Sins of a Solar Empire : Real-Time Strategy. Unrivalled Scale.
© 2003-2016 Ironclad Games Corporation Vancouver, BC. All rights reserved.
© 2006-2016 Stardock Entertainment

Artificial Unintelligence (Mod v0.62 + AI info)

Improving the AI one parameter at a time

By on January 9, 2014 8:04:31 AM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Delnar_Ersike

Join Date 04/2013
+20

Let's face it: we all know that Sins of a Solar Empire's AI is stupid. Unfortunately, their most fundamental parts are locked into an exe or dll file, their source code most likely never to be released for modders to examine, alter, and fix. However, there are still ways to push the AI, however dumb it may be, to still put up a fight without resorting to ludicrous resource or stat bonuses. This post contains the following: my attempt at an AI-improvement mod for vanilla Sins of a Solar Empire, observations for players and modders about various in-game AI settings, and observations for modders about various moddable settings and a trick or two that influence how the AI behaves.

Even though my notes may seem thorough, both the research and the mod are very much a work-in-progress: as I continue to expand and double-check my notes, others will undoubtedly also start reporting their own observations, which I will try to aggregate and verify as much as possible. If I am missing something or you think a certain observation/conjecture is incorrect, let me know and I'll see what I can do.

Quick links:

Section 1: The Mod
Section 2: In-Game AI Info
General Info Presets Factions
Section 3: Modding the AI
Missions gameplay.constants Abilities
Planet Modules and Ships Fleets and Army Management Production and Resources
Research
 
Word of Warning: There is a lot of text. If reading white font on a dark background tires your eyes, just copy-paste the thing into Notepad or a word document. You won't get fancy formatting, but at least your eyes won't bleed.



Section 1: Artificial Unintelligence (mod v0.62 for Rebellion v1.82)

This mod takes advantage of the settings and tricks mentioned in Section 3 to make the vanilla Sins of a Solar Empire AI more competent; every single parameter mentioned has been altered, and every single trick mentioned has been used to some extent. The mod also has two gameplay-altering addons that help the AI (they are optional because they actually change gameplay) as well as a standalone addon that changes the way it behaves:
The Quick Start modification addon adds a Culture Center to the Quick Start buildings. This was added to make it harder to quickly steal planets near an enemy capitol early in the game, something that the AI is very vulnerable to, and also because AIs will start building culture buildings much quicker if an opponent has already built one, so the Quick Start modification also enables them to get into the culture game faster. Ideally, I would have made the Home Planet improvement add flat culture in addition to its regular bonuses, but I have not found a way to do this, so for now, the Culture Center Quick Start is my best solution.
The Kostura modification addon makes Kostura Prototype (the research that unlocks the Kostura Cannon) require 2 levels of NME Warheads. This not only brings the Kostura more in line with TEC and Advent superweapons in terms of how much research is required to unlock it (eg. you can't just build a ton of military labs and rush it, only TEC Loyalists can do that now), but also makes the AI more likely to research phase missile upgrades, since it is hardcoded to prioritize techs that eventually unlock new ships or buildings.
The Rabid AI addon forces the AI to use some fallback, faction-generic routines for vanilla Sins factions instead of the usual, faction-specific ones. It is called "rabid" because these routines make the AI both a lot more aggressive and a lot more illogical, as well as introducing bugs into its handling of capital ships. Ship production, tactical structure production, and fleet management are all governed by fallback routines, while faction-specific routines still govern all other aspects (eg. research, starbases, logistical structures). This addon functions independently of other parts of the mod, including the main mod. Enable this addon at your own peril.

Sins Artificial Unintelligence v0.62 [for Rebellion v1.82] can be downloaded here (mirror).

Changes since the last version (v0.60):

  • Increased possible pirate upgrades per raid from 2 to 4; this allow pirates to make better use of high bounties and maps with multiple pirate bases.
  • The AI will now only wait a minimum of 1 minute (from 5) before offering a player (human or AI) new quests or new treaties/pacts. This allows AI players to team up with each other much more quickly and not be forced to wait forever to make pacts with each other.
  • With the exception of disciple vessels, all factions' antifighter frigates, long-range frigates, carriers, heavy frigates, and light frigates have been given the AntiMine frigateRoleType due to the AI not knowing how to use anti-mine abilities, but still building AntiMine frigates when going up against a large amount of mines. Disciple vessels were skipped because their transfer antimatter ability's autocast targetting depends on other ships not sharing frigateRoleTypes with the disciple vessel.

Known issues:

  • For the Kostura modification addon, the arrow in the Vasari Warfare tech tree indicating that Kostura Prototype requires NME Warheads is bugged and points to a blank spot instead of extending to the Kostura Prototype button.
  • The AI-only ability used to force capital ships to retreat is still visible as "Explore" under capital ships' tooltips and under the level up menu.


 

Section 2: In-Game AI Info

In the absence of any official information on how the AI's operate, all info had to be acquired through weeks of running test games in various configurations.

 

General:

  • Quick Start:
    If you are playing solo, it is recommended you enable Quick Start. The main reason is that the AI almost never upgrades their capitol population immediately, so they start at a huge disadvantage income-wise. Since Quick Start starts everyone off with maxed population at their capitol, this helps the AI a metric ton in singleplayer games without giving it an unfair advantage. Note that playing on a higher difficulty level with Quick Start disabled is also viable, since the AI's income advantage earlygame is mitigated by it delaying its capitol population upgrade.
  • Pirates:
    The AI's handling of pirates are a mixed big. A single pirate raid can be a death sentence for a Normal Economist or Researcher AI earlygame, but they are not much to Aggressor AIs. Lategame, the AI's will never give the Pirate faction missions, even though they have the necessary research unlocked, resorting to still do bounty races. However, pirates can spice the game up a lot, and they are especially handy for leveling your capital ships. I tend to turn pirates off in my games, though you might still consider keeping them turned on.
  • Victory Conditions:
    The AI is often oblivious to victory conditions. Left to their own devices, AIs ignore occupation planets almost entirely (save for scouting). Though AIs recognize when flagship victory is enabled, they are incredibly stupid at handling flagships: they will treat enemy flagships as top priority ships even if their army wouldn't be able to defeat it (leading to them sacrificing fleets), while they are incredibly cowardly with their own flagships and will never actually take advantage of its offensive capabilities. The AI will never actively pursue a diplomatic victory, but they can still inadvertently win one. The AI will occasionally start off a research victory, but if they do, the other AIs will not respond, happily leaving their opponent to win a research victory; they will behave in a similar fashion towards human players progressing towards a research victory. Capitol victory is the only special victory condition that the AI truly understands: it will place a huge emphasis on eradicating opponents' capitols and will immediately rush their fleets to defend their own capitol if threatened. If Allied Victory is turned off, AIs will be just as inclined to form alliances; the only difference in their behavior is that they will automatically unally a player if the player is nearing a non-research victory or starting a research victory, and if an AI has no opponents, they will automatically unally the player with whom they have the lowest standing.
  • AI Advantages:
    The AI will always enjoy some advantages over human players, even at lower difficulty levels. They can start researching techs even before they have labs built, as they only need to initiate the construction to start the research; however, the effects of techs researched in this way will only trigger once the labs are complete (eg. if an AI queues up a military lab at the beginning of the game, it can have the first damage upgrade techs researched within the first 30 seconds of the game, but the actual damage increase only takes place once the first military lab is built). They can also see all phase jumping ships in the game even without the phase jump detection researches, though their algorithms only ever pay attention to jumps up to 2 planets away (like the maxed TEC or Advent phase detection techs). The AI can send and receive both missions and resources without unlocking the necessary techs. Finally, the AI can buy upgrades for its capital ships past the upgrade limit of level 3 (4 with tech), though they will only do so once all their capital ships have reached the upgrade limit (thanks Lavo_2!); the amount of levels they will buy depends on their difficulty level, with Vicious AIs buying up to level 7.
  • Unlimited Starbase Capacity:
    With enough resources, the AI can go way beyond starbases' module limit. AIs spend resources on starbase upgrades in spurts: they will queue up all the upgrades they want and can afford in an instant, just like how they increase bounties in spending spurts. The problem is the game only checks if a starbase has maxed its upgrade capacity after the AI has done its spurt, meaning that if the AI has enough resources, it can queue up more upgrades than the starbase's capacity would allow. As a result, it's quite common for AI starbases to have 10-14 upgrades.
  • Multi-star Maps:
    The AI is usually terrible at multi-star maps. This is mainly because when the AI determines targets for its missions (see Missions subsection of Section 3), map distance is the only factor it will consider; other factors, such as fleet comparison or diplomatic missions, are only used to alter existing missions' priorities. Since planets in other solar systems are much further away than those in an AI's home system, the AI will almost always only attack or colonize extrasolar planets once it has completely conquered its home system. Even when the AI does expand into other systems, it will prioritize poorer systems that are close to its home system over richer systems that are further away. If you do plan on playing on a multi-star map against AI opponents, pick maps where each AI begins in their own system (eg. Voruk's Labyrinth) over ones where systems are occupied by multiple players. Note that if you are a modder, you can make the AI slightly more competent at multi-star maps by making interstellar travel time instantaneous: AIs are very reactionary and might often need to redirect fleets from their original goals. Since unmodded interstellar travel can often take minutes, AIs are left very vulnerable when they commit their fleets to interstellar travel; making it all instantaneous removes this commitment, awarding the AIs much need flexibility for handling their fleets.
  • Jump-in, jump-out Looping:
    AIs often exhibit a loop whereby they will send parts of their fleet back and forth between an safe and an unsafe gravity well. This is caused by unsafe pathfinding. After retreating from a fight, the AI will give new move orders to the just-retreated fleets. Often it will tell those fleets to move to a target gravity well to which the shortest path requires moving through the gravity well the fleet had just retreated from. Instead of finding a safe way around or choosing a different target, the AI will blissfully send its fleets back into the danger zone, only to have them order a retreat the moment the fleet completes its jump... back to their starting gravity well, where the AI once again orders the fleet to move through the enemy-filled gravity well. This will repeat ad infinitum until the AI chooses a different target planet or the enemy fleet moves to a different gravity well.


