[Sub-Mod] Improved Campaign AI v4 relased- Enhancement Mod for APE:TI + Discussion

[Matte979]

The goal of this SUB-MOD is to create a stronger AI and an AI thats more adjusted to the game mechanics of APE:TI.

Version 5

updated for 1.0.5, The mod is back to just having the CAI file as 1.0.5 have the additional line infantry recruitment as standard.


Version 4 has been released. This mod is save game compatible but for best effect a new campaign is recommended.

Changes in v4:

AI will defend better
AI will keep more troops to combat unhappiness (Help the AI to hold regions when conquered)
AI relation adjustments toned done, less angry on other countries (Changes from Darthmod AI)
AI will build a few less ships (With the v3 changes all the AI players had navies even if they needed ground troops more)
AI build larger stacks and have higher priority to take back lost regions and less priority to raid

Details in Readme.


To summarize changes in v3:

AI should build more troops
AI should have larger stacks
AI should create stronger stacks
AI should be a more reluctant to declare war if it does not have the strength to fight the war.
AI should have slightly more artillery
AI should defend better.

For installation instructions and details check the readme.

[Jet Jaguar]

Thanks Matte979, these changes look fantastic; I'm surprised these settings weren't originally set to how you've changed them. Maybe CA thought it'd be too challenging.

I think the best and most important change might be the "WAR_AND_PEACE_MANAGER_PROPORTION_OF_REQUIRED_FORCES_TO_START_WAR". Meaningless and wreckless declarations of war by the AI (against both the player and the other AIs) is my biggest problem with 1.3; it was making me think of just waiting for 1.4. I had no idea this could be modded. There are lots of threads about the overly warring AI in 1.3; I'm sure that many players would like to know that something can actually been done about it.

also: I intend to use your changes with APE:TI, but I also wanted to do some vanilla 1.3 testing; is your sub-mod compatable with vanilla 1.3 as well as APE:TI?

Thanks again; your efforts have saved me from having to wait for 1.4 to enjoy E:TW.

[Matte979]

The file is compatible with the standard 1.3.

Yes that parameter is one I would like to explore more as reducing wars a little would not hurt.

Currently in my test it did slowed the "Std" declaration of war with a few turns but as the AI build more troops now they get the required strength faster so in my opinion its still Total War out there and AI seem to declare war not because they have an objective but just because their alliance total strength is stronger does not matter how many they are at war at the same time. I hope the really put some effort into the CAI in the next patch.

One thing I might do is to increase it further and see the effect. I am out of town this weekend so cant test more until next week.

But the changes above do reduce wars a little and the AI have more troops to commit to battle, specially minor factions seem to be able to hold their own better and factions that loose troops in battle seem to replace them faster so they dont get overrun. I play with vh (-2) script setting.

[Jet Jaguar]

The file is compatible with the standard 1.3.

Yes that parameter is one I would like to explore more as reducing wars a little would not hurt.

Currently in my test it did slowed the "Std" declaration of war with a few turns but as the AI build more troops now they get the required strength faster so in my opinion its still Total War out there and AI seem to declare war not because they have an objective but just because their alliance total strength is stronger does not matter how many they are at war at the same time. I hope the really put some effort into the CAI in the next patch.

One thing I might do is to increase it further and see the effect. I am out of town this weekend so cant test more until next week.

But the changes above do reduce wars a little and the AI have more troops to commit to battle, specially minor factions seem to be able to hold their own better and factions that loose troops in battle seem to replace them faster so they dont get overrun. I play with vh (-2) script setting.

Thanks for your reply. Where can I find "WAR_AND_PEACE_MANAGER_PROPORTION_OF_REQUIRED_FORCES_TO_START_WAR"?
Can I mod it with a hex editor?

Thanks again for sharing this sub-mod.

[Yarkis de Bodemloze]

@Matte979

Interesting thread. I also started to experiment with some of the values this week. I did soemthing extreme with WAR_AND_PEACE_MANAGER_PROPORTION_OF_REQUIRED_FORCES_TO_START_WAR and set it to 999 just do see the effects more clearly. I can confirm that the AI, and especially the minor nations won't start any wars so easily with higher settings (actually i didn't witness a minor starting a war with these settings at all). However, the expected side effect is that the AI won't invade that easily even if at war, but instead builds up very strong armies. So currently I am trying to find a good value beyond the 150 you propose, because 150 leads to a delay but not to a fundamental change I would like to see. These experiments also confirm that the AI likes to start wars if it has enough forces, and it seems that it adds not only it's own forces to the equation but also the armies of it's allies. But due to the fact, that many allies don't follow the war declarations, especially if a minor is declaring war, the AI most of the time miscalculates it's true strength, which leads to those insane wars between a single minor and a major, like Savoy against France or Curland against Prussia.

