AGCEEP: AI improvement

[Bodders]

EDIT: Thinking about it, these events might be a little deterministic, and handing out a core on Mascate is sorta questionable. Perhaps it might be better to just give a core on Zanzibar and a CB on Oman and make the player/AI do the rest?

My own personal feeling is that the ai needs all the help and determinism it can get

The most important part of Daywalker's mod, in my opinion, is not the change to the ai files but the 'help' events for the ai that he uses. Of course, it's these which make it pretty much unworkable with the AGCEEP setup but I see nothing wrong with coming up with our own specific ai only events - it could help a lot, in fact.

I would therefore stick with the events as are with ai=yes added. The player will then have to do everything on his own unless there's an ai=no command that stops the ai having the event - I'm pretty sure there isn't. The other option is to make additional events which fire based on what a player accomplishes but which won't interfere with these ones. If the ai inherited a country for example then it wouldn't matter that a player event gave cb against this country that wouldn't then exist to the ai.

Got to be careful, of course, but I think judicious use of the ai=yes trigger is one definite way to go.

As to my tests, thanks for the information on the countries where I was incorrect to have them as counterreform=yes. Other than that, it's promising. In addition, I've removed the core in Switzerland from Tyrol (as discussed in the beta development thread on Germany and Austria) and given it peaceful ai. In most cases, all countries that are at the start in the HRE and Italy apart from Holstein exist 30-40 years down the line which is better than I've seen before After that it gets trickier - with peaceful ai, the one province minors will still attack eventually and one will win and become the annexing badboy leading to the usual war spiral in the HRE or Italy.

Specific problems arising from the changes so far :-

(1) Burgundy, despite being on peaceful.ai these days often goes to war, dragging a lot of other countries with it. I assume this is due to the size of its starting armies. On the other hand, this doesn't usually lead to much territory change, particularly as the likes of Lorraine (which often gets swallowed) will come back when Burgundy is inherited. A minor problem which I think I'd probably ignore.

(2) Scotland even with peaceful ai still often declares war on Eire in the first twenty to thirty years - however, England doesn't usually get involved and this war is usually settled for money so again a minor problem.

(3) Due to longer periods of peace, diplo annexing seems more common - in the first thirty years, France usually annexes Savoy, England annexes Eire and Venice annexes Athens. Some random events for the ai which decrease relations by a small amount to everyone would be useful (if that's possible).

(4) Denmark attacking Holstein - I can't seem to stop this, however I alter the ai file in the first monarch. I'm experimenting with the standard peaceful ai at the moment, though I don't hold out much hope. I have no idea how to stop this.

Cheers

 

[Registered]

unless there's an ai=no command that stops the ai having the event - I'm pretty sure there isn't.

I don't know much about events, but wouldn't it be possible to have an event at the beginning of the campaign, that sleeps all "player"events?

 

[Bodders]

I don't know much about events, but wouldn't it be possible to have an event at the beginning of the campaign, that sleeps all "player"events?

Oh yes, that would work. The ideal is that any ai only events wouldn't conflict with any of the other events then this is unnecessary. If, however, there's a situation where the ai needs a lot of help for historical behaviour and players need a little then it is possible. The colonisation and removal of native countries is about the only possible area I can think of. Even then, hopefully it could be avoided but if not then this is certainly possible.

Of course, someone could then cheat by starting as another country to sleep the player events for the one they wanted to play and then taking that over on a reload

 

[Khephren]

The player will then have to do everything on his own unless there's an ai=no command that stops the ai having the event -

Cheers

There is, and there is couple of sequences that use it. The one that comes to mind is havard's Anjou inheritence events which have a seperate sequence for ai and human player. (Using ai=yes and ai=no)

 

[Bodders]

There is, and there is couple of sequences that use it. The one that comes to mind is havard's Anjou inheritence events which have a seperate sequence for ai and human player. (Using ai=yes and ai=no)

Ah, OK, that's even better then, ignore the post above

Shows how I only fiddle with ai files and don't do events

The big expansionists and colonisers could do with direct help for the ai from events then, in my opinion. Of course, this has to be done carefully so as the gameplay for other nations isn't ruined. If you were playing the Incas and lost all your provinces the minute a Spanish unit discovered you it might be annoying for example

Worth checking out Daywalker's approach and adapting it to AGCEEP specifics - the biggest help went to Portugal, Spain, England, France (on colonising), Russia and the Ottoman Empire (last two for expansion).