Behavior Presets:

Keep in mind that these descriptions only fit well on lower difficulty levels: at higher difficulty levels, the fact that one preset is better or worse at building their economy is made pointless by the fact that they're showered with resources anyway. When I mention "economy upgrading", I mean building, researching, or upgrading anything that increases income or decreases ship/module cost.

  • Aggressor:
    These AIs are situational. They expand rather quickly and spend most of their income on building ships; however, this means that they don't tend to upgrade their planets or their economy, so even though they focus on building ships, they tend to lag behind AI presets that do focus a bit more on economy upgrading. They are actually worse than Economist AIs with the vanilla Gameplay.constants settings, but with a few tweaked settings, the situation improves greatly, though they still lose to the Fortifier preset 80% of the time. I would only recommend using them on tiny or small maps (or medium maps with pirates turned on) where their quick expansion and larger earlygame fleet size is the most important. The disadvantage of not having a properly upgraded economy also slowly fades away as you increase the difficulty level, so they are just as viable on larger maps at Cruel or Vicious.
  • Fortifier:
    These AIs are your Aggressors that are still capable late-game. Contrary to their name, they are actually quite aggressive with their expansion (they expand just as fast as the Aggressor preset) and don't focus nearly as much on tactical structures as the Aggressor preset, and although their planets are more easily taken earlygame since they don't concentrate as much on building ships either, they are much more able in the lategame due to their economic upgrades, and are much harder to defeat due to their use of upgraded Starbases (other AI presets don't tend to upgrade Starbases nearly as much as the Fortifier preset). They are more research-focused than Aggressors, and although they are generally less research-focused than the Researcher preset, the economic advantage that comes with early expansion can get them to end up having more techs unlocked than the Researcher preset. In short, they are like a hybrid between the Aggressor and the Researcher presets, with an extra emphasis on starbases and mines. I would recommend using them on medium or larger maps.
  • Researcher:
    These AIs are useless without any modding, but with modding, they are much more powerful and behave like your typical late-game Boomers. They are not as fast with their expansion as Aggressor or Fortifier presets, but still a bit faster than the Economist preset. Unmodded, their fleet sizes are laughable and they are easily taken out by other presets, but with some modification, they can hold their own against other AIs earlygame while they build up their prowess. They make up for their lack of units by concentrating a lot on research: it is not uncommon for Researcher AIs to have twice the amount of total researches as Aggressor or Economist presets in the first 15-90 minutes of the game (depending on map size), especially if they're Advent. In some cases, this advantage makes them snowball when factions start fielding larger fleets, and at normal difficulty levels, they are often the only AI preset that can stand toe-to-toe with a human player's offense. Their research and their upgrades will push their economy to levels where they will usually have the largest fleets out of all the AI presets after the first hour or so. However, their slow expansion makes them vulnerable to a containment strategy on smaller maps, especially without any modifications to Gameplay.constants, so once again, I would recommend using them on medium or larger maps.
  • Economist:
    These AIs are OK without any modding, but they improve the least from modding, causing them to perform the worst out of all four presets after they have all been tweaked. They are generally on par with the Aggressor AI in terms of fleet size, but lack in economy compared to Fortifier and lack in research compared to Researcher. The "Economist" part of their name only applies to their earlygame income, which they spend on ships anyway. In all other aspects, other presets tend to have a much better economy. Economists tend to either be very passive and defensive or overly aggressive to the point of sending their earlygame fleet to harass an enemy's capital instead of taking neutral planets. They expand incredibly slowly and spend most of their income on building ships and tactical structures; if the Researcher preset has been modified to not be such a wuss, Economists tend to be the first to get wiped out in AI FFA's due to their ridiculously slow expansion and suicide mission antics. I recommend you only use this preset sparingly.


Factions:

The AI has a lot of faction-specific behaviors associated with it; not only do AIs of the three/six vanilla factions behave differently based on the factions themselves, but they follow different routines as well. However, if the AI cannot determine which faction it is controlling (eg. as a result of controlling a non-vanilla faction or extensive modding of a vanilla faction), it will instead resort to fallback, faction-generic routines that make it completely differently, to the point of being considered its own "faction" entirely. Note that the AI will not always switch 100% to fallback routines: in some cases, only certain fallback routines will be used, while faction-specific ones will still govern other parts of the AI.

  • TEC Loyalist:
    AIs are a beast when playing this faction. Not only do they have the economy advantage inherit from being TEC, but they also take advantage of it much more quickly than TEC Rebels (eg. building Trade Ports, conquering neutral planets), since they don't get distracted as easily. Though they don't attack as often as TEC Rebels, when they do, they attack with a massively greater force, especially due to them spending more of their income on bolstering their economy. Their bonuses are also used a lot easier by the AI: it knows how to rush Novalith Cannons and research survivability-boosting techs (I haven't really seen multiple starbases though), and since the AI is very reactionary in its behavior, it will take advantage of the "if enemies in gravity well" bonuses as well. In fact, their only real downside compared to TEC Rebels is that because their cannons don't need missile tech, they won't upgrade missile tech as much in the unmodded game as TEC Rebels. TEC Loyalists is the best AI faction in vanilla Sins, period. I recommend using them with the Fortifier or Researcher presets, though they are so good at bolstering their economy that even the Aggressor and Economist presets are viable.
  • TEC Rebel:
    AIs playing as this faction are usually incredibly aggressive, sometimes overly so: when presented with the opportunity, they are much more likely to start harassing enemy planets, even if they have no siege ships, than to conquer neutral planets. If they start near an opponent, they are much less likely to bolster their economy as TEC Loyalists, plus they don't take advantage of their specialties nearly as much, either (since most, like Truce Among Rogues and Pirate Mercenaries, require tactics that aren't implemented into the AI). I do not recommend you use them on smaller maps, since they are much more likely to send suicide missions to other players' capitols instead of capturing neutral planets while there are still neutral planets to capture. On larger maps, they can get quite formidable, especially if they don't suicide their entire fleet halfway through the game; TEC Loyalists are still better overall though. If you do use TEC Rebels, I recommend the Aggressor or Researcher presets.
  • Advent Loyalist:
    The AI suffers from the fact that it cannot recognize how Advent units' abilities work together, so it will play Advent significantly worse than any competent human player.  This goes both ways however, as enemy AIs are also unable to recognize which ships should be killed first. Unmodded, Advent AIs also field poor unit composition, delaying Illuminator deployment and spamming Solarises instead of carriers with bombers. Despite all this, as long as the Advent AI plays against only other AIs, it should be fine. AIs playing as Loyalists, due to the faction's reliance on lategame techs for their economy, are often handicapped even more unless they are the Researcher preset; the lack of unlockables tied to economic techs is also a huge problem in unmodded Sins, since this means the AI will tend to research useless techs instead of economy-improving ones. If you do want to use Advent Loyalists, I recommend the Fortifier or Researcher presets due to the importance of lategame techs, but both Aggressor and Economist presets are viable on smaller maps due to Disciple Vessels' cost-efficiency and said presets' tendency to spam them.
  • Advent Rebel:
    The AI suffers from the fact that it cannot recognize how Advent units' abilities work together, so it will play Advent significantly worse than any competent human player. This goes both ways however, as enemy AIs are also unable to recognize which ships should be killed first. Unmodded, Advent AIs also field poor unit composition, delaying Illuminator deployment and spamming Solarises instead of carriers with bombers. Despite all this, as long as the Advent AI plays against only other AIs, it should be fine. I tend to find that Advent Rebels perform much better than Advent Loyalists, sometimes better than even TEC Rebels or Vasari Rebels, due to their unique techs being both mostly unlockable earlygame, requiring a lot less specialized tactics than the Advent Loyalist techs, and synergizing quite nicely with the whole Disciple-vessel-spam tactic Advent AIs usually employ at the beginning of the game. The fact that said techs generally help Advent players' economies earlygame adds to this advantage on lower difficulty levels. I tend to find that for Advent Rebels, the Fortifier preset concentrates the most on overall research while still properly expanding and maintaining a fleet (the Economist preset concentrates on Civilian Research, while the Researcher preset concentrates on Military Research, but both are eventually beaten in their focused categories by the Fortifier preset), so I would highly recommend using the Fortifier preset.
  • Vasari Loyalist:
    This faction comes in a close second to TEC Loyalists in terms of AI performance, and for similar reasons. Except for the whole mobile labs/planet stripping thing, none of their unique techs really require any "specialized" tactics to use; they even provide a tiny economic benefit by getting credit income from capital ships. This is in addition to how powerful Kosturas are compared to other superweapons and how Vasari don't need to send out colony ships to capture neutral extractors (it's not that this ability is so powerful, it's that the AI knows what to do with scouts more than it does with colony ships); since the AI tends to build colony ships instead of getting a colony capital ship, the fact that Vasari use colony ships for starbase construction only makes things easier for Vasari AI compared to other factions' AIs. However, in longer games, Vasari Loyalist AIs lose out to AIs from other factions due to not upgrading their economy well enough to maintain proper fleets and Vasari ships in general being the least economically efficient. It should be noted that Vasari AIs also do not know how to use starbases for offensive purposes, giving them a fairly large handicap against human players. Similarly to TEC Loyalists, Vasari Loyalist AI is good with any preset, though it is still generally better with the Fortifier and Researcher presets.
  • Vasari Rebel:
    AIs playing as this faction are either as capable as Vasari Loyalist AIs or worse than TEC Rebel AIs, depending on whether they have allies or not. Though they don't know how to use Starbase mobilization, arguably the most overpowered tech in the entire game, they know how to use all their other special techs, especially since AIs are usually quite eager to join allies in battle, and they move between phase lanes quite a lot. Of course, Vasari Rebel AIs still have all the advantages of being Vasari for AI (Kosturas, scouts capture extractors, one tech improves both metal and crystal extraction, etc.). Without allies though, a lot of the special techs that the AI knows how to use are worthless, hence why they perform worse than TEC Rebels in those cases. It should be noted that Vasari AIs also do not know how to use starbases for offensive purposes, giving them a fairly large handicap against human players. I recommend using the Fortifier or Researcher presets for reasons similar to Advent Rebels.
  • Fallback (faction-generic) AI:
    Fallback routines result in an odd AI behavior; the best word I can use to describe it is "rabid". The routines make the AI both extremely aggressive and incredibly irrational. Though fallback AIs still follow the behavior of their presets somewhat, they usually concentrate more on building ships and attacking enemy planets. The extent of "rabidness" depends on how much different parts of a faction were modded: for example, if the AI cannot find its faction-specific tactical structure, it will resort to fallback routines for ship production, tactical structure production, and fleet management, but it may still use faction-specific routines for things like research and starbases. When the fallback AI controls fleet management, attacks on enemy capitols within the first 10 minutes of a game are not uncommon. Fallback AIs maintain much fewer scouts than faction-specific AIs and are very buggy when handling capital ships: in the best case, the AI will usually idle capital ships around neutral planets until they are colonized, but in the worst case, capital ships might stand idly in the AI's capitol until they die (reported by GoaFan77).


