General questions on AI behaviours

[ComradeCrimson]

I've always wondered if AI were ever affected behaviour wise beyond trying to maximize/pick social policies befitting of their focus, for example like if an AI that emphasizes on support or green behaves differently than another in terms of how they fight, build and interact with other players. I am not talking of preferred social policy here explicitly and anti-thesis policies but whether or not social modifier emphasis actually encourages specific behaviours to develop.

So say I got a action that has Emphasis on Police, or Emphasis on Support, or Economy, etc; do these factions exhibit social behaviours based on their emphasis or does the AI behaviour of Pacifist, Aggressive and Normal+ their social preferences tend to dictate everything?

Does social modifier emphasis affect the way the AI operates at all beyond what social policies they tend to pick or focus on?

[Dio]

I can say yes in specific circumstances. The actual pro-ideology or anti-ideology and associated social value itself does not directly influence specific behaviors beyond social model selection and reactions towards running a social model in the same catergory as his or her prefered model. The limitations on this involve the fact that the actual values from specific social areas influence behavior. For example, the fact that the Spartans prefer MORALE does not specifically influence the behavior of this faction. However, the increase in inclinations to run social models that increase the MORALE score of this faction can influence behavior. This further means, for example, that the Morale bonuses from +4 MORALE that the Spartans receive from running Power (the prefered social model of this faction) does affect AI behavior because the Morale level of the units in a faction influences the AI decision making process for a variety of decisions that include a request to stop a vendetta against an ally or to betray an ally. In short, the direct decision to select a certain social model or social area does not have an effect on the AI decision making process. The decisions that stem from this increase in tendencies for certain social areas does influence AI behavior.

[ComradeCrimson]

I can say yes in specific circumstances. The actual pro-ideology or anti-ideology and associated social value itself does not directly influence specific behaviors beyond social model selection and reactions towards running a social model in the same catergory as his or her prefered model. The limitations on this involve the fact that the actual values from specific social areas influence behavior. For example, the fact that the Spartans prefer MORALE does not specifically influence the behavior of this faction. However, the increase in inclinations to run social models that increase the MORALE score of this faction can influence behavior. This further means, for example, that the Morale bonuses from +4 MORALE that the Spartans receive from running Power (the prefered social model of this faction) does affect AI behavior because the Morale level of the units in a faction influences the AI decision making process for a variety of decisions that include a request to stop a vendetta against an ally or to betray an ally. In short, the direct decision to select a certain social model or social area does not have an effect on the AI decision making process. The decisions that stem from this increase in tendencies for certain social areas does influence AI behavior.

Interesting; so its the overall social modifier compositions that result from the social policies that result in specific behaviours? Like if the policy choices they made results in 4+ morale and some support/industry bonuses but a lack of research perhaps they are more likely to be aggressive? Or if they have lacking morale and have more commerce related stuff they'll be more amicable towards trade?

How would one then aim to create specific personalities and patterns by this method? I am intensely interested in this in creating unique AI without extensive exe coding.

[Dio]

I can say yes in specific circumstances. The actual pro-ideology or anti-ideology and associated social value itself does not directly influence specific behaviors beyond social model selection and reactions towards running a social model in the same catergory as his or her prefered model. The limitations on this involve the fact that the actual values from specific social areas influence behavior. For example, the fact that the Spartans prefer MORALE does not specifically influence the behavior of this faction. However, the increase in inclinations to run social models that increase the MORALE score of this faction can influence behavior. This further means, for example, that the Morale bonuses from +4 MORALE that the Spartans receive from running Power (the prefered social model of this faction) does affect AI behavior because the Morale level of the units in a faction influences the AI decision making process for a variety of decisions that include a request to stop a vendetta against an ally or to betray an ally. In short, the direct decision to select a certain social model or social area does not have an effect on the AI decision making process. The decisions that stem from this increase in tendencies for certain social areas does influence AI behavior.

Interesting; so its the overall social modifier compositions that result from the social policies that result in specific behaviours? Like if the policy choices they made results in 4+ morale and some support/industry bonuses but a lack of research perhaps they are more likely to be aggressive? Or if they have lacking morale and have more commerce related stuff they'll be more amicable towards trade?

How would one then aim to create specific personalities and patterns by this method? I am intensely interested in this in creating unique AI without extensive exe coding.

Yes. The most prominent examples I can think of right now appear with MORALE modifiers, SUPPORT modifiers, and ECONOMY modifiers (especially postive).
The problem with unique "AI" code involves the fact that the various sections involve independent operators that pertain with specific file names. It remains possible, but extremely difficult, to create additional spaces, and the most logical method without extensive .exe coding to create different behaviors remains to look for different combinations of AI mandates (EXPLORE, DISCOVER, BUILD, CONQUER), FIGHT values (Pacifist, Erratic, and Aggressive), and Faction bonuses (including SOCIAL Bonuses) from the official factions.

[ComradeCrimson]

I see; this is most useful.

So say hypothetically what would I use to make the following:

A relatively science focused/economic AI?

A warmongering conqueror who attacks relentlessly and without mercy?

And an relatively peaceful expansionist who spreads far and wide?

What combos would you say to achieve these the best?