Changing the spending policy in favor of the military may be problematic on the long run. Actually I favor the opposite approach by trying to improve the economic power of the AI.

[Jet Jaguar]

I wonder what the effects would be if "WAR_AND_PEACE_MANAGER_PROPORTION_OF_REQUIRED_FORCES_TO_START_WAR" was raised, but "BASIC_SPENDING_BIAS_RECRUITMENT_ARMY" was lowered?

Maybe the A.I. would be more historically reasonable about it's amount of war declarations because it'd need a higher military ratio to declare war; but it wouldn't be favoring spending on it's military over it's economy. So the number of turns it'd take to reach that military ratio needed for a declaration of war, would be greater.

At the same time, I think the A.I. really needs to build up it's military to make a campaign challenging. I think 1.4 will improve the CAI; hopefully signifigantly. Looking around the forums, the CAI seems to be most player's main complaint about 1.3; it's certainly mine. But in the meantime, I'd like to keep playing. That's why I was really pleased when I saw this thread.

[Tariq]

Well i tried this mod with APE:TI and IS and i can see quite difference. I mean , you will still incur negative penalties like before , but they are more manageable, if you act with combination of different actions.

And you are right , mileage will depend from mod to mod , with Vanilla being least prone to have any difference.

With both mods i was able to avoid multiple wars , even when i broke alliance with Denmark as Russia.Sweden was very red , but they are still not declaring war in this case , which actually benefits them , considering that i would easily take 2-3 provinces from them at this stage.

And peace agreements look more achievable and sustainable.It's far from perfect , but at least it's much more playable.

[notger]

Sorry to ask, but I had to reinstall my game and I am unsure now: In which directory do I have to put The Creative Assembly/Empire/scripts/user*.txt ? What is the base dir?

[l33tl4m3r]

Sorry to ask, but I had to reinstall my game and I am unsure now: In which directory do I have to put The Creative Assembly/Empire/scripts/user*.txt ? What is the base dir?

VISTA C:\Users\(Your User Name Here)\AppData\Roaming\The Creative Assembly\Empire\scripts

XP C:\Documents and Settings\(Your User Name Here)\Application Data\The Creative Assembly\Empire\scripts

[Matte979]

@Matte979

Interesting thread. I also started to experiment with some of the values this week. I did soemthing extreme with WAR_AND_PEACE_MANAGER_PROPORTION_OF_REQUIRED_FORCES_TO_START_WAR and set it to 999 just do see the effects more clearly. I can confirm that the AI, and especially the minor nations won't start any wars so easily with higher settings (actually i didn't witness a minor starting a war with these settings at all). However, the expected side effect is that the AI won't invade that easily even if at war, but instead builds up very strong armies. So currently I am trying to find a good value beyond the 150 you propose, because 150 leads to a delay but not to a fundamental change I would like to see. These experiments also confirm that the AI likes to start wars if it has enough forces, and it seems that it adds not only it's own forces to the equation but also the armies of it's allies. But due to the fact, that many allies don't follow the war declarations, especially if a minor is declaring war, the AI most of the time miscalculates it's true strength, which leads to those insane wars between a single minor and a major, like Savoy against France or Curland against Prussia.

Changing the spending policy in favor of the military may be problematic on the long run. Actually I favor the opposite approach by trying to improve the economic power of the AI.

Good, thats what I thought and this is the information I would like us to find out with this thread, so we can identify a few things that then can be used in APE:TI to create a better campaign game.

The reason I lowered economic building was due to the AI just building everything and not building enough troops even when they need troops for fighting a war. Not sure I agree that the AI need to spend more money on economical development, I think the vanila setting makes them focus to much at the expense of defences.

As I stated I can sadly not test my changes more until middle of the coming week. But all feedback is great.

[Matte979]

Well i tried this mod with APE:TI and IS and i can see quite difference. I mean , you will still incur negative penalties like before , but they are more manageable, if you act with combination of different actions.