 

[Khephren]

I think Daywalker's approach should be used sparingly, for colonization help at the most. That was the thing I liked most about it. But other things, like Muscovy automatically annexing the Golden Horde every game, were a bit too much to swallow. Too deterministic.

 

[hansmaa]

You need to keep one thing in mind folks. If this is going to be the MP-mod we all hope for, you need to create your events with care.
In a 16-player game there is often a bit of lag, especially if there are wars being fought.
That means that you can have an event forking one way on one 'puter and another way on another 'puter. That will make a haywire of it for sure.
Another problem is in peace-deals. Usually you won't respond to a peace-deal the moment it is offered, and if - in that moment of pondering on the deal, the days ticking away - the Lorraine you have been offered has been ceded by an event to Burgundy, then... then that mp-game is well and truely up the famous creek...

I'm not sure if adding ai=yes/no to all events that move provinces around will help, but do try to minimize the risk, eh?

 

[Bodders]

Both good points but don't worry I have no real power and was merely throwing out the thought that this could be a way to go

I accept I'd probably be more deterministic than the mod in general should go

For colonization though, a known ai problem area where the seceding of provinces from minor nations to the ai would also cause less conflicts between players in mp, I think there are definite advantages.

 

[unmerged(2456)]

I think Daywalker's approach should be used sparingly, for colonization help at the most. That was the thing I liked most about it. But other things, like Muscovy automatically annexing the Golden Horde every game, were a bit too much to swallow. Too deterministic.

That is actually a good thing...historical stuff should happen most of the time. I do want to know were you using daywalkers mod or just his AI because if its the former, he did a lot of stuff to make it more deterministic in certain areas. His AI generally is good, though the values for garrison need to be tweaked moreso than the rest. There are always exceptions though in indivisual AIs.

 

[ForzaA]

I would think that expansionists states (OE, Russia mostly) would be best served with maybe one or two countries on their "to-fight" list.. and then after they have conquered the territories they're supposed to get (at that time) they'd switch to the next target(s).
Ofcourse, they shouldn't get stuck for too long in this if someone else takes the target out (for example: Albania annexing the Duchy of Athens) so something like:

OE starts with OE_1420.ai containing Duchy of Athens and Candar (randomly picked )
and they get an event triggered something like:

OR = {
AND  ={ owned { province=xxx data=-1#athens}
            owned { province=xxy data=-1#candar} }
year = 1435 }

switching to some new targets..
ofcourse, they may need a "mop-up" AI from time to time. (For countries that somehow managed to survive being on the hit-list)

 

[unmerged(2456)]

for the OE at times this isn't possible as at certain times they expaned in ALL directions, FE the beginning of the game.

 

[Zander]

I would think that expansionists states (OE, Russia mostly) would be best served with maybe one or two countries on their "to-fight" list.. and then after they have conquered the territories they're supposed to get (at that time) they'd switch to the next target(s).
Ofcourse, they shouldn't get stuck for too long in this if someone else takes the target out (for example: Albania annexing the Duchy of Athens)

Why not base it on the existence of those countries, rather than on the province ownership?

 

[ForzaA]

for the OE at times this isn't possible as at certain times they expaned in ALL directions, FE the beginning of the game.

I think that an AI having 15 years for 2 countries, and 15 years for two other countries will do better than one AI for 30 years for 4 countries, but that's just what I think.
@Zander: they may not need to destroy a country at a certain time to gain the lands they are supposed to (for example, Venetian Islands)

 

[unmerged(35036)]

AI Issues and Disbalances

If you play this mod a lot like me, and analize the resultant map at 1810, you will see like me that its pretty diff that it should by that time.

This are some suggestions and problems i found, if we can solve this, maybe we could bring some more balance to the game.

1) Small Nations Problems
The old problem of small nations eating each other and countries like france and austria eating em all.

Well france most of the times could reach austria borders somewhere in the middle of germany and north of italy, all the smaller duchies are always eated either by conquest or diplomacy.

Key nations that would help to keep the balance in the map would be Savoy, Helvetia (Swiss), Bavaria, The Palatinat, Hessen, Hanover.

I think lot of independence events should be made for this countries.

Also it should be wise to lower their WAR variable in the AI, maybe to 5 or 0 instead of the default 10 that comes with the default.ai

This nations should have more a bit more of power but without giving em a chance to eat their neithboring nations.

