A Grand List of AI Problems with the hope they will be fixed

[redmark]

the AI is NOT HISTORICAL. it will not make historical mistakes, it will do whatever it thinks is best for itself. Major wars against major powers that it thinks it cannot win are not. period.

The AI should not accurately assess the strength of itself and neighbours to such a degree. It should overestimate it's own strength, believe God is on it's side, believe in the natural superiority of it's own people and inferiority of others, be influenced by the relative ability of it's leader, etc.

The AI should make mistakes, including starting wars it cannot win. Not a one-province minor attacking France or Burgundy, but a major taking a gamble on a sudden opportunistic expansion and getting the assesments horribly wrong.

 

[jhhowell]

I was talking about major-vs-major wars: in that respect the AI acts like a human player does, and that's fine by me. It levels the playing field because everybody (supposedly) acts the same.

See, that's just it. In my opinion, and clearly that of others in this thread, the AI is not acting like a human player would. If I'm playing France, I guarantee you that I'll put together a plan to take down Burgundy, and I'll make sure to get Calais off the English at some point. AI France is happy picking on Lorraine and then picking its nose for a century.

Merrick, a slight quibble with your claim that blob stabilization occurs when the surviving powers don't have cores on each other. In my experience, that never happens - despite what the manual says, cores don't go away after a certain delay (or if they do, the delay is much longer than a century). So the starting cores, say France on Burgundy, or either on England, will last longer than the Blob Wars themselves. And Border Incident type events fire often enough that I would expect most majors to have cores on their neighboring majors. I'm pretty sure Portugal and Castille didn't start with a web of overlapping cores! The AI does (rightly) value a CB (could be a pointless embargo by a nation owning no CoTs), but there's a lot more to the DoW decision making than that.

 

[Inferis]

The AI should not accurately assess the strength of itself and neighbours to such a degree. It should overestimate it's own strength, believe God is on it's side, believe in the natural superiority of it's own people and inferiority of others, be influenced by the relative ability of it's leader, etc.

The AI should make mistakes, including starting wars it cannot win. Not a one-province minor attacking France or Burgundy, but a major taking a gamble on a sudden opportunistic expansion and getting the assesments horribly wrong.

Well, that's not a "bug" or problem, but a design decision then. Clearly, Johan et al had other things in mind when they designed/created the AI...

 

[Inferis]

See, that's just it. In my opinion, and clearly that of others in this thread, the AI is not acting like a human player would. If I'm playing France, I guarantee you that I'll put together a plan to take down Burgundy, and I'll make sure to get Calais off the English at some point. AI France is happy picking on Lorraine and then picking its nose for a century.

So, what you want is the AI to be emotional, and not rational like it is now? (Or "I hate burgundy, let's kill it" vs "Burgundy is as strong as me, let's not touch it because that would probably be a worse situation than now")

 

[Winterborn]

My AI issue of the day: When I try to sell provinces to my allied vassals they refuse to take them... even when I'm asking for 0 ducats. I've tried many prices, many provinces but they just have no desire at all. It's very frustrating.

 

[Darth Tracid]

My AI issue of the day: When I try to sell provinces to my allied vassals they refuse to take them... even when I'm asking for 0 ducats. I've tried many prices, many provinces but they just have no desire at all. It's very frustrating.

are they their cores, culture and religion, and adjacent to their provinces? if not, not a chance.

 

[Winterborn]

Adjacent and religion yes, unsure on the culture.

It seems to me that they should want free provinces anyway.

 

[Inferis]

Why are you so sure about that? Would you accept every province somebody else would throw at you, even for 0 ducats?

 

[jpd]

I would not. It depends on how well said province fits into my nation. Slider tolerance wise and what not.

 

[Rwn]

Make Badboy depend on the overall power and "danger" (number of provinces, income, attitude towards other countries, number of armies, technology, lack of internal trouble...) instead of just agressivity.

A 20-provinces, rich France should be perceived a as much bigger threat, even if (and even more !) if it was in peace for the last 50 years than some 1-province minor which happened to gobble two of its neighbours. High badboy nations should on this respect be much more concerned about other's badboy (i.e. a "dangerous" nation, aka major nation, should be more likely to fight another major).

Oh, and make AI agressiveness not depend solely on cores. I had a 2-provinces Lithuania who kept declaring war to my giant Russia which stopped the day I made it renounce all its core, I then insulted them many times, kept my relations very low, had a high badboy, they never reacted, until they finally had another Boudary dispute event 50 years later and restarted to keep declaring war on me.

 

[unmerged(67170)]

Alternatively make badboy limit be inversely proportional to size.
If you are very large you will have very little room to maneuver before other countries will get concerned, but if you are small, they will be less concerned.

 

[mlepkows]

are they their cores, culture and religion, and adjacent to their provinces? if not, not a chance.

I can't agree with that. At least not with the "not a chance" part. I've sold some provinces in HRE and Kurland areas to nations with either wrong culture (non-national but accepted, I think) or no cores on said provinces. It sometimes took a few attempts but someone finally agreed.