 

Section 3: Modding AI Info

Since I do not have access to the AI's source code, all statements are based either on in-game observations from tweaking various values and/or on experimentally verifiable observations made by other players and modders.

As a modder trying to get into the AI's head, the Player Info menu ([d] inside the dev menu) is your friend: the four items [a] Show Mission Queue, [b] Show Fleet Comparison, [e] Show Spending, and [f] Show Build Lists are invaluable in relaying what the AI is thinking at any given time.

The AI operates using AI ticks: all AI orders are given out at regular time intervals. To clarify, this means all orders, from selecting techs to retreating fleets, are all given out at the same time, and the AI will wait a few seconds afterwards before giving out its next batch of orders. AI ticks happen roughly once every 10 seconds (9 to 11 income ticks), but they are out of sync between different AI players. To my knowledge, the time interval cannot be modded. 


Missions:

The AI gives orders to its ships based on missions the AI creates for itself to complete: it will populate its mission queue with missions, then produce and assign ships based on its current mission queue, rated by priority. Modders can view an AI's mission queue via [d] Player Info > [a] Show Mission Queue from within the dev menu, then selecting a unit, module, or planet belonging to the AI in question. Scout frigateRoleTypes will be sent on Explore missions with targets on enemy, neutral, or undiscovered gravity wells, Colony frigateRoleTypes will be sent on Colonize missions to nearby neutral, colonizable planets, combat ships might be sent on ClearOrbitForColonize missions to nearby colonizable, neutral planets with hostiles in orbit, etc. Note that the AI will never remove missions, only alter their priority; this is why AIs tend to be more and more schizophrenic the longer a game lasts, as their mission queue expands to sizes it can no longer handle properly. The only way for a mission to be removed from an AI's mission queue is if it is completed or if its objective is no longer valid (eg. attack a planet that is now allied, clear orbit for colonization of a planet that was just colonized by an enemy).

Mission .entity Files:
Some missions have their own .entity files, allowing modders a limited amount of control over them. Each mission has its own entityType, but all mission entityTypes have only one parameter, missionType. For the game to recognize a mission, its name must match its entityType, so multiple entities corresponding to the same mission are not possible. The missionType parameter is used to direct the AI to the appropriate algorithms. The framework means that, in theory, modders could redirect AI missions. For example, having the MissionAttackPlanet entity have a missionType of Explore would redirect the AI to its Explore routine whenever it wants to attack a planet. In practice however, the different mission types use different internal arguments, so redirects often cause the AI to lock up in the best case (eg. when redirecting MissionAttackPlanet to MissionFosterRelations) or the game to minidump in the worst case (eg. when redirecting MissionBuildMines to MissionBuildStarbase).

 

Gameplay.constants:

  • aiRetreatThreshold:
    Represents how quickly the AI will retreat: a lower value makes it retreat less often and a higher value will make it retreat more often.
    More precisely, it is the ratio of total allied time-to-kill to total enemy time-to-kill below which the AI will order a full retreat. Zero or less will mean the AI never retreats, while values higher than one make the AI retreat even when it faces a weaker force.
    Actual time-to-kill calculations are highly flawed: for example, things not taken into account include hit chance, miss chance, regeneration, squadron movement between attacks, weapons being on separate banks (this is mainly why AIs overvalue capital ships and Titans), and allied tactical modules. Decreasing this value does make the AI more consistent and reckless, willing to keep fighting even when it would be better to just cut losses and regroup. Increasing the value makes the AI preserve its ships a lot more, but it also makes its attacks more inconsistent; it is also more likely to get into jump-in, jump-out loops. As side effects, increasing this value also slightly slows down AI expansion rates, but simultaneously increases AI Capital Ship longevity and overall fleet sizes. Lower values are generally better on smaller maps, while higher values are generally better on larger ones.
    The AI will never retreat individual units: retreats are always full retreats, hence why the AI tends to lose its capital ships so often. Possible ways to get around this include exploiting Explore-type abilities.
  • pirateRaidDef:
    All settings here control the way pirates work, from the amount and types of ships they get to how often they go out on raids. For the purposes of this guide, I will only highlight the two entries that control planet targeting and upgrades; all other parameters control fleet composition, fleet size, and various cooldown times.
    • maxNumUpgradesPerRaid:
      Pirates have techs, just like all factions. This parameter controls the maximum amount of techs the pirate AI will unlock between raids, including levels within techs. Techs still require credits to be spent though, and the pirate AI will spend credits to build a properly sized raid fleet before it starts spending on research. The default value for this parameter is 2, possibly causing a significant bottleneck when pirates gain access to a lot of credits early on (either through multiple pirate bases or through high bounties).
    • juicyPlanetModuleThreshold:
      The pirate AI has deterministic when choosing what planet to raid: from the target player's owned planets, it will choose the one that is the least amount of phase jumps from a pirate base. However, if there is more than one planet that is closest to a pirate base in phase jumps, the pirate AI chooses one randomly, but it will always choose "juicy" planets over non-"juicy" planets. A planet is deemed "juicy" if it has at least as many modules built (both tactical and logistical count, as do extractors) as the value set by juicyPlanetModuleThreshold.
  • PlayerAITable: (under playerAISharedDef)
    These are a list of four tables used by the four different AI presets for various stuff. All parameters except for MaxBountyBidCount work in a similar fashion.
    • All values except MaxBountyBidCount: (eg. BuildShip, BuildModuleTactical, etc.)
      Most likely update frequency values: the higher a parameter's value, the more often the AI checks if it can/should perform that action. This process is independent of actually giving out orders; even if the AI checks whether it wants to build ships twice every second, it will still only give orders to build ships every AI tick (roughly 10 seconds). The higher these values go, the less of a difference you will see between them. For example, there is little observable difference between 4 digit values, since they indicate time intervals faster than once every second.
      All parameters are rather self-evident in what they control (BuildModuleTactical controls building tactical structures, BuildModuleResearch controls building labs, UpgradePlanetInfrastructure controls building bombing defenses, etc.), but BuildShip only affects military Frigates, Corvettes, and Cruisers and does not affect capital ships or civilian ships. Values zero or less makes the AI never perform the given action.
      These values also affect the maximum size of AI build lists (see Black Market and Resource Allocation subsection); when a build list can contain items from different categories (eg. the build list for planet improvements), these values also affect item priorities within the build list.
    • MaxBountyBidCount:
      The amount of "bidding spurts" the AI is willing to do between an immanent pirate attack and a launched pirate attack; the AI will only spend on bounties during this time. Changing the value to a different non-zero value will not change the amount of money the AI is willing to spend on bounties, only the number of times it is willing to overtake the existing bounty leader. Setting the parameter to zero or less will have the AI never spend money on bounties. Since the AI only gives out orders every AI tick, setting MaxBountyBidCount too high is pointless, since the bidding window will close before the AI can take advantage of the large amount of times it could potentially bid.
  • PlayerDiplomacyAIDef
    This section is filled with various settings related to diplomacy and the AI. For the purposes of this guide, I will only highlight the entries that do not affect human players, only AI players.
    • breakRelationshipThreshold:
      This determines the value at which pacts or treaties will automatically be revoked; each pact/treaty has its own value. Though this affects both human and AI players alike, the reason I am mentioning this value is because the AI will never revoke a pact or treaty, only if it would automatically happen by passing below breakRelationshipThreshold.
    • aiDiplomaticPointBonusScalar:
      This gives a bonus to the AI's accumulated diplomatic points depending on the AI's difficulty level; the bonus at each level can be set manually. The regular amount of diplomatic victory points received is multiplied by (1 + bonus) to get the actual value visible in-game.
    • timeBetweenTreatyOffers:
      The amount of time, in seconds, an AI player will wait before offering a player a treaty or a pact after a previous offer. Note that in the case of AI-AI alliances, this is usually the cause of AIs taking so long to make pacts with one another even though they have the necessary techs unlocked and are at a high enough relationship level.
    • timeBetweenQuests:
      The amount of time, in seconds, an AI player will wait before offering a player a quest after a previous offer. Note that this is based on when the quest was offered, not when it was completed.
    • aggressivenessModifierMin, aggressivenessModifierMax, aggressivenessModifierIncrement:
      Whenever an AI player is spawned, it will randomly get assigned an aggressiveness modifier. The aggressiveness modifier will be added to all of the AI player's relations with other players and will show up in the UI as "diplomatic inclination". Aggressiveness modifiers can range from aggressivenessModifierMin to aggressivenessModifierMax in increments of aggressivenessModifierIncrement; remember that because the modifier is added to relationships, a lower modifier equals more aggressiveness. An AI's aggressiveness modifier is independent of its preset: an aggressor preset is just as likely to have the highest modifier possible as an economist preset having the lowest modifier possible.
    • aiBonusMax:
      AI players will get a small relationship bonus with other AI players; the bonus starts out at 0 and increases over time until it reaches the value set by this parameter. Rate depends on the value set by aiBonusRate.
    • aiBonusRate:
      AI players will get a small relationship bonus with other AI players; the bonus starts out at 0 and increases by the amount set by this parameter every AI tick (about 10 seconds) until it reaches the value set by aiBonusMax.


Abilities:

The AI will automatically enable autocast and use certain abilities on ships and buildings of certain roleTypes. For example, abilities will automatically have autocast enabled for CANNON planetModuleRoleTypes. Scout frigateRoleTypes will automatically be assigned to Explore missions (see Missions subsection), even if they lack an Explore-type ability. Colony frigateRoleTypes (and possibly capital ships with the COLONY roleType) will automatically use Colonize-type abilities for Colonize missions. Otherwise however, if an ability does not have autocast turned on by default, the AI will never use it.