2) Prussia Creation

Something really imporant is the creation of Prussia as the nation it was, its really hard to see Brandenburg become Prussia, its mainly becouse that either Polland or Lit eat The Litavian order and Prussia or Teutonic Knights always with some russian help in the process.

I think both Prussia and Brandenburg should need a bit more help in the process.

(In this game i saw the creation of the Kingdom of Italy many times, but never prussia.)

3) Austria

Well Austria becomes a real mess, instead of taking Bohemia and Hungary most of the times it takes all the smaller duchies and become france borders.
I don't know that should we do with this, but Austria never becomes what it did in real.
Mostly becouse of Turks lack of power, and diff to get the bohemian trone maybe.

4) France

Overpowered too much, it makes france go eat all countries it can either politically or by force, all allies you see in france alliances end pollitically eated, its enemies, eated by force, and at some point of the game, when france have taken from Britany to North Italy and north Germany, it collapses and at that moment you could see Polland and austria eating the map, while france explodes between french catholics and Hugnotes.
I think france shouldn't grow that big, so it will be easier for it to stand the religious disorders, being a bad boy isn't allways that good.

5) France Economic Balance

France have tons of Provinces in this game, i don't doubt France was one of the top powers in post Medieval history, but there is an unbalance of economic power that is really easy to see.
Its a question of numbers of provinces, france have like 26 really rich provinces of National Provinces (at some point it can get a bit more), that is the territory that belong to em without any conquest.
This is pretty odd, as Spain itself have 16 provinces maybe at most.
Maybe we could lower tax incomes of provinces in france, making it more normal.
(To be honbest i played france once but i quit becouse it was to damn easy, i could raise armies as i wanted and maintain it even without increasing any Inflation at all.)

6) Ottoman Empire (The Turks)

They need help bad, its pretty hard to see them taking over hungary as they did in Real Life, thus making impossible for Austria to get its coronation unless austria itself attacks Hungary at the same time that the Turks are at war with em and capture it capital (for the AI thats really hard to came up with).
I guess we should give a bit of more gunpower to this guys, either by events or something.

7) England and Scotland

Well i think most of the times England take over Scotland pretty easy, and i can't really come up with any idea on this, becouse if we give more power to the scots they will take over England Itself.

8) North Amerindians

Why the North American tribes where wipped out? even when you needed more tags and i know tags are a problem, the North Americans Natives played a big part in the fight betwen England and France over the North.
It would be wiser to put em back, at last the most reprensative of em and maybe take out some south asian cultures? remember this is Europa Universalis after all.
Its pretty depresing to get up to the colonization part in this mod, when north america is unpopulated.
I bet we could spare some tags to bring some of em back.
Whats most important i don't know but iroqueses might be one.

9) South Amerindiands

The southern parts of South America, in Argentina and Chile where almost inhabitated, but there was few nomads tribes that were so fierce they couldn't be conquested until the lates 1800's and that conquest was made by the recent created nations in look of more territory for landlords and generals.
I bet we could increase the quantity of natives a bit and make them more agressive, as i dont think we should spend a tag in the Patacon Tribes (there where several of em in the Patagonia).

10) Netherlands Creation

In pretty weird case netherlands wasn't created i can't inform you a lot about this couse i can't remember its factors, but was sure becouse either france or spain or austria didn't got that territory at that time, i guess it was becouse of Flanders and cleves or others like berg where holding that terrain.

 

[Garbon]

This isn't really related that much, but perhaps we should put in the FAQ the reasoning behind the removal of the Native American nations.

 

[dharper]

This isn't really related that much, but perhaps we should put in the FAQ the reasoning behind the removal of the Native American nations.

Yes, I'd like to know why myself - they were in the EEP when I left, but aren't in the AGCEEP now that I'm back (admittedly, 2 years later).

It's a pity, because I have a lot of event ideas involving them.

 

[kurtbrian]

threads merged.

 

[unmerged(17856)]

This isn't really related that much, but perhaps we should put in the FAQ the reasoning behind the removal of the Native American nations.

They'll be back. Eventually. I know I say that every year...but...yeah.

Anyways the facts have born out, we've all seen the evidence; the bottom line is that successful colonization pattern depends upon appropriate expansion areas, expansion values and # of settlers available to the country. There are obviously other factors, like money, but ENG's fine with that, as opposed to say a country like POR.

The NA states, especially in their proposed reduced form would produce an immeasurably tiny impact on colonization. With them, we could bring some fantasy events for NA tribe unification, finally model the beaver wars and F&IW the best we can.