[Dio]

The first faction would depend upon the final purpose of the faction. The Economic faction would likely focus on wealth accumulation (i.e. Morgan) with the INTEREST, COMMERCE, ECONOMY, HURRY, and ENERGY bonuses. This faction could have a BUILD focus and a Fight value of 0 or -1 depending upon the aggressiveness of money spending (e.g. reduced hurry unit production costs, Mindcontrol, or loan accumulations). The Research focused faction would employ RESEARCH, Number (# of free labs per base each turn), SOCIAL Research, TECHCOST , TECH, and SHARETECH. This could allow the faction to have DISCOVER and Fight value of 0 (i.e. University). The best combination involves ECONOMY, RESEARCH #, SHARETECH, and INTEREST. This essentially combines the best of both without additional overconsumption in other areas.
The second faction I would likely make with CONQUER, DISCOVER for high technology, and a FIGHT value of 1. I would likely make the AI employ the Hive file name because they do not surrender or accept pacts as easily as other factions. I would give the faction postive INDUSTRY, SUPPORT, and/or MORALE modifiers for the additional unit production and better units.
The third faction I would definitely want postive GROWTH, TALENT, and EFFICIENCY modifiers with a EXPLORE mandate on and a Fight mandate of -1. I would think the Believer faction file name might work best because it makes the AI produce additional colony pods. The rest depends upon the other qualities of the faction.

[ComradeCrimson]

The first faction would depend upon the final purpose of the faction. The Economic faction would likely focus on wealth accumulation (i.e. Morgan) with the INTEREST, COMMERCE, ECONOMY, HURRY, and ENERGY bonuses. This faction could have a BUILD focus and a Fight value of 0 or -1 depending upon the aggressiveness of money spending (e.g. reduced hurry unit production costs, Mindcontrol, or loan accumulations). The Research focused faction would employ RESEARCH, Number (# of free labs per base each turn), SOCIAL Research, TECHCOST , TECH, and SHARETECH. This could allow the faction to have DISCOVER and Fight value of 0 (i.e. University). The best combination involves ECONOMY, RESEARCH #, SHARETECH, and INTEREST. This essentially combines the best of both without additional overconsumption in other areas.
The second faction I would likely make with CONQUER, DISCOVER for high technology, and a FIGHT value of 1. I would likely make the AI employ the Hive file name because they do not surrender or accept pacts as easily as other factions. I would give the faction postive INDUSTRY, SUPPORT, and/or MORALE modifiers for the additional unit production and better units.
The third faction I would definitely want postive GROWTH, TALENT, and EFFICIENCY modifiers with a EXPLORE mandate on and a Fight mandate of -1. I would think the Believer faction file name might work best because it makes the AI produce additional colony pods. The rest depends upon the other qualities of the faction.

Roger that. Thank you very much on the input; it should help me figure out some means to produce some faction personalities I am aiming for with the factions I'm making.

My only other question for now revolves around the question of; "What is the max limit of aggression, and how can I make a faction hyper aggressive?"

Reason being is I intend to make a relative monster faction in terms of how violent it is, very rarely making pacts or treaties and completely relying on warfare to further its aims. Would offensive bonus or Psi bonus to troops also affect AI behaviour?

[Dio]

The PSI faction bonus would make the AI produce more Alien Lifeforms than average (unlike PLANET Bonuses alone). The Alien Offense and Fanatic bonuses play a minimal role in the actual aggressiveness of a faction. The most prominent factors that I currently recognize in AI aggressiveness involves the Faction FIGHT level, Difficulty level of the game, and the past actions of a faction against each other. I also know that at the beginning of each new scenario, or map, that the game randomizes certain values that include the supposed "combat encounters" (I remain uncertain except that it increases by 1 for each combat encounter between two factions) between factions, the "wrongs" against each faction, and various other features that increase with Difficulty level. The maximum aggressiveness involves two factions that have "Fight to the death" on, have a FIGHT value of 1, the "intense rivalry" rule on, and transcend difficulty.

[ComradeCrimson]

The PSI faction bonus would make the AI produce more Alien Lifeforms than average (unlike PLANET Bonuses alone). The Alien Offense and Fanatic bonuses play a minimal role in the actual aggressiveness of a faction. The most prominent factors that I currently recognize in AI aggressiveness involves the Faction FIGHT level, Difficulty level of the game, and the past actions of a faction against each other. I also know that at the beginning of each new scenario, or map, that the game randomizes certain values that include the supposed "combat encounters" (I remain uncertain except that it increases by 1 for each combat encounter between two factions) between factions, the "wrongs" against each faction, and various other features that increase with Difficulty level. The maximum aggressiveness involves two factions that have "Fight to the death" on, have a FIGHT value of 1, the "intense rivalry" rule on, and transcend difficulty.

Hmm... not quite what I want with the Psi bonus thing given that the said faction its intended for is the Unity; which is an incredibly anti planet faction.

What is the max level for faction fight level? Is 1 the max or can it go higher?

[Dio]

The highest Fight remains 1. As far as I know, any value over 1 converts into a value of 1 for the purposes of equations.

[ComradeCrimson]

Alright. Thank you; I've learned a fair bit.

I will have to think of some sort of bonus to give them... what is the difference between morale social modifiers and the morale +modifiers in the faction editor, by the way? Do they essentially perform the same function?

[Dio]

The MORALE social modifiers perform essentially the same task as MORALE, # modifiers. The only differences between these two values exist in very specific and circumstantial cases. The MORALE # modifiers stack with other morale modifiers. For example, +4 SOCIAL MORALE and MORALE, 1 give a unit +4 base morale. The morale modifiers of Morale, Number do not receive halved bonuses from -2 or lower MORALE (although the morale penality of, say, -2 MORALE would cancel this bonus for a result of 0 morale modifiers and halved bonuses from military unit facilities).

[ComradeCrimson]

Alright, so if I gave a faction a non social morale modifier it would stack with whatever social policy that is chosen?

So say that if a faction had morale modifier 1, and then chose social model Power, Thought Control and maybe Fundamentalist they would get maximum morale?