  • aiUseTime:
    Controls when an ability is autocast. There are multiple values it can take, you can read up on them in Zombie's soaseplugin wiki here; the situation each value represents is self-evident from its name. "Invalid" is only used if the ability cannot be autocast, while "NoEnemyShipsInGravityWell" will occasionally get bugged and still trigger if the enemies are militia or civilian ships.
  • aiUseTargetCondition:
    Controls under what conditions the ability is autocast, provided the aiUseTime condition is also met. There are multiple values it can take, you can read up on them in Zombie's soaseplugin wiki here; the situation each value represents is self-evident from its name. "Invalid" is only used if the ability cannot be autocast and "IsDifferentRoleType" will choose a target that has no common *RoleType (eg. frigateRoleType, planetModuleRoleType) values with the caster entity. Keep in mind that as of now, there is no way to logically manipulate aiUseTargetCondition parameters (ie. use "and", "or", "not", etc.), so you must choose only one, pre-defined value and work with that.
  • isAutoCastAvailable:
    Enables or disables autocasting. If set to FALSE, autocasting is disabled, and both aiUseTime and aiUseTargetCondition must be set to "Invalid". By default, the game only really sets this to FALSE for passive abilities, even for abilities with bugged autocasting (eg. Advent's Resource Focus).
  • isAutoCastOnByDefault:
    Controls whether the ability has autocasting turned on by default. The only reason I am mentioning this is that the AI will automatically use and enable autocast for abilities on ships and buildings of certain roleTypes (eg. Cannon, Scout, Colonizer) regardless of whether this parameter is set to TRUE or FALSE; otherwise however, if an ability does not have autocast turned on by default, the AI will never use it (thanks GoaFan and Seleuceia!).
  • ignoreNonCombatShipsForAutoCastTarget:
    Controls whether the ability ignores civilian ships (Construction Frigates, Trade Ships, Refinery Ships, Colony Frigates, Envoys, and Starbase Constructors) for autocast targetting.
  • onlyAutoCastWhenDamageTakenExceedsPerc:
    Decimal parameter that controls when an ability is autocast depending on the hull damage the caster has taken as a fraction of its total hull. Only hull damage is taken into consideration. Values zero or less will disable this parameter, while values one or greater will make the ability never autocast (the caster will be destoyed before the ability is cast).

Leveling Capital Ships and Titans:
The AI will choose a random ability from its unlocked abilities that are at the lowest level, unlocking new abilities only if it cannot do anything else. Abilities with isUltimateAbility set to TRUE will be prioritized over those that are not. As a result, AI capital ships in the base game will always end with two random abilities and their ultimate maxed and will never have a non-ultimate ability more than one level higher than their other unlocked abilties. The same logic appears to govern the way Titans are leveled.

 

Planet Modules and Ships:

AIs strongly depend on frigateRoleType, planetModuleRoleType, roleType, and role values assigned to ships and planet modules. "Strength" calculations are never done when the AI is deciding what to build. Modifying these values will affect what ships/buildings the AI will create, also influencing its army composition. For example, setting the culture module's planetModuleRoleType to "REPAIRPLATFORM" should have the AI building it as if it were an orbital repair platform, while adding the frigateRoleType "AntiModule" to carriers will have the AI building carriers to try to counter starbases and static defense in addition to actual antimodule frigates like the Ogrev. Keep in mind though that altering these values may have other gameplay effects, notably on autocasting when the "IsDifferentRoleType" aiUseTargetCondition value is used; ships with the "AntiModule" frigateRoleType can only attack structures, so use it only sparingly, even if the AI does love it. For strike craft, the "role" value for the strike craft's squad is used, while the roleType value for capital ships is barely used outside of determining which capital ships can colonize.
Note that messing around with modules' planetModuleRoleType is a lot more risky than messing around with frigates' and capital ships' roleTypes: the game seems to be very particular about certain roleTypes, and will crash if modules of a certain planetModuleRoleType do not comply with its expectations. For example, structures with the WEAPONDEFENSE planetModuleRoleType must have a weapon, otherwise the game will crash when the AI wants to build one.

Frigates with Bombing Damage:
The AI handles these in an odd fashion: if a frigate can deal damage to enemy planets, the AI will order them to attack the planet upon entering the gravity well, even if they don't have the "Siege" frigateRoleType and/or there are high priority targets (ie. Colony Ships) nearby (thanks Seleuceia!). The game will crash if an AI is playing a faction with no frigates that have a particular frigateRoleType, such as "ResourceCollector" (thanks Seleuceia!); this is likely due to it searching for particular frigateRoleType ships when queuing possible missions (see Missions subsection).

Building Capital Ships:
The AI is quite straightforward with building capital ships: if it has no capital ships or its lowest level capital ship is level 3 or higher, it will queue up a random capital ship at its home planet (thanks Lavo_2!). This queue is not the same as the in-game build queue, and the capital ship type will randomly change every AI tick until it actually gets queued in-game. In-game stats have so far shown that capital ship choice is truly random, both for the first capital ship built and for any subsequent capital ships. Giving the AI the means to reach capital ship level 3 quickly, eg. by making the XP cost of the first two levels 0 via gameplay.constants, will force the AI to spam capital ships. Note that if the AI has no capital ship factories, it is hardcoded to prioritize a new capital ship factory in a similar manner; the implementation is a bit buggy though, see the Build List Bugs subsection under Overall Production and Resources.

Building Titans:
The AI is prioritized to research titan-unlocking techs; however, it will only actually start construction of a titan once it has enough capital ship slots. Since the AI only upgrades its capital ship slots if it currently has 0 slots, Titan construction is only possible once the AI has researched the first capital ship supply tech that gives 2 capital ship supply, or if it loses a capital ship while having the necessary techs unlocked. By default, this is capital ship supply tech 4, meaning the AI must either lose a capital ship after all titan-unlock techs have been researched or have at least 3 capital ships level 3 or higher before it will even consider titan construction. Once a titan is in a build list, it is hardcoded to have a higher priority than even capital ships, so gathering enough resources for it usually is not a problem for the AI.

 

Fleets and Army Management:

When managing armies, AIs make constant use of fleets. Though AIs can manage individual ships just as well as fleet groups, fleets will be an AI's smallest army element 99% of the time, so AIs will usually move to target planets as a fleet, retreat as a fleet, and move to locations within gravity wells as a fleet (attacking is a bit different). All AI fleets will adopt a cohesion range of All Standard, regardless of what is set as defaultCohesionRange in Fleet.entity. AI ships will organize into fleets regardless of whether or not autoJoinFleetDefault is set to TRUE or FALSE. However, AI ships that only have certain frigateRoleTypes will never make or join fleets, even if autoJoinFleetDefault is set to TRUE: these frigateRoleTypes are Scout, Colony, ResourceCapturer, ModuleConstructor, Cargo, Envoy, and StarBaseConstructor; Scout is a special case outlined two paragraphs down. Flagships are still included in fleet management processes, but the AI forces them out immediately afterwards, so they are practically exceptions.

Making and Rearranging Fleets:
AIs will fire off fleet creation and rearrange routines throughout the entire game. Most creation and rearrangement routines will only fire if any ship in the gravity well has no current order. When a non-fleet combat frigate (any frigate that is created for a frigateRoleType that isn't Colony, Scout, ResourceCapturer, ModuleConstructor, Cargo, Envoy, or StarBaseConstructor) with no current order is in a gravity well or phase jump region that has no fleets and at least one other combat ship, it will initiate a fleet creation function at the next AI tick (roughly once every 10 seconds): either the originator of the creation function or a higher ranking (Titan > highest level Capital Ship > lowest level Capital Ship > Frigates and Corvettes), non-fleet combat ship in the same gravity well (ships in phase space count as being part of the gravity well to which they are jumping) will create a new fleet and invite all other non-fleet combat ships to that fleet.

When a combat frigate enters a gravity well or phase jump region that already has at least one fleet, it will initiate a fleet rearrangement process on the next AI tick (roughly once every 10 seconds); every ship that takes part in the fleet rearrangement process (ie. has at least one frigateRoleType that is a combat frigateRoleType) will be part of a fleet after it is completed. All ships in the same gravity well assigned to the same mission (see top of Section 3) will be placed into the same fleet. Note that ships already in the same fleet can still be (re)assigned to separate missions. The AI will periodically check if ships in a fleet are still both in the same gravity well and on the same mission and if not, they will fire off a fleet rearrange. As a result, large fleets phase jumping to a gravity well will often be split up, since even though the ships are on the same mission and in the same fleet, the ships currently in phase space technically are not in the same gravity well as those that are preparing to jump; this triggers a fleet rearrange, splitting up the large fleet,

Scout frigateRoleType Issues:
Although the AI treats the Scout frigateRoleType as a non-combat frigateRoleType, the moment a ship with it enters into a fleet with one other ship, even if that other ship is also a scout, it will switch over to a combat frigateRoleType for that ship. Since the AI will not attack with ships that do not have auto-attack turned on, those scouts are dead weight for the rest of the game. The AI will also not build any new scouts, thinking its existing ones (that are dead weight) are enough. A small workaround is to have an explore ability with autocast turned on by default; this will force the AI to still try to keep scouting with its dead weight scouts. However, the AI still recognizes those ships as combat ships, so they will often try to jump back to a gravity well containing normal fleets. If aiUseTime is set to Always on the explore ability, the scout ships will waddle back and forth between trying to scout a gravity well and trying to assist an allied fleet; though this is still no solution, it is better than nothing, and there is little modders can ultimately do to fix the scout grouping bug.

Giving Orders:
The AI will only give fleet leaders explicit orders, making non-leader fleet member behavior fairly predictable; the sole exception is any frigate with planet bombing capabilities, who will always be ordered to bomb a planet. Unless the fleet is planning to jump to another gravity well, fleet members will move with the fleet leader at all times, matching the fleet leader's speed. This means that if a fleet member is faster than the fleet leader, it will move slower than usual. If the fleet leader is targeting an enemy that is not in the same fleet as a member's target, the member will move into formation with the leader (matching speed as well) until an enemy that is part of the fleet the leader is targeting enters the member's attack or autocast range; if the member cannot attack (eg. carriers) or doesn't have auto-attack turned on by default (eg. scouts), it will simply stay in formation with the fleet leader.

