1.7 dilemma

[OKE DOKE]

I can see many people are extremely excited about 1.7 update, however I am personally a bit pessimistic over the upcoming update given the history of previous ones. First of all, I am the player who mostly uses directly controlled investment pool due to various reasons, however I don't mind it going away and I feel like its a correct step to go. Nevertheless, in a current state of the AI and its logic this is gonna be a disaster. So that without of the rework of AI its not gonna work in a desinged way.

1. MAPI is a representation of logistic ingame, where states with appropriate input goods have a slight advantage compared to states wihtout. AI completely ignores that. State with no sulphur or lead? Please i wanna build paper, fertilizer and munition plant because there are shortages of that goods

2. Economies of scale have important bonuses such as throughput, and AI again completely ignores that, so we end up with 10 lvl 3 steel factories instead of 1 lvl 30, which is more historical given that certain areas like Ruhr or Bohemia developed as certain production hubs of heavy industry, while other areas were more agricultural or consumer goods hubs.

3. Military related goods should strictly be controlled by the player. I dont wanna see 5 lvl 1 munition plants all over the world just because I am at war and there are temporary shortages. Russia starts with 1 artlillery foundry in Ingria and a shortage, AI, instead of expanding the one in Ingria, is building many just beacuse it is profitable. In addition government was giving licences for production of a certain goods or industries to a specific states/families. One of the reason Us war of independence happened was because they were not allowed to trade autonomously and to produce high marginal goods. Even in 19th century it was impossible in Europe to come and start extracting iron or coal without state allowing to do so. With construction being mostly in control of the AI it feels like player as a representative of a state should have the possibility to prohibit certain buildings in certain states, while incentivising another, which might help to create a more encouraging and dynamic system.

4. Some wars and diplomacy should be railroaded. There were billions of post about German unification, unless you're playing with OPB mode, it is either not happening or hapenning in 1910s as a Super Germany. Even after winning Borthers war, Prussia nor Austria never attempts to liberate Schleswig Holstein. Even Opium wars are not happening after 1.6 in some games, not even talking about India or Indonesia unification or China explosion, AI is almost never trying to use its claims, even France would not try to unite Rhineland even though it had received a country-specific DLC.

5. Colonization needs a major rework, it should cost some money, as historically some of the colonial ventures were just bankrupted, it should aim towards one state - one country, and not ending up divided between 5 countries. Maybe some meachanic of decaying colony if you control less than 50% of the state.

To sum up, even though upcoming patch looks promising on paper, it seems to be a flop without significant AI improvement. I am happy to be wrong, but in a current state of the AI it seems to be facing a lot of problems and player dissatisfaction as result.

 

[Swamphill]

1/2 While I agree that AI should consider MAPI and economy of scale there i a certain realism to capitalists setting up a competing manufactory to the ones already set up earlier by other capitalists.

3 To fix this we need stockpiles, that is actual goods and not just the idea of goods like we have now.

4. Just no, there should be a high chance of opium wars, but not the certainty. Or do you argue that the UK could not have elected to act differently? Same with unification, the AI should have a high probability to do it, not certainty.

5: Remember that state areas are not God/nature given but arbitrary lines on maps.

 

[Dragnar]

I agree that the AI in its current form is too weak. Considering Paradox had to rewrite parts of economic AI anyways to work with the new ownership rules, I hope they took the opportunity to generally improve it.
The optimistic outlook is that we didn't get serious AI improvements in 1.6 because they were already working on the 1.7 ownership AI and didn't want to do work that would be obsolete within 3 months.
Fingers crossed.

 

[Revolution 11]

I agree that the AI in its current form is too weak. Considering Paradox had to rewrite parts of economic AI anyways to work with the new ownership rules, I hope they took the opportunity to generally improve it.
The optimistic outlook is that we didn't get serious AI improvements in 1.6 because they were already working on the 1.7 ownership AI and didn't want to do work that would be obsolete within 3 months.
Fingers crossed.

You would think that one of the upcoming dev diaries would be focused on AI then....

 

[47pik]

You would think that one of the upcoming dev diaries would be focused on AI then....

AI is a not a feature, it’s an implementation detail. Unlikely to get a dev diary. But I think it’s reasonable to think that “subject interactions” are probably going to involve manipulation of AI behavior since being able to exert influence on your subjects is kinda the whole point of power blocs

 

[OKE DOKE]

1/2 While I agree that AI should consider MAPI and economy of scale there i a certain realism to capitalists setting up a competing manufactory to the ones already set up earlier by other capitalists.

