Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Here's hoping the siege target and join army controls available in CK2 are added soon, as well as some logic tweaks to the warfare AI.
Which is exactly why the CK2 commands need to be added back in.
Put them at the end of the decision tree if nothing else. Let the AI do its thing but if it does decide to help let us give it some suggestions about what to do that will actually be helpful. Even if it decides in the end to ignore them.
Right now it's like waging a war in which none of the participants ever bother speaking to each other.
As it stands currently, the AI doesn't do much of anything for the player that isn't centered around the player's own military activities. (In before jokes which cut my quote off at "player.")
You can do that anyway by avoiding enemy armies and letting the nearby Ally AI engage instead.
That aside it's why I say "end of the tree". The point isn't to override/force the AI to do something only to apply higher priorities to things it might decide to do based on what's needed at the moment.
At least you're still exposing your forces to the possibility of harm and, more importantly, paying full upkeep though.
It either decides to help or it doesn't and THEN takes into account what you want help with. As is it's freaking near worthless.
And it's also sometimes buggy, I think. Or am I the only one experiencing AI units at sea endlessly starting to move towards land, back to sea, to land, to sea, to land, an ever-repeating cycle of movement arrows shuttling back and forth whilst the enemy beats me?
I've had that happen from time to time. My guess is the AI is in a Catch 22 between "I want to land for x or y reason" and "if I land, I will probably die".
Splitting armies is likely a side effect of not wanting to take attrition. To fix it, the AI will need to get better at recognizing when to converge on an ongoing battle. I don't often see the AI doing this, myself. Usually if there's a battle within four baronies or so, they'll send everyone to it. Usually when they don't, it's because they're using the smaller army as a diversion while they lay siege to one or two holdings.
Standing idle is the same thing, they're avoiding attrition. Allegedly 1.1 told the AI it was okay to join a nearby battle the player was in even if the numbers seemed to be in the player's favor, but if this is still happening, evidently it needs more refining.
And part of the problem is that all-AI coalitions stack together, so all-AI coalitions co-ordinate better, but human-AI coalitions generally don't stack or in any case they stack less. This leads to human-AI coalitions losing to all-AI coalitions despite numerical advantage. And it means AI co-ordinates well with AI but has a difficulty co-ordinating with a human player, especially when it comes to stacking.
AI is also clueless in its decisions, choosing neither to join its 4K with your 4K for a battle against the enemy 8K, nor to run/sail away but to just get its army slaughtered after leaving your army unassisted to be slaughtered. This means there is a reactivity problem.
My other impression is that AI has deteriorated after one of the last patches, having been better before, though I'm not certain. There have always been problems.
For the record, I would be ready to accept that some of AI's blunders are the good roleplaying of bad commanders by an AI well-designed to roleplay.
However, overall I'm not inclined to believe that the AI is roleplaying or that it is well designed.
For a moment I thought that perhaps the AI might be well-designed but just momentarily suffering from a relatively simple, though unfortunate problem — like a single issue frustrating a comprehensive good design. If this was the case, however, the problem would long have been fixed by now. And it hasn't. Either the skill or the care is lacking.
There is a reason or two programming school exists for (not to mention practical experience), and these are the sort of challenges intelligent teenage modders fascinated with coding are often quite capable of resolving, or people with some background in which they rely on logic. These issues may have challenging practical aspects to them, but they are no Gordian knots, no insoluble conundrums.
You're frankly inflating the difficulty level of a task that's not as complex as you make it in order to unfairly defend Paradox's lack of sufficient skill and care in the AI department (which is not the only such department in the CK2–3 series), and you're also being unfair to the people whose legitimate and valid criticism of the bad state of AI in this game (actually partially recognized by Paradox itself if you look at the comments in upcoming patch notes/DDs) you choose to discount so cheaply, in a way that doesn't hold water.
