Civ VI is SO close to greatness - A call to the developers

[Gedemon]

I think you are wrong here. The entire game as it is now has been designed with 1UPT in mind. Yes there is some nuances that are implemented here and can be expanded upon. But if the key point is resources and development time, you are mistaken.

Consider that to use your approach you should first implement a new UPT system and then rebuild the AI including many systems that already work fine in order to just start to implement war strategies. Also this affect the balance of the entire war system, from production costs to unit strengths.

The alternative of just improving and iron the current system is no doubt much less time consuming and much less likely to need debugging, testing, QA and patching. Those will require alone more resources than the necessary to just improve and deepen the AI.

If you are aware of how software development works, you can easily see that changing in this stage of the game a core mechanic would lend to months of developing and insane costs, in addition to the actual AI development.

This is just not a viable solution and even more probably not a solution most players want since it removes the fundamental strategic elements of combat. Yes making it more simple for the AI to handle but also more boring for the player to use.

"improve and deepen the AI", is months of development time for one man if VP is a reference.

Yes, obviously 1(c)UPT won't change for civ6, the base game, as you say it is a core mechanic that most of the players like, but I can say it's not because of time development to rebalance everything, and for that I do speak from experience.

I've 6000+ hours of play with civ5, most of them where with n UPT, and were actual (fun) playtime. I've logged 4000+ hours in civ6, most of them being autoplay to test stability, but the few actual hours of playtime were also done with n UPT.

I my mods, adjusting production cost and units strength were just a few days of development time, not even weeks.

And the vanilla AI could handle it BTW, no need to re-code anything, even with heavy modifications of the combat system, I'm not sure that VP tactical AI specialized on 1UPT would perform much better than the vanilla AI in my civ5 WWII total conversion with 2UPT, ranged units attacks limited to adjacent tiles, counter-fire and offensive/defensive support fire.

I did made some ajustement to the tactical AI for that mod, but it wasn't related to nUPT or the new combat rules (which were helping the AI by simply existing) just some hardcoding of fighter's (rebase, intercept, air domination) and destroyer (hunting subs) behavior.

Ho, and go and try the Stalingrad scenario, and come back to tell me how boring/easy it was, please

This is a key point. I don't think the game has to be substantially changed here. Just tweaking trade, resource distribution and resource requirements will help a lot no doubt. But if resources are too easy to get, the core game mechanic would be lost and I think it is one of the best ideas in the game and should not be simplified too much. In my opinion the path to follow is to implement AI strategies to judge the resources it needs and effective ways to pursue them.

Balance in design, again, making something that can both please the human players and not penalize the AI too much. They should be important but not too much. They should give bonuses in producing/moving/healing units requiring them, but not completely prevent their construction if you stick to the few deposits per map.

Stockpiling is a good idea (can't say otherwise !), IMO they could expand on it a bit more, then it would be easier to "implement AI strategies".

Maybe i expressed myself badly. When i said "game state" i was merely expressing the ability of the AI to judge the situation of the game. IE, how is my map, how are other civs doing and what opportunities do i have according to my current situation. The AI already does this. And this is pretty much the approach taken by the Real Strategy mod. In that mod the AI commits to one or other strategy depending on civ and leader preference, but also on the current situation of the civ and its known rivals in that victory path. Also commits longer to a strategy instead of flipping constantly and takes into account only the available victory conditions of the game. This is building on the AI systems that the game already has. This mod has shown how good this approach is in making a AI more competitive and FXS should take note

I understand your point, mine is that I want an AI that can build an empire along mine, not one that act as a human player and will focus on a victory type and rush it, as it should with the current victory conditions. To rephrase, I'd like the victory conditions to be more balanced, and so I'd want an AI capable of building a balanced empire and survive the full game's length.

That said, Real Strategy concept is indeed good, and from what I've understood of it, should be able to play the "build an empire" game when no victory conditions are enabled by judging other players advancement and adjust it priorities based on its neighbors.

 

[oSiyeza]

"improve and deepen the AI", is months of development time for one man if VP is a reference.

And the vanilla AI could handle it BTW, no need to re-code anything, even with heavy modifications of the combat system, I'm not sure that VP tactical AI specialized on 1UPT would perform much better than the vanilla AI in my civ5 WWII total conversion with 2UPT, ranged units attacks limited to adjacent tiles, counter-fire and offensive/defensive support fire.

Yea, transition from 1UPT to 2UPT should not be hard. The game does already implement the support units to bend the 1UPT rules. However how easy or hard to adapt it would be for Civ 6, depends on how it is actually coded. You are probably right, I was thinking in a more drastic change.

That would only improve path-finding, though. Still a lot of work for the AI to be improved. And I'm not even convinced path-finding cannot be done right with 1UPT, that should already be solved in Civ V.

I understand your point, mine is that I want an AI that can build an empire along mine, not one that act as a human player and will focus on a victory type and rush it.

Strategies to pursue a Victory path, are not actually so restrictive in Civ 6. Any Civ will build a balanced empire despite the strategy they use. AI is actually a very complex system, that most user complains fail to realize.