Maciek

 

[Carewolf2]

Alternatively make badboy limit be inversely proportional to size.
If you are very large you will have very little room to maneuver before other countries will get concerned, but if you are small, they will be less concerned.

Sounds reasonable

 

[JScott991]

Sounds reasonable

Except major powers don't react to badboy anyway so a solution would have to be more comprehensive.

 

[Slyspy]

I suspect that EU3 is designed as little more than a canvass against which to play multiplayer games, which is why the single player game is rather disappointing.

Many features, for good or ill, have been added in order to benefit the multiplayer game.

Talking of which, does multiplayer actually work properly?

 

[unmerged(12921)]

Alternatively make badboy limit be inversely proportional to size.
If you are very large you will have very little room to maneuver before other countries will get concerned, but if you are small, they will be less concerned.

This is a TERRIFIC idea. Hope the "powers that be" are listening

 

[unmerged(12921)]

I suspect that EU3 is designed as little more than a canvass against which to play multiplayer games, which is why the single player game is rather disappointing.

Many features, for good or ill, have been added in order to benefit the multiplayer game.

Considering that studies still show that the great majority of computer games are played by individuals, doing this without calling it EUIII MULTIPLAYER would be a really shabby thing to do to their customers. Let's hope it's not the case.

 

[Palle]

Considering that studies still show that the great majority of computer games are played by individuals, doing this without calling it EUIII MULTIPLAYER would be a really shabby thing to do to their customers. Let's hope it's not the case.

I also have a feeling that the game was balanced more for Multiplayer than single player. In MP blobification and supreme cavalry doesn't matter since human player's can respond to those easily. However, majority of the Paradox players only play SP, I think. It would be shame if Single Player is not as well balanced as MP.

 

[unmerged(24599)]

A rather belated point here (this has been quite a busy thread!).

Those people defending the passivity of AI majors are arguing against a strawman. Nobody is suggesting that majors of equal power should blindly attack each other. AI majors should attack each other when one perceives an advantage against another, just as a human would. This could be an absolute advantage (as Castille will usually have over Aragon or Portugal, should they all happen not to be allied in a game) or a temporary advantage (France attacking Burgundy while the latter is mucking about in a war against a bunch of HRE minors). Past Paradox AIs have been able to do this, so I'm not sure why the EU3 AI doesn't.

I wouldn't be this categoric about the EU3 AI. It does attack you in the latter case you mention - if you get embroidered in wars elsewhere. In my recent Castile game Aragon jumped on me "out of the blue" when I was tangled up in a prolonged war with Morocco and Algiers. They would not have doen this had I been at peace.

The problem is to evaluate the former correctly, and how to make the AI recognice a good opportunity. I fail to see the "strawman" in saying that

The overall balance of power problem is tricky. BB is the zeroth order mechanism to control overly powerful nations, but as previously noted it's easy enough to create (say) a super-Germany without significant BB. Lord Vader's idea for an event may be workable. The key to constraining an emerging superpower is forming a strong coalition. Thinking out loud, let's say there's an event triggering on one nation having a military score 50% greater than any other nation in their capital region (Europe, for example). If there exist three other nations in the same region whose combined military is 50% greater than the emerging superpower, trigger events for each saying "Nation X is growing too powerful, we must work together to maintain the balance of power" creating alliances between all three and bumping up relations by +200 (to make it more likely alliance calls would be honored). Then have a follow-up event for when neither trigger exists, either through the alliance trouncing the superpower in war, or just catching up in strength, with the effect "Our work is done, go our separate ways" undoing the alliances and dropping relations by some amount. Precise numbers and effects are certainly open for tweaking, I'm just presenting a first concept of how this might work. It really doesn't seem like a crazy or unreasonable thing to do - nations should (and did) form such coalitions at times. It would be nice if the AI could be made smart enough to form such coalitions naturally, but absent that, an event makes good sense to me.

It will probably be easier to tweak the AI then to make any events like this...

 

[JScott991]

I wouldn't be this categoric about the EU3 AI. It does attack you in the latter case you mention - if you get embroidered in wars elsewhere. In my recent Castile game Aragon jumped on me "out of the blue" when I was tangled up in a prolonged war with Morocco and Algiers. They would not have doen this had I been at peace.

I have never once been attacked as Austria starting in 1510. I've been involved in 1-2 year wars with Poland, 3-4 year wars with France, constant wars in the HRE it seems (stupid allies), 2-3 year wars with Turkey (it takes a long time to siege every province between Buda and Constantinople), and not once has the AI seen fit to declare war on me. I'm probably in my sixth complete game (1510-1680 or so, since the map is too stagnant to play much further; the longest I've played is 1707) and not one DoW against me. I really hope they fix this, but the moderators' and betatesters' posts on this thread suggest otherwise, meaning the suggestions we are making here are probably just dandelions in the wind.