And you are right , mileage will depend from mod to mod , with Vanilla being least prone to have any difference.

With both mods i was able to avoid multiple wars , even when i broke alliance with Denmark as Russia.Sweden was very red , but they are still not declaring war in this case , which actually benefits them , considering that i would easily take 2-3 provinces from them at this stage.

And peace agreements look more achievable and sustainable.It's far from perfect , but at least it's much more playable.

Good, I think an increase of the strength to declare war should be the first change. Maybe from 150% to 200%, I dont want to stop to many wars. But as APE:TI reduce the strength of the player comparable to the AI we should still see enough attacks on the player.

[Tariq]

Good, I think an increase of the strength to declare war should be the first change. Maybe from 150% to 200%, I dont want to stop to many wars. But as APE:TI reduce the strength of the player comparable to the AI we should still see enough attacks on the player.

Yeah , you probably should be able to maintain decent balance with it, but it will take some testing to confirm that.

I made little bit of adjustments to various values(based on what i have noticed with Russia) and checking that with UP,only 40 turns into campaign so far though.

[Society]

Have you checked out the:

DB->Campaign_AI_Manager_Behavior_Junctions file?


There's a few simple "War_And_Peace" ones in there, all of which are set to "1000" ...

Perhaps saving it to a different number would do the trick? FOr some reason when I try to create a pack file my pack manager crashes, so perhaps you could test it? Try setting it to 1 and then try setting it to like 10,000 or something.

[alpaca]

Just a heads up: We're not sure which values are actually used by the game and which ones are pre-parsed into the startpos file. So for all changes you do you will need a statistically significant number of games that show a difference - seeing that the effect of AI is largely subjective you'll have to play a lot of campaigns before changing anything more. This is the primary reason why we have left AI untouched for now.

[aeoleron9]

I got some arbitrary results after fiddling with these values. Required force to assistance/war was at 100/150 but the AI is still dow'ing like hell. On the other hand I raised the basic spending bias diplomatic to 0.2 and actually saw Poland and Russia ally with Denmark. But no new trade agreements. I think 1.3 messed that up, because I recall much more trade going on in 1.2.

[l33tl4m3r]

I have noticed the absense of trade agreements as well; let us know how it goes with that!

[notger]

Though it is only one campaign and not representative, I feel that the declaration of war is closely tied to the number of troops you have in a province.
An unguarded province seems to trigger an attack immediately, no matter what the diplomatic status is. (Could be a hint for the weights.)

Still, I see less DOW's with this mod, though again, this is not representative, more of a feeling.
(Two prussian campaigns, one with and without your sub-mod.)

[aeoleron9]

OK false alarm, Denmark was allied to Poland and Russia from the start. And I think it's official that AI can't make trade partners in 1.3: I let Ireland rebel because it's easier to observe, and new factions used to make trade partners quickly. Now, Ireland is trading with nobody after five turns, while the other factions are full of free slots. I had even upped the diplomatic bias to 0.35 for this new campaign, still no luck.

So basically, I think the AI's diplomatic ability is completely borked atm, and tuning the values don't seem to have much effect. I hope Kieran means what he says in the interview and will really focus on AI balance in 1.4...

[erasmus777]

I noticed there's an AI entry that prioritizes what ships focus on: defending the coast, defending ports, patrolling neighbor's coast. They're numbered 3,2,1 respectively and I assume 3 is the highest priority. Anyone tried changing these so the AI will prioritize defending their ports?

[Matte979]

OK false alarm, Denmark was allied to Poland and Russia from the start. And I think it's official that AI can't make trade partners in 1.3: I let Ireland rebel because it's easier to observe, and new factions used to make trade partners quickly. Now, Ireland is trading with nobody after five turns, while the other factions are full of free slots. I had even upped the diplomatic bias to 0.35 for this new campaign, still no luck.

So basically, I think the AI's diplomatic ability is completely borked atm, and tuning the values don't seem to have much effect. I hope Kieran means what he says in the interview and will really focus on AI balance in 1.4...

I assume the trade issue is true, that is a major bug. *SIGH* I guess we have to put some hope to 1.4. I still would like to play this game now not in 1 month.

Is there anyway we can create script that will create trade agreements to tie us over until 1.4?

I am back in town and plan to do some more testing on some of the variables above.