I probably shouldn't even respond to something like this, but it was hard to pass it by without retorting. And I once again want to stress that such attitudes by a lot of posters in community forums are responsible for enabling the sort of situation CK2–CK3 and Paradox in general is in, with very little QA/QC and sometimes issues that shouldn't exist even in a product that never went through QA and QC.
Happens all the time to me and no amount of apologetics from fanboys can justify it. It's simply incompetent AI design/programming, which has been this series's glaring problem for years now. (Referring also to the state of AI in CK2 when CK3 was released.)
Catch 22's need a tie solver. A tie solver is a very basic concept in just about any area of life that includes voting or decision-making under uncertainty, which includes people flipping a coin when you need to have a democracy of two voters. I really don't want to be mean, but it's really not something that should be a foreign concept to an IT-school graduate.
Sometimes it's moving. But attrition generally seems to be probably in the centre of it, though unequally judged by human-AI coalitions (which don't stack) and AI-AI coalitions (which do stack, and thus win despite numerical inferiority).
IIRC AI used to be relatively good at joining ongoing battles earlier in the patching history… perhaps more so than in CK2. I actually recall noting that AI appeared to be smart in helping you.
Well, I'm glad to see you aren't taking the 'everything is perfect, you guys are just toxic/ungrateful' approach, after all, but 'needs more refining' doesn't cut it. And especially not after what we've been through with CK2 AI ever since 2012.
Paradox needs to get more serious about AI design and scripting. More skill, more care. Either sit down and make an effort and finally get it right or hire someone new who can, or get some additional training. But too long is too long for AI to always be largely inept and fraught with really silly problems such as indecision loops, inability to make basic choices or not knowing how its own WS rules work, or standing next to raised ships for months without embarking in CK2, or standing immobile in a single province, allowing smaller enemy armies to pass by any siege all the way to 100% (also CK2) or AI insisting for years on hiring and sending holy orders versus people of the same religion as the order (or on not hiring them when available and when having enough piety and needing just the order or at least a band of mercs in order to defend and survive — CK2). The problems are too baffling, too many and too basic-looking for a professional game. And having to wait so long for fixes is not acceptable, and especially wouldn't be if it was something as simple as attrition checks, which could easily be reverted to pre-patch status with a quick hotfix.
But enemy AI, they have no problem meeting up to create a huge stack and having their allies siege down two different sides of the map to force me to divide my forces.
Indeed. I don't think enemy AI has any problem getting its act together to join battles when it makes sense to. I may occasionally see some silliness going on between two AI belligerents, but in my own wars enemy AI is certainly more competent than ally AI most of the time.
I've been trying over and over to do the Daurama Daura starter and this is the main reason why I never seem to be able to survive to the 3rd generation without significant territory loss
from the huge gains I make from Daurama and her first heir which are generally very strong, even with the difficulties of dealing with tribal government
In my last attempt I managed to take all of the Yoruba land and Igboland with Daurama's heir, conquered everything up to the the like 3-4 kingdoms that eventually congregate then
I get curbstomped because while I'm trying to finish off the Gaurama lands(not sure I'm spelling this correctly) the other green kingdom (don't remember the name) declares war and calls freaking Ghana (obviouslyt the strongest kingdom in the region at this point)
and bring in a stack from Ghana of 10K soldiers with their 6K soldiers
while my armt is stuck on the other side of the map dealing with the army I was actually trying to defeat. They attacked at the perfect time, even without being allies (I checked on every combatant involved) and creamed me completely
Maybe I suck at this game, but I could never count on my allies to be this efficient
I could have beaten back the green kingdom and Ghana alone but both, while I was stuck in a war with another kingdom, no way
Isn't refinement technically what we're all asking for, though? :) The unfortunate thing about CK2 AI is it doesn't apply to CK3 because the algorithm for the AI was rewritten from the ground up for the new game. They can of course take general lessons learned from CK2's AI work over the years, but they can't literally pick up where they left off. They're more or less starting over from square one.