 

Overall Production and Resources:

To determine what it wants to spend resources on, the AI keeps track of six total spending queues, or "build lists": one for ships of all types, one for research, one for building research labs, one for building logistical structures that aren't labs, one for tactical structures, and one for planet upgrades of all types. These can be viewed at any time from the dev menu via [d] Player Info > [f] Show Build Lists. Build lists are where the AI queues up stuff it wants to buy but cannot. Every AI tick, the AI will choose a build list to focus on (highlighted in red), trying to buy whatever it can from the focused build queue. The build list to be focused on is determined by a combination of chance, AI preset, current spending allocation, current fleet stance, and the priority of the highest priority item in each build queue. Priority within build queues is largely dependent on hardcoding and AISharedDef values from gameplay.constants. The maximum size of each build queue is determined by AISharedDef values from gameplay.constants, but not in a 1:1 fashion (eg. a value of 3 for BuildModuleResearch will not force the AI to only have 3 items in its labs build list).

Resource Tracking:
Credits are the main currency for the AI: they are what the AI keeps track of, while Metal and Crystal are only considered to the extent that the AI can afford what it wants to do. For instance, while Credits spent on buying Metal and Crystal on the Black Market are kept track of in the AI's spending log, Metal and Crystal sold for Credits are not.

Resource Allocation:
Current resource allocation can be seen under the [e] Show Spending in the player info menu, while ideal resource allocation is driven by the AI's preset. An Aggressor AI is hardcoded to spend more Credits on ships, while Researcher AIs are hardcoded to spend more Credits on research. Spending Credits on the Black Market is tracked independently of everything else, so an AI that spends twice as many Credits on a tech due to having to buy Crystal from the Black Market will still only note down the tech's credit cost as "spending on research". The ideal allocations set by each preset don't appear to change over time, resulting in things like the Researcher preset researching every tech up to level 3 in the first few minutes of the game.

Things get really wonky with ships that build structures or have abilities that cost resources. Specifically, the AI will not keep track of Credits spent on abilities or structures built by ships instead of from planets: in particular, this applies to starbases and non-TEC mines (TEC mines are built from the tactical structures menu, so the AI logs its spending as spending on tactical structures). Spending on ships will always be logged down as such, even if the sole purpose of the ship being built is to construct something, eg. starbase constructors. Starbase upgrades, on the other hand, are kept track of in their own category with its own allocation expectation. This is why Aggressors tend to have a lot of unupgraded starbases (their hardcoding prioritizes spending on ships, so they will spend lots on starbase constructors, not log down the cost of building a starbase, then not upgrade those starbases because they are hardcoded to not spend as much on starbase upgrades). The AI also has no resource allocation for diplomacy: spending on envoys is recorded as ship spending, spending on unlocking diplomatic actions and abilities is recorded as research spending, and sending resources to other players to complete missions is not kept track of at all.

Black Market:
The AI is very liberal when using the Black Market. When considering what it can afford, the AI will always take the Black Market into account, even if it would result in spending three or four times as much on something. More often than not, the AI will decide what to do and try to find a way to afford it rather than choose something to build or research that it can already afford without splurging on the Black Market. This is why Aggressor Vasari AIs tend to spend mounds on buying metal (Vasari LFs have a fairly high metal cost for their build times and supply usage, so Aggressor AIs who keep spamming LFs will run out of metal real quickly, and their drive to build more ships results in their Metal buying sprees), while Researcher AIs, especially Advent, will tend to spend loads on Crystal.

AI Biases From Resource Cost:
Due to set-in-stone ideal resource allocations and the fact that only Credit spending is kept track of, the AI will prefer ships and structures that cost less credits but more metal and/or crystals over those that cost more credits. For example, if a ship costs 100 Credits and 1000 Metal, the AI will spend 4500 Credits or so on buying the Metal, build the ship, and log it all down as 100 Credits spent on ships and 4500 credits spent on buying Metal. If the AI is an Aggressor AI, it will want a rather large chunk of its spending in the ships category, so it will keep buying the ship until it thinks it has spent enough credits on ships, theoretically building about 45 more ships than normally desired. The extra credits spent on buying metal are logged as Black Market expenditures, which lower the percent of expenditures in all other categories, even the one in which the bought resources were spent (in the case of the example, this would be ships).

Extractor Build Order:
The AI will always build extractors in the same order, no matter their current income ratios. They will always build metal extractors first and crystal extractors second. When there are multiple asteroids available, the AI will alternate between building the two, but if possible, one metal extractor will always be built before a crystal extractor. This does not appear to apply to capturing neutral extractors (since autocast targeting is what controls AI behavior there).

Build List Bugs:
Normally, the AI is programmed to only place items on build lists that are available to it at the time. For example, it will not place frigates in its ship build list that it has not unlocked yet, it will not place social upgrades on its planet upgrade build list when it has already chosen industrial upgrades for all its current planets, etc. There are, however, a few exceptions: the AI does not check if research prerequisites are met before queuing up extractors, frigate factories, capital ship factories, military labs, and civilian labs. While most of these will simply be queued up and forgotten, capital ship factories with research prerequisites will actually cripple the AI until it finally builds one. This is because capital ship factories are given hardcoded priority, meaning the AI will barely build any non-labs logistical structures, including extractors, until it finally satisfies its lust for a capital ship factory.

 

Research:

AIs have a very set-in-stone method of researching techs, and there is little modders can do to alter the process. Though the AI will not consider a research's cost, it will not try to research anything that it cannot afford outside of the hardcoded stuff mentioned later and the Research Victory tech. Due to the AI's reliance on resource allocation when determining whether it should research a tech or not, AIs will usually research lower tier techs first. Observations indicate that when the AI wishes to research something, it will queue up all techs that are currently available to it in its Research build queue (see Production and Resources subsection), though never queuing up more than one level of a tech at a time. It will then treat its research build list just like any other: rearranging it every AI tick, giving it focus based on the highest item on the build list and its current resource allocation, etc.

Hardcoded Priorities:
The AI is hardcoded to greatly prioritize certain techs under certain conditions (they will always occupy the top of the research build list and will often cause the AI to shift focus to it). AIs are hardcoded to research every single tech and its prerequisites that unlock ships or planet modules, though it will only do so once it has enough labs to research all the necessary techs required: for example, this includes the first few levels of the extraction bonus techs for Vasari (unlocks Orbital Refinery), all the regular culture techs for Advent once 8 civilian labs are built (unlocks Deliverance Engine), and all missile weapon upgrades for TEC Rebels once 8 military labs are built (unlocks Novalith Cannon). However, since the AI can research techs before their labs complete, it may appear that the AI is unlocking techs even before it can research a prototype. Other techs included in AI hardcoding are as follows: supply cap and capital ship cap techs if the AI has maxed out the respective supply, techs that unlock the colonization of planet types the AI has discovered and wishes to colonize, wormhole travel unlocking techs if the AI has discovered wormholes, and interstellar travel unlocking techs if there are multiple star systems. All vanilla faction AIs in Sins are also hardcoded to research the relationship-improving techs from the Diplomacy tech tree. The AI is hardcoded to only ever research pacts if it has a ceasefire treaty with at least one other player, but said pact unlock techs get hardcoded priority when this is the case.

The priority parameter:
This parameter in a tech's .entity file dictates its priority in the research Build List: a tech with a priority of 2 will always be placed above techs with a priority of 1. Techs with a priority of 0 will never be researched by the AI unless they have hardcoded priorities. The game will accept both negative and decimal values for priority (possibly up to the limits of a double-type variable), but it will treat negative priority values as if they were 0. Techs with hardcoded priority are always placed at the top of the list irrespective of their priority parameter value.  The AI will rearrange its build list every AI tick, meaning that two techs with equal priority have an equally random chance of being researched before the other. Note that giving a single tech a higher priority value than others may effectively stop the AI from researching completely: since the AI will only ever compare the top items in each build list, and giving a single tech higher priority than all others will place it at the top of the research build list (under hardcoded techs), if the AI never wants to research mentioned single tech, it will never give focus to the research build list, stalling all research.

Dummy Research Ladder:
A while back, a coding trick for prioritizing certain techs was developed by a group of enterprising modders, including Lavo_2, GoaFan77, and Zombie. The trick takes advantage of the AI being hardcoded to prioritize techs that unlocks ships/buildings, being able to define 2 research prerequisites for each tech, and being able to place tech buttons in an off-UI area so that human players will never see them, but the AI can still research them. It involves making a ladder of 0 cost dummy techs, with the bottom of the ladder populated with actual techs (endpoints for each tech chain are enough) and the top of the ladder containing a tech that unlocks a dummy ship or building that is too expensive for the AI to ever actually make in-game. Since the AI is hardcoded to prioritize unlocks, it will try to unlock the dummy ship, even if the ship is actually too expensive to ever purchase. To do so, it needs to climb up the dummy research ladder, and to do that, it needs to complete every single research in the game. Though the trick requires a long, ardous setup time (you need to set up a different ladder for each tech tree, requiring about n-1 dummy techs for n total "real" techs), and the effects usually only kick in once the AI has enough labs to research the highest tier tech at the bottom of the ladder, the payoff is that the AI will highly prioritize certain techs, regardless of its preset. Modders who use this trick must choose between two options: either having a smaller ladder, which results in the AI greatly prioritizing certain key techs and relying on the usual process for all others, or a large ladder, which results in the AI prioritizing all techs regardless of preset, but losing focus as a result.




Once again, the contents of this post are subject to change as current observations are reevaluated and new ones are noted down. If you see something missing or not aligning with your own in-game observations, please let me know.

Locked Post 113 Replies +7
Search this post
Subscription Options


Reason for Karma (Optional)
Successfully updated karma reason!
January 12, 2014 1:26:30 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

This is awesome information, thank you very much I wish games would move towards a more programmable AI, but this is something at least. 

Reason for Karma (Optional)
Successfully updated karma reason!
January 12, 2014 1:48:49 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

That's extensive   I honestly thought that Economic was the best AI type (which may be the case unmodded, your research is more focused on the changes you made). Either way what you've put together is extremely thorough.

Reason for Karma (Optional)
Successfully updated karma reason!
January 12, 2014 2:04:36 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Very, very nice list here. Good information, in particular for newcomers to Sins modding.