3 To fix this we need stockpiles, that is actual goods and not just the idea of goods like we have now.

4. Just no, there should be a high chance of opium wars, but not the certainty. Or do you argue that the UK could not have elected to act differently? Same with unification, the AI should have a high probability to do it, not certainty.

5: Remember that state areas are not God/nature given but arbitrary lines on maps.

I would argue that some events like a North German unification and opium wars are a must. Obviously as a player you can prevent this from happening, but otherwise it should happen. The problem is not that AI does not want to do it, the problem that it cannot do it. Uk does declare war on Qing, but sometimes it can not land naval invasion with current system, and it happens like 3-4 out of 10 times. Uk is scripted to do that, but it just can't sometimes. Same goes for german leadership play, Prussia can win this war, but it usually either make stupid secondary goals like on screenshot, or it will not declare on schleswig holstein liberation. Prussian AI want to do so, but system is too broken for it too happen without player intervention.

 

[OKE DOKE]

AI is a not a feature, it’s an implementation detail. Unlikely to get a dev diary. But I think it’s reasonable to think that “subject interactions” are probably going to involve manipulation of AI behavior since being able to exert influence on your subjects is kinda the whole point of power blocs

I am not talking only about subject interaction, but ai logic in general. it has to become more complex than 'hey we have a shortage, lets build stuff'. It would make the game more complex, harder to play and interesting. How much time does it take to outscale Britain on Japan/Ottoman/Russia not even talking about qing or Germany and France. Ai has to become more challenging than simple input-output simulator.

 

[Revolution 11]

AI is a not a feature, it’s an implementation detail. Unlikely to get a dev diary. But I think it’s reasonable to think that “subject interactions” are probably going to involve manipulation of AI behavior since being able to exert influence on your subjects is kinda the whole point of power blocs

Yes and looking at it historically, PDX has released several AI-focused diaries for several of their games when they acknowledge that there is a AI problem or issues with the state of the game. The same is true of performance-based diaries when that becomes an issue for the players.

 

[OKE DOKE]

I agree that the AI in its current form is too weak. Considering Paradox had to rewrite parts of economic AI anyways to work with the new ownership rules, I hope they took the opportunity to generally improve it.
The optimistic outlook is that we didn't get serious AI improvements in 1.6 because they were already working on the 1.7 ownership AI and didn't want to do work that would be obsolete within 3 months.
Fingers crossed.

Mapi introduction did not make ai better, i would even say the opposite. It made this game a bit harder for player, but ai cant just cope with it for now, thats why i am a bit pessimistic, unless they make a sweet surprise

 

[Swamphill]

I would argue that some events like a North German unification and opium wars are a must. Obviously as a player you can prevent this from happening, but otherwise it should happen. The problem is not that AI does not want to do it, the problem that it cannot do it. Uk does declare war on Qing, but sometimes it can not land naval invasion with current system, and it happens like 3-4 out of 10 times. Uk is scripted to do that, but it just can't sometimes. Same goes for german leadership play, Prussia can win this war, but it usually either make stupid secondary goals like on screenshot, or it will not declare on schleswig holstein liberation. Prussian AI want to do so, but system is too broken for it too happen without player intervention.

Ok, I've seen the UK not triggering the Opium wars, 5% of the time or so, maybe 10% tops, but I doubt it's been that much. I think that is an acceptable rate. I'd be hard pressed to remember a game where they start the war and fail it, though if I don't play in the area I tend not to keep tabs on it.

Regarding unifications I'm in favor of improving the AI ability to do so. But if every game had Germany and Italy forming the same year that would feel too static for my taste.

 

[Swamphill]

Regarding leadership war goals, do they have an infamy cost?

If they don't, or if it is low, perhaps a highish one would reduce the Preussian conquest war goals in those wars?

 

[OKE DOKE]

Ok, I've seen the UK not triggering the Opium wars, 5% of the time or so, maybe 10% tops, but I doubt it's been that much. I think that is an acceptable rate. I'd be hard pressed to remember a game where they start the war and fail it, though if I don't play in the area I tend not to keep tabs on it.

Regarding unifications I'm in favor of improving the AI ability to do so. But if every game had Germany and Italy forming the same year that would feel too static for my taste.

German unification was one the reasons British empire failed, it should happen in every game, unless player is specifically blocking it with early agressive wars. Leadership has 0 infamy, and it has to stay so, the AI should be aiming towards leadership only + maybe reparations. When Uk declares on qing it only targets trade port and reparations or something, i bet it is doable in terms of programming. But as I said Germany should be a must in every game. There is an achievement for France to prevent Germany from forming, which is marked as hard, but in a current state you just dont have to do anything, Germany would not form.