AFAIK, besides goal strategies to guide the Civ towards a victory condition (Macro strategies that guide long term goals) and are dynamical. Civ has like a dozen of supporting strategies, more situational, for defense, army building, exploration, situational naval guides, science and culture catch up strategies and even era based strategies that modify for example science priorities based on the current era (higher on the Information era and lower in Middle Ages, for historical flavor). Those act upon 27 AI subsystems that rule building, yield evaluation and other complex AI needs. And those act on top on probably hundreds of behavioral trees. It is actually a quite flexible, elegant and very complex system that allows to pursue short-term, middle-term and short term-goals, while keeping leader differences and overall balance empire building.

One of the difficulties of the system is in fact, that detecting a bug or testing a simple modification is very difficult cause the AI will behave mostly in a sound way even if a part is broken, cause all the AI systems works together in an interconnected way. For example if victory macro goals don't work, or support assault city operations never activate due to an impossible condition on the tree, that may be unnoticed in a single play-trough unless you are consistently looking for a specific behavior, and even then.

 

[Trav'ling Canuck]

I understand your point, mine is that I want an AI that can build an empire along mine, not one that act as a human player and will focus on a victory type and rush it, as it should with the current victory conditions. To rephrase, I'd like the victory conditions to be more balanced, and so I'd want an AI capable of building a balanced empire and survive the full game's length.

I consider this a really important factor for me, immersion wise, and I really like the way you stated this, Gedemon. I don't care personally care whether the AI is getting bonuses or, better yet, if the player is getting handicapped at the higher levels. Whatever it takes to allow the AI civs to develop generally in parallel with the human. "Keeping up with the Joneses" is a big part of "one more turn" for me, and if the AI isn't at risk of getting ahead of me, this is lost. Diplomacy's also more interesting, to me, if I need to worry about a two-vs-one attack, for example, which isn't just the AI's military competence, it's the AI's general ability to field and maintain an army (and ability to repair and replace after a lost war, which is another whole topic).

Regarding balance, one of the problems with the seventeen flavours of victory approach is that it rewards specialization. It's a fine balancing act to structure the victory conditions such that they encourage enough specialization to make them feel different without making whole other aspects of your economy irrelevant. I personally prefer an approach where there's only one victory condition, but many different ways to get there, with the best approach depending on your circumstances and those of the other civs in the game. I think it's easier on the AI, while still allowing players to try out different approaches with different leaders.

 

[acluewithout]

I consider this a really important factor for me, immersion wise, and I really like the way you stated this, Gedemon. I don't care personally care whether the AI is getting bonuses or, better yet, if the player is getting handicapped at the higher levels. Whatever it takes to allow the AI civs to develop generally in parallel with the human. "Keeping up with the Joneses" is a big part of "one more turn" for me, and if the AI isn't at risk of getting ahead of me, this is lost. Diplomacy's also more interesting, to me, if I need to worry about a two-vs-one attack, for example, which isn't just the AI's military competence, it's the AI's general ability to field and maintain an army (and ability to repair and replace after a lost war, which is another whole topic).

Regarding balance, one of the problems with the seventeen flavours of victory approach is that it rewards specialization. It's a fine balancing act to structure the victory conditions such that they encourage enough specialization to make them feel different without making whole other aspects of your economy irrelevant. I personally prefer an approach where there's only one victory condition, but many different ways to get there, with the best approach depending on your circumstances and those of the other civs in the game. I think it's easier on the AI, while still allowing players to try out different approaches with different leaders.

I'd mostly agree with this, but two small points

First. I want some AI to keep pace - to be big enough and scary enough to keep the pressure on.

But I don't want all the AI to keep pace. Some should fall behind, because there needs to be AI to bully and pillage ... but also to form alliances with etc.

That's why I don't like real rubberband mechanics. Because you just end up with everyone being basically the same and that's no fun. IRL the world is made up of hyperpower(s), regional powers and lots of other countries. I think the Civ map should generally end up the same.

Second. I don't think Civ needs just one victory condition per se. But I do think Civ has made too mistakes. One: there are too many victory conditions. There should never have been a Diplomatic Victory and Religious Victory is silly. Having these victory conditions just burdens these mechanics with unnecessary balance issues that stop the the wider related mechanics being meaningful. Two: it's too narrow making all victory conditions balanced . Some VCs should be harder than others. That way, if you go for say a harder culture victory, one thing you have to balance with that is someone stealing a victory with the easier science victory. Perhaps that needs to be somehow balanced with some victories being worth more than others (eg they give a higher score victory). But having them all be "the same" is very boring.

 

[tab]

Lets not pointlessly complain about how bad the game it is. That is not constructive at all.

Like i said the aim of each (of MY) post is to let people and the developer know the current state of the game. The present situation indicates the improvements in patches/dlcs to date, however incremental, is just NOT improving the AI,period. And after such a long period after release a customer expects that this would change which is not the case here ergo the complaints. The aim of all such post is to consolidate into an outcry which the developers might finally respond to.

But saying things like well the AI makes air/land units at least, that good enough! well that is NOT constructive IMHO. The developers need to know what is wrong with their product instead of being coddled.

Lets not use biased observations to support a preconceived conclusion.

The AI being currently lacking does not support at all your claim that has not improved at all or that FXS does not care about the AI. A lot of complains from the initial state of the AI have been addressed.

Biased observation??? Are your