One thing to note with the AI and research is that even if you do things such as making research 100% free for the AI, and putting in things that encourage it to research, does not prevent that often times, the AI flat out stops researching. The timing is variable, ranging anywhere from 1 to 2.5 hours into a game, but there hits a point where the AI won't research anything, not even to unlock new abilities.

I would also like to add that the AI knows when strikecraft are on ships, frigates in particular, even if the ship does not have a Carrier roleType or FrigateCarrier statCountType. Not only does it know if a ship has strikecraft, but it absolutely adores frigates that have strikecraft.

For example, lets look at a ship with the roleType Heavy and statCountType FrigateHeavy. Usually, the AI does not care for this combo of Types. However, the moment you give that ship strikecraft, the AI will build large numbers of the ship. This effect, with that specific combo of Types, has been replicated in two mods, with the Empire's Vindicator in SoGE, and the UNSC's Athens carrier in SotP.

Reason for Karma (Optional)
Successfully updated karma reason!
January 12, 2014 2:42:49 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

It is important to keep in mind that the AI will automatically set all abilities to autocast

I don't believe this is the case...repulsion is an excellent example, I have never seen the AI use it in vanilla where autocast is not set on by default....if you mod repulsion to have autocast on by default, the AI will use the ability (albeit not intelligently)...

SB construction and ultimate abilities may be exceptions -- the AI may still use those even if autocast is not on, but for normal abilities I believe that autocast has to be on or the AI will not use it...

Might also want to add a few things about how the AI uses frigates....for example, IIRC the AI must have at least have one frigate in its arsenal with the colonize roletype, the resource capturer roletype, the siege roletype, and the antifighter roletype....also, any frigate that has planet bombing weapons will always immediately go to bomb out the planet regardless of what other weapons/roletypes it has (caps and titans obviously do not behave this way)...

This is an excellent thread, extremely useful for new people...

Reason for Karma (Optional)
Successfully updated karma reason!
January 12, 2014 3:44:19 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting Lavo_2,
One thing to note with the AI and research is that even if you do things such as making research 100% free for the AI, and putting in things that encourage it to research, does not prevent that often times, the AI flat out stops researching. The timing is variable, ranging anywhere from 1 to 2.5 hours into a game, but there hits a point where the AI won't research anything, not even to unlock new abilities.

I know we've talked about this before Lavo, but I've definitely noted several cases where the AI is able to research everything, even beyond the 2.5 hour time your claim. In all cases though this only occurred once an AI maxed out its fleet supply and managed to not take many loses for a period of time. Of course I only noticed this in AI versus AI tests, against human players it is doubtful the AI will manage to do so well so perhaps in practice this rarely happens.

Quoting Seleuceia,
I don't believe this is the case...repulsion is an excellent example, I have never seen the AI use it in vanilla where autocast is not set on by default....if you mod repulsion to have autocast on by default, the AI will use the ability (albeit not intelligently)...

This is my experience as well. I've never notice the AI to use abilities with autocast disabled by default with the exception of some of the ones Seleuceia is talking about (also mine building abilities an probably a few other that tend to involve resources being spent).

From what I've seen, the AI is terrible at multi-star maps. It is incredibly slow at expanding to non-home star systems and is even worse at protecting said expansion systems. This is mainly because it prefers to keep its main army in its home system until there are no more enemies left. I would recommend playing either on a single-system map or a map with a large amount of star systems (eg. Voruk's Labyrinth), since having the AI easily conquer its home system makes sure it will properly deploy forces to expansion systems.

I've definitely noticed this, however I'm wondering if this applies on those few premade maps where the multiple solar systems are connected by normal phase lanes.

Reason for Karma (Optional)
Successfully updated karma reason!
January 12, 2014 4:22:42 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting GoaFan77,



From what I've seen, the AI is terrible at multi-star maps. It is incredibly slow at expanding to non-home star systems and is even worse at protecting said expansion systems. This is mainly because it prefers to keep its main army in its home system until there are no more enemies left. I would recommend playing either on a single-system map or a map with a large amount of star systems (eg. Voruk's Labyrinth), since having the AI easily conquer its home system makes sure it will properly deploy forces to expansion systems.

I've definitely noticed this, however I'm wondering if this applies on those few premade maps where the multiple solar systems are connected by normal phase lanes.

I have made a map that is entirely connected via phase lanes despite having multiple stars and I can confirm the AI still researched the tech to allow them to jump between stars despite the fact all planets could be accessed by standard phase lanes.

Reason for Karma (Optional)
Successfully updated karma reason!
January 13, 2014 2:12:39 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

I'm interested to know what researches I can delete from the research list without the AI crashing. I tried purging everything but the crew/supply limits and planet unlocks but, yeah. My goal is to purge every single research I possibly can so that nothing affects stats in any possible way. I've manually changed a lot of researches, but there's still a bunch I've probably missed. It'd just be cleaner to gut them from the start.

Reason for Karma (Optional)
Successfully updated karma reason!
January 16, 2014 3:04:38 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting WOEaintME,
That's extensive   I honestly thought that Economic was the best AI type (which may be the case unmodded, your research is more focused on the changes you made). Either way what you've put together is extremely thorough.

I actually thought the same thing, too, until I started running my test games. Turns out that Aggressor and Fortifier have better economies because they colonize more, while Researcher has a better economy because it researches eco techs more. The other presets' drawbacks can be helped by altering check frequencies, but there is no way to really change Economist's slow-tech, slow-expand behavior.

Quoting Lavo_2,
One thing to note with the AI and research is that even if you do things such as making research 100% free for the AI, and putting in things that encourage it to research, does not prevent that often times, the AI flat out stops researching. The timing is variable, ranging anywhere from 1 to 2.5 hours into a game, but there hits a point where the AI won't research anything, not even to unlock new abilities.

Quoting GoaFan77,
I know we've talked about this before Lavo, but I've definitely noted several cases where the AI is able to research everything, even beyond the 2.5 hour time your claim. In all cases though this only occurred once an AI maxed out its fleet supply and managed to not take many loses for a period of time. Of course I only noticed this in AI versus AI tests, against human players it is doubtful the AI will manage to do so well so perhaps in practice this rarely happens.

I've noticed a top-out thing as well, but from what I've seen the AI might still research everything, especially with modding the check frequency for Research and using the Fortifier or Researcher presets. As I mentioned in the original post under the "priority" section in Research, I think the AI selects non-hardcoded techs to research by randomly selecting a researchable tech (the priority tag is the weighting during random selection), then checking whether it wants to research the randomly selected tech or not; how it decides if it wants to research a tech or not is unknown to me. The "top-out" behavior would mean that no matter what tech is selected, the AI will never want to research it.

Quoting Lavo_2,
I would also like to add that the AI knows when strikecraft are on ships, frigates in particular, even if the ship does not have a Carrier roleType or FrigateCarrier statCountType. Not only does it know if a ship has strikecraft, but it absolutely adores frigates that have strikecraft.

For example, lets look at a ship with the roleType Heavy and statCountType FrigateHeavy. Usually, the AI does not care for this combo of Types. However, the moment you give that ship strikecraft, the AI will build large numbers of the ship. This effect, with that specific combo of Types, has been replicated in two mods, with the Empire's Vindicator in SoGE, and the UNSC's Athens carrier in SotP.

Hmm, interesting, I did not know that. I'll start playing around with this later, possibly also doing things like removing all strike craft in tandem with adding strike craft to frigates to see what the AI actually looks for (ie. strike craft capabilities specifically, ships that can build more ships in general, or something else).

Quoting Seleuceia,
I don't believe this is the case...repulsion is an excellent example, I have never seen the AI use it in vanilla where autocast is not set on by default....if you mod repulsion to have autocast on by default, the AI will use the ability (albeit not intelligently)...

SB construction and ultimate abilities may be exceptions -- the AI may still use those even if autocast is not on, but for normal abilities I believe that autocast has to be on or the AI will not use it...

Quoting GoaFan77,
This is my experience as well. I've never notice the AI to use abilities with autocast disabled by default with the exception of some of the ones Seleuceia is talking about (also mine building abilities an probably a few other that tend to involve resources being spent).

Explore starts with Autocast disabled, but the AI still uses it. I've tested turning off Autocast by default for the colonization abilities, yet the AI still used them properly. In addition, I did a bit of testing with my DEFCON mod, which essentially lets you spam superweapons like there's no tomorrow, and not only did the AI use superweapon abilities even though autocast by default was turned off, it automatically turned autocasting on; however, when I changed the superweapons' planetModuleType to WEAPONDEFENSE, the AI never used the cannon ability, nor did it turn its autocasting on (caveat: the DEFCON mod makes the AI play like the generic faction AI, so this might not be the case for vanilla AIs). Looks like you are correct, but it appears the hardcoding is based on ships' and buildings' roleTypes, not the ability itself; I'll update the post nevertheless.

Quoting GoaFan77,
I've definitely noticed this, however I'm wondering if this applies on those few premade maps where the multiple solar systems are connected by normal phase lanes.

Quoting WOEaintME,
I have made a map that is entirely connected via phase lanes despite having multiple stars and I can confirm the AI still researched the tech to allow them to jump between stars despite the fact all planets could be accessed by standard phase lanes.

The fact that the AI researches the star jump tech only proves that it is hardcoded to research that tech in multi-star maps. However, now that I think about it, I've also rarely seen the AI use wormholes, which might offer a clue as to why the AI is so bad at multi-star maps: it's time/distance. Traveling from one star system to another takes much longer than individual phase jumps, ditto for wormholes. Connecting stars via phase lanes should therefore have little to no effect on the way the AI behaves in multi-star maps. Possible ways to test to confirm: making stars in a multi-star system extremely close to each other vs. further apart, drastically increasing interstellar phase jump speed, having a Vasari AI with Phase Stabilizers built in two separate systems and see if it uses them. Since I'm not that good with the Galaxy Forge, I'll probably try my hand with the latter two (second one by modifying the first fleet research to increase interstellar jump speed, third one via messing around with the dev console).

Quoting IskatuMesk,
I'm interested to know what researches I can delete from the research list without the AI crashing. I tried purging everything but the crew/supply limits and planet unlocks but, yeah. My goal is to purge every single research I possibly can so that nothing affects stats in any possible way. I've manually changed a lot of researches, but there's still a bunch I've probably missed. It'd just be cleaner to gut them from the start.