 

[Aliquam]

as I said Germany should be a must in every game.

No, the game should not be railroaded.

 

[Madzai]

Mapi introduction did not make ai better, i would even say the opposite. It made this game a bit harder for player, but ai cant just cope with it for now, thats why i am a bit pessimistic, unless they make a sweet surprise

I think MAPI and local prices need work. While no local prices were weird, currently we have a lot of issues, that historically wasn't also. And they cannot be solved within current system, forcing people to hyperfocus even before taking economy of scale into consideration. Especially stuff like two neighboring states not affecting each other at all, and there is being no way, past LZ and some tech to make local prices better, if you're not producing stuff locally.

 

[Citizen Caen]

Having ai construction really isn’t that bad. Do they sometimes make dumb decisions? Yeah, but all this results in is a slightly less optimal productivity and in the grand scheme of things I find it barely noticeable. As long as you as the player build up certain industries favorably it’s really not a big deal. You concentrate that 51 level steel factory in the iron and coal state, and make massive profits, and the ai builds one steel factory in the middle of the desert and makes a moderate profit.

So what if you have a little bit extra steel in your market, not a big deal.

 

[Woifee]

I would argue that some events like a North German unification and opium wars are a must. Obviously as a player you can prevent this from happening, but otherwise it should happen. The problem is not that AI does not want to do it, the problem that it cannot do it. Uk does declare war on Qing, but sometimes it can not land naval invasion with current system, and it happens like 3-4 out of 10 times. Uk is scripted to do that, but it just can't sometimes. Same goes for german leadership play, Prussia can win this war, but it usually either make stupid secondary goals like on screenshot, or it will not declare on schleswig holstein liberation. Prussian AI want to do so, but system is too broken for it too happen without player intervention.

No, German Unification is not a must, but at the moment AI won't form it ever. This is my problem with it. AI should be able to form it but German unification was not predestined.

 

[OKE DOKE]

No, German Unification is not a must, but at the moment AI won't form it ever. This is my problem with it. AI should be able to form it but German unification was not predestined.

I dont want to argue whether it was destined or not, if you make a decent system, give ai a tendency to form it, it will form like 8/9 times out of 10, that will do for me

 

[OKE DOKE]

Having ai construction really isn’t that bad. Do they sometimes make dumb decisions? Yeah, but all this results in is a slightly less optimal productivity and in the grand scheme of things I find it barely noticeable. As long as you as the player build up certain industries favorably it’s really not a big deal. You concentrate that 51 level steel factory in the iron and coal state, and make massive profits, and the ai builds one steel factory in the middle of the desert and makes a moderate profit.

So what if you have a little bit extra steel in your market, not a big deal.

It is not a big deal for you as player, beacuse you will have a bunch of lvl 51 steel factories in a coal iron states, but it is a deal for ai countries, who has 30-40% of their steel production in a middle of the desert. How much time would it take for you to outscale Britain in GDP as USA? 10-20 years i would assume, IRL it took them almost 60 form the game start. Those changes would make the game more challenging than build a bunch of factories and hope for good laws rng

 

[Aliquam]

I dont want to argue whether it was destined or not, if you make a decent system, give ai a tendency to form it, it will form like 8/9 times out of 10, that will do for me

90% isn't tendency, it's near sure outcome.
I agree there should be tendency, Germany forming in around half of the games would be ok.

 

[OKE DOKE]

90% isn't tendency, it's near sure outcome.
I agree there should be tendency, Germany forming in around half of the games would be ok.

Imagine creating the game about victorian era, where one of 2 main contestants of the world leader is not attending. How do you feel about US winning almost 100% wars against Mexico? You want to remove country who had 16% of world GDP by 1900? Show me the fun then. Are you ok with Japan or Russia never modernizing? Thats like another 15% of world GDP. This game is about competition, for colonies, for ideas, for world leadership etc. And you want to make this competition easier for you, I don't. In EU4 you have to make specific moves to counter Ottoman from blobbing early on. Is it railroading? It is. Does it make game worse? It does not, in fact, it gives a player extra challenge, because it is quite hard to fight Ottoman/France blobs in 1550-1700. By the way, the problem is not the tendency, as Prussia declares leadership war in 100% games, the problem is that it stops afterwards, so Germany, a place with most coal-iron states on the map, is just a bunch of OPMs