 or more religions are present in a city can make religion a far more flexible and changing element in the game. It could really make religion more interesting if implemented well, but... I really don't think it belongs in this mod.

What is suggested by Yakk is of course far less game changing, but in my opinion it is not a move to a better AI but more of a change to the religion system.
 
An AI that doesn't know the other AIs care about religions shouldn't care about what religions the other AIs use.

But that isn't how the AI works -- it forms friendly blocks based on their current religion.

Players then have to react to this fact that AIs react diplomatically to religion choice, and are forced to pick sub-optimal religions in order to get diplomatic results they like...
 
An AI that doesn't know the other AIs care about religions shouldn't care about what religions the other AIs use.

But that isn't how the AI works -- it forms friendly blocks based on their current religion.

Players then have to react to this fact that AIs react diplomatically to religion choice, and are forced to pick sub-optimal religions in order to get diplomatic results they like...

Which is a nice way to introduce some "sacrifice" to the player in the name of diplomacy. I like it as it is.
 
Speaking of stupid religious behavior, in my current (multiplayer, 3 humans 6 AIs) game, Elizabeth is really hosing herself because every 20 turns or so she switches religions back and forth between Hinduism (whose founder is on one side of her) and Buddhism (whose founder is on the other). She's already at the bottom of the leaderboard but these frequent turns of anarchy are really not helping. I don't know how high the threshold is for deciding to change state religions, but I wonder if it being a smidge higher would prevent these two from constantly leapfrogging each other in her decision calculus.
 
Logically, everyone would change to the same religion, but then that would make for a boring game.

I think the current way works OK, no need to mess with it.
 
I don't think logically everyone would change to the same religion. But there are some religion choices that are just suicidal for an AI player, and this mod is all about making the AI smarter. I think there are very few players who would continue using a religion that all of their AI neighbors are not using if that player doesn't also control the holy city. There would be no point unless there's no other religion in the player's cities. Should we have different AI's adopting deleterious behaviors for the sake of making things more dynamic?

The AI doesn't take diplomacy into account currently. Nor does it take into account civics... if you aren't running any civics that give state religion bonuses for each city, the number of your cities that have the religion is less signficant. But these things do matter, and diplomacy should be weighed in along with the factors it already takes into account.
 
Top Bottom