I don't see why you wouldn't be able to remove all research (besides fleet techs) from the game: the AI will still be functional, though it will probably default to the Generic Faction AI. As for what techs you can remove without having it revert to the Generic Faction AI, I haven't done any testing, though I guess you can safely remove techs that do not unlock new ships, abilities, or diplomacy options.

Quoting Seleuceia,
Might also want to add a few things about how the AI uses frigates....for example, IIRC the AI must have at least have one frigate in its arsenal with the colonize roletype, the resource capturer roletype, the siege roletype, and the antifighter roletype....also, any frigate that has planet bombing weapons will always immediately go to bomb out the planet regardless of what other weapons/roletypes it has (caps and titans obviously do not behave this way)...

Tested, the AI can function without any ResourceCapturer, Siege, or Antifighter frigateRoleTypes, but does not work properly without a Scout frigateRoleType (since it will only ever move non-scout ships to explored planets) or a Colony frigateRoleType unless it builds a colony capital ship. Naturally, if it has no frigates with a certain frigateRoleType, the AI will never make any frigates for a certain task (eg. capturing neutral extractors for no ResourceCapturer, harass-sieging undefended planets for no Siege, etc.). As for the planet bombing thing, I did not notice this before, but testing has confirmed what you said, so I'll update the post later.

 
 
Reason for Karma (Optional)
Successfully updated karma reason!
January 16, 2014 7:25:32 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting Delnar_Ersike,
Looks like you are correct, but it appears the hardcoding is based on ships' and buildings' roleTypes, not the ability itself; I'll update the post nevertheless.

This is very interesting, and quite useful...I'd prefer to have cannon abilities NOT on autocast by default, but I was always afraid the AI wouldn't use them...

Quoting Delnar_Ersike,
Tested, the AI can function without any ResourceCapturer, Siege, or Antifighter frigateRoleTypes, but does not work properly without a Scout frigateRoleType (since it will only ever move non-scout ships to explored planets) or a Colony frigateRoleType unless it builds a colony capital ship

I took away the resource capturer roletype from the 3 ships when I started a planet mod...I didn't have any neutrals so I figured I might as well remove the roletypes...everytime I started a new game, I MDed about 20 or so seconds in....it happened again and again and again even though it worked just fine before I took away the roletypes...I added back the roletypes and the MDs went away...this, at least, has been my experience, and when I asked others about it I was told the same thing happened if you omitted the other 3 roletypes I mentioned...

 

Reason for Karma (Optional)
Successfully updated karma reason!
January 17, 2014 10:08:52 AM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting Seleuceia,
I took away the resource capturer roletype from the 3 ships when I started a planet mod...I didn't have any neutrals so I figured I might as well remove the roletypes...everytime I started a new game, I MDed about 20 or so seconds in....it happened again and again and again even though it worked just fine before I took away the roletypes...I added back the roletypes and the MDs went away...this, at least, has been my experience, and when I asked others about it I was told the same thing happened if you omitted the other 3 roletypes I mentioned...

Did another round of testing: lack of ResourceCapturer does indeed cause a crash, but the AI worked fine if I changed the frigateRoleTypes of all antifighter ships to "Light" (so none of the factions had a frigate with the AntiFighter frigateRoleType), same case with setting all siege types to "Light". Maybe it only crashed for the latter two because you had frigates with no frigateRoleType set and/or you tested this in Trinity, but the AI changed in Rebellion?

I'm also currently testing Lavo's strike craft statement: so far, adding strike craft to colony ships does not make the AI build more colony ships and adding strike craft to all Utility A and Utility B ships does not make the AI build them more. Still to test: what happens if strike craft are removed from carrier frigates, what happens if the frigateRoleType of carrier frigates is changed to "utility", what happens if strike craft are added to SB constructors, what happens if strike craft are added to scout frigates.

Oh yes, and my previous suspicion about why the AI is bad at multi-star maps is correct: it is because of how far away other star systems are, so the AI is much more likely to attack a planet that it can move its entire fleet to in a minute than to colonize planets that are 5-10 minutes away. It's the similar case with wormholes: unless the planets surrounding a wormhole exit are extremely undefended, the AI is much more likely to send its fleet to a well-defended planet that is closer than through a wormhole to a group of less defended planets further away. After speeding up interstellar phase speeds by 100x the original speed, the AI essentially started treating other star systems as if they were part of the AI's home system, rapidly sending through fleets to colonize if they had planets close to the stars in their actual home system. I'll be updating the main post with my findings.

 
Reason for Karma (Optional)
Successfully updated karma reason!
January 17, 2014 2:43:09 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

With regards to your research into priority for research entities, is it possible to put in a priority level such that the AI will look at every tech item and not get into a loop of checking for the same tech that it thinks that it does not want? Alternatively, is there a list of techs that the AI is hardcoded (or effectively so) to ignore?

Quoting Delnar_Ersike,
so far, adding strike craft to colony ships does not make the AI build more colony ships and adding strike craft to all Utility A and Utility B ships does not make the AI build them more.

There's one flaw with this specific test; those statcounttypes are not primarily combat ships.

Quoting Delnar_Ersike,
Still to test: what happens if strike craft are removed from carrier frigates

This much I have tested; the AI builds less of them. Have a ship in SotP which had the StatCountType of FrigateCarrier, and the frigateRoleType of Heavy and Carrier. When it had a single strikecraft squadron, the AI loved it. Removed the strikecraft squadron, and the AI stopped building them as much as they used to.

Quoting Delnar_Ersike,
After speeding up interstellar phase speeds by 100x the original speed, the AI essentially started treating other star systems as if they were part of the AI's home system, rapidly sending through fleets to colonize if they had planets close to the stars in their actual home system. I'll be updating the main post with my findings.

This is extremely interesting. If this is reproducible, I will certainly increase interstellar phase jump speed to make the AI more competent in multi-star maps.

Reason for Karma (Optional)
Successfully updated karma reason!
January 17, 2014 4:07:29 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting Delnar_Ersike,
Maybe it only crashed for the latter two because you had frigates with no frigateRoleType set and/or you tested this in Trinity, but the AI changed in Rebellion?

I have only personally experienced the resource capturer MD, I was simply told about the other 3 roletypes in a discussion after rebellion was released...

Reason for Karma (Optional)
Successfully updated karma reason!
January 17, 2014 7:42:46 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting Lavo_2,
With regards to your research into priority for research entities, is it possible to put in a priority level such that the AI will look at every tech item and not get into a loop of checking for the same tech that it thinks that it does not want? Alternatively, is there a list of techs that the AI is hardcoded (or effectively so) to ignore?

As far as I know, there are no techs that the AI is hardcoded to ignore, and I have not found a way to influence the random tech selection process other than to increase the priority of techs the AI is more likely to favor (stat increases, for example). However, as mentioned in the priority section, setting a tech's priority to 0 effectively makes it so that the AI will never research said tech unless it is hardcoded to do so; for example, this is why the AI will never research phase detection techs (their priority value is set to 0) and why it will only ever research planet colonization access techs if it has found a nearby planet that requires colonization tech (the techs have a priority value of 0, but the AI is hardcoded to research colonization prerequisite techs when it wishes to colonize a planet). A couple more examples of techs set to 0 priority in unmodded Sins: fleet research techs, wormhole travel tech, population cap increase techs, ship/building unlock techs, interstellar travel tech.

Quoting Lavo_2,
There's one flaw with this specific test; those statcounttypes are not primarily combat ships.

I know, I know, hence why I said I was still testing. I also found out later that my test setups for the strike craft thing were biased towards earlygame units, so I'll need to run the utility frigate tests again. To be fair though, the AI does use Support frigateRoleTypes in combat (Hoshikos are quite common for TEC AIs, for example), so if strike craft always increases a frigate's value in the AI's eyes, I should see a noticeable increase in utility frigate production.

Quoting Lavo_2,
This is extremely interesting. If this is reproducible, I will certainly increase interstellar phase jump speed to make the AI more competent in multi-star maps.

The difference isn't night-and-day, it's more like night-and-dawn: with instantaneous interstellar travel, the AI treats planets in other solar systems roughly as if the two stars represented 3-4 neutral gravity wells (this is due to the time it takes to move around in stellar gravity wells). Therefore, the AI will only really start to colonize planets in other solar systems if it is willing to colonize planets that are at least 3 phase lanes + [number of phase lanes between a star and their nearest owned gravity well] away, which usually means that the AI will only start expanding into other solar systems if it owns at least one planet (the more, the better) located next to a star in its home system. It is also worth noting that with the Stellar Phenomena DLC, certain star types (eg. Black Holes) will greatly discourage the AI from expanding into other solar systems due to slower phase drive warmup. In one test game running 6 AIs in 3 home systems and 1 expansion system, two of the home systems ended up being Black Holes, and the superior AI from the home system with the regular star already had 3 planets colonized by the time either of the four AIs from the two black hole home systems started sending their own fleets through.

 
 
 
Reason for Karma (Optional)
Successfully updated karma reason!
January 17, 2014 10:19:36 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting Delnar_Ersike,
The difference isn't night-and-day, it's more like night-and-dawn: with instantaneous interstellar travel, the AI treats planets in other solar systems roughly as if the two stars represented 3-4 neutral gravity wells (this is due to the time it takes to move around in stellar gravity wells). Therefore, the AI will only really start to colonize planets in other solar systems if it is willing to colonize planets that are at least 3 phase lanes + [number of phase lanes between a star and their nearest owned gravity well] away, which usually means that the AI will only start expanding into other solar systems if it owns at least one planet (the more, the better) located next to a star in its home system. It is also worth noting that with the Stellar Phenomena DLC, certain star types (eg. Black Holes) will greatly discourage the AI from expanding into other solar systems due to slower phase drive warmup. In one test game running 6 AIs in 3 home systems and 1 expansion system, two of the home systems ended up being Black Holes, and the superior AI from the home system with the regular star already had 3 planets colonized by the time either of the four AIs from the two black hole home systems started sending their own fleets through.

You know, I was already toying with increasing the max speed of ships at stars via the star ability, since it takes forever to move through them. If the AI really uses real time to reach the destination (which I suppose might be more logical), that may help too. Though I suppose that depends on the AI realizing the ability will allow it to move faster...

Reason for Karma (Optional)
Successfully updated karma reason!
January 18, 2014 7:30:08 AM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting Delnar_Ersike,
Hmm, interesting, I did not know that. I'll start playing around with this later, possibly also doing things like removing all strike craft in tandem with adding strike craft to frigates to see what the AI actually looks for (ie. strike craft capabilities specifically, ships that can build more ships in general, or something else).

The SotYR mod uses a lot of strike craft... Scouts with fighters are not overly built. The Centauri faction does not have many fighters and then only mid game and we experienced some very strange AI behavior both for the Centauri and in how the other AI dealt with them in battle. The worst AI behavior noticed is not finishing off heavily damaged ships. Whole fleets of heavily damaged ships would be ignored while ships in good shape would be engaged with gusto. Not having fighters (strike craft in general) seems to deprive the AI of the ability to "clean up" a gravity well during a battle.

The AI also seems to prioritize ship speed, frigates, corvettes and capital ships that are noticeably faster tactically will be built more often.

Edit: I should probably also add, we use much heavier damage per shot than Vanilla Sins and this also impacts AI behavior.

Reason for Karma (Optional)
Successfully updated karma reason!
January 18, 2014 7:31:29 AM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting GoaFan77,
You know, I was already toying with increasing the max speed of ships at stars via the star ability, since it takes forever to move through them. If the AI really uses real time to reach the destination (which I suppose might be more logical), that may help too. Though I suppose that depends on the AI realizing the ability will allow it to move faster...

I still need to check if it's path time it checks or path distance. Test setups with instantaneous interstellar travel are promising on smaller maps with multiple stars, but on larger maps, the AI still pretty much ignores expansion systems. I also need to perform tests on the black hole setup to make sure it's reproducable.
EDIT: Yup, both the black hole thing and the faster instellar travel were red herrings. I still have yet to test path distance, though I suspect that is what is actually responsible for poor AI behavior: Voruk's Labyrinth is the key, where AIs will always go for the closest expansion system, even if ones further away have planets that could be accessed more quickly with instantaneous interstellar travel.

Re-tested utility frigates with strike craft, the AI still does not start making more of them. However, adding AntiModule and AntiFighter frigateRoleTypes to carriers has resulted in the AI almost doubling their production (reasoning: Bombers are AntiModule, Fighters are AntiFighter) and halving production of actual antifighter and antimodule frigates, while adding the Light frigateRoleType to Illuminators has resulted in the AI producing a lot more of them (with no gameplay effects, since Illuminators don't use AM, so Advent LF's wouldn't autocast on Illuminators anyway). Oh yes, and I also tested adding Light frigateRoleType to siege frigates and it resulted in the AI roughly tripling siege frigate production. It still seems like the AI only really takes frigateRoleType into account when producing units, though I still need more testing to confirm that strike craft are not taken into account.

I have also found an additional advantage the AI has: Starbase upgrade capacity. The AI performs starbase upgrades in spurts, as in it will instantly select all the upgrades it wants and can afford in a single "spurt". However, the game only checks starbases' upgrade capacity after AI spurts, so the AI will often have starbases with 10-14 upgrades instead of the maximum 8, since it is unhindered by upgrade capacity until the final spurt. I'll update the main post to mention this AI advantage, since it's actually pretty huge in my eyes. Also, although the AI can research techs even before a lab is completed, the tech effects only take place after the lab is complete (eg. the AI can complete Terran population cap increase in the first 30 seconds of the game if it queues up civilian labs, but the actual population cap increase will only happen when the first civilian lab completes).

 
Reason for Karma (Optional)
Successfully updated karma reason!
January 22, 2014 9:22:57 AM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Hi, thought I would post this here.

 

Is there any mod that greatly increases the ability of the AI?

 

Is there any reason why the devs don't make the AI substantially easier to Mod? I'm thinking the StarCraft AI tournament or Sorian AI for SupCom

 

Cheers

Reason for Karma (Optional)
Successfully updated karma reason!
January 22, 2014 5:35:55 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

While I haven't mentioned it in the past, the AI can train more levels than human players can. Even if you remove the ability for humans to train capital ships entirely, the AI will still be able to, and does, buy them.

Reason for Karma (Optional)
Successfully updated karma reason!
January 22, 2014 6:43:14 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting Lavo_2,
While I haven't mentioned it in the past, the AI can train more levels than human players can. Even if you remove the ability for humans to train capital ships entirely, the AI will still be able to, and does, buy them.

Noticed this as well...IIRC, it's one level per difficulty over hard...so, at vicious they can buy up to lvl 7...

Reason for Karma (Optional)
Successfully updated karma reason!
January 23, 2014 2:50:16 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting StormingKiwi,

Is there any mod that greatly increases the ability of the AI?

Is there any reason why the devs don't make the AI substantially easier to Mod? I'm thinking the StarCraft AI tournament or Sorian AI for SupCom

If you just want an AI-improving mod, you can always try mine, Artificial Unintelligence; you'll find it in Section 1 of the post. If it's still v0.4 when you're reading this, keep an eye out for an updated version I'll be uploading soon.

As for why the devs haven't made AI's more moddable, I assume it's a combination of the following factors: 1) Making an AI is hard. 2) Making a good AI is harder; most "better AI" mods have the same flaws as their unmodded counterparts, since they rely on running hard-set AI scripts, it's just that they had years of figuring out the best game strategies to make their scripts successful more often. 3) Making a modular AI is hardest, especially when you need to work with giant, complicated systems that aren't directly connected, and if you aren't a huge development studio. Even Firaxis, a considerably larger and older developer, had to revert back to hardcoding-extensive AI scripts for Civ5 due to how many more systems they had to get the AI to keep track of compared to previous titles. For example, in Civ4, upkeep was the only disadvantage to building too many cities at once (it was quite a major one, too), but upkeep is tied to finance income, something the AI was already handling, so it could base decisions on whether it should settle more cities simply by checking income figures; plus since it only really changes when you found a new city or move your capitol, something the AI would have direct control over, there is no need to calculate for potential future upkeep when placing cities in the present. In Civ5, placing cities affects your policy growth rate and your ability to build key national wonders, while the main city-spam preventer, global happiness, depends on your city population, something that can vary quite a lot with time. Not surprisingly, Civ5's AI is laughably stupid, even with the latest expansions, and oddly enough, Firaxis' way of simplifying the Civ formula is partly to blame.

Quoting Lavo_2,

While I haven't mentioned it in the past, the AI can train more levels than human players can. Even if you remove the ability for humans to train capital ships entirely, the AI will still be able to, and does, buy them.

Hmm, it hadn't occurred to me to check that, and I didn't notice it before due to running my test games at 8x speed. I'll look into it, thanks for the pointer.

In the meantime, I finished my strike craft testing: approximately 40 games run in total in various configurations of adding/removing strike craft from ships and seeing how the AI, running with different presets and factions, would react. Verdict: for production purposes, the AI does not care if a ship has strike craft or not, the main thing (possibly the only thing) that it cares about are the ships' frigateRoleTypes (StatCountType is irrelevant). Caveat: I did not test the generic faction AI, so it is possible that it actually forgoes frigateRoleType in favor of some sort of combat strength calculation for combat ships. Besides the generic faction AI thing, other possible explanations for observed behavior of the AI favoring strike craft: the test AIs were running with different presets, one of the AIs had a starting spot that favored a lot of ship production, small test pool meant that the significance of a single test case was overestimated, one of the AIs was playing against opponents that were pushing it to use the Heavy and Carrier frigateRoleTypes.

I've also slowly been updating my mod internally to keep up with my findings (eg. frigateRoleTypes are the key to what the AI wants to produce) and will be uploading a more current version hopefully within the next few hours.

 
 
 
Reason for Karma (Optional)
Successfully updated karma reason!
January 25, 2014 12:13:53 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Could you host this mod somewhere that doesn't want to put malware on my computer??? Christ..

Reason for Karma (Optional)
Successfully updated karma reason!
January 25, 2014 3:46:38 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

did you try the mirror link in the op then the direct download from the mirror site?

when I did this it was JUST a zip, no exe downloads.

harpo

 

Reason for Karma (Optional)
Successfully updated karma reason!
January 25, 2014 4:33:56 PM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

I did end up getting the mod.  Although the AI seems a lot more inclined to suicide capital ships into gauss turrets now.

Reason for Karma (Optional)
Successfully updated karma reason!
January 26, 2014 9:04:10 AM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Quoting StormingKiwi,

Is there any reason why the devs don't make the AI substantially easier to Mod? I'm thinking the StarCraft AI tournament or Sorian AI for SupCom

 

Quite simply, due to laziness. Hardcoding the AI is the fastest, easiest way to get something done. It just so happens it makes expanding the game infinitely more difficult, not simply for developers but for anyone else. Sins was never built with expandability in mind, hence the expansion content being what it is.

There are possible ways to get past this, of course, but the sins community doesn't have the talent to do it nor would I ever encourage someone who did have the talent to waste their time doing it on this game.

The BWAPI is essentially a hack that allows one to inject various types of code in place of the existing AI engine. It "plays" the game for the human, so to speak (at least it did when I was still doing BW stuff), and wasn't fit for general single player gameplay.

Sorian's AI takes advantage of the fact that Supreme Commander is mostly lua. Supreme commander was built with expandability in mind. That said, the game is still extremely hardcoded in many ways.

 

Apparently stardock is making some new engine that uses Mantle and might be decent for modders. If you're interested, you should spam them relentlessly to avoid hardcoding super basic stuff like AI build orders.

Reason for Karma (Optional)
Successfully updated karma reason!
January 26, 2014 10:03:15 AM from Sins of a Solar Empire Forums Sins of a Solar Empire Forums

Keep in mind IskatuMesk that it was Ironclad that made the game and they are still a small indie company with one game to their name. I don't think it was laziness that made them decide to hardcode but rather I don't think they expected this game to be so popular so they didn't feel the need to make the AI moddable.

Reason for Karma (Optional)
Successfully updated karma reason!
Stardock Forums v1.0.0.0    #108433  walnut3   Server Load Time: 00:00:00.0001000   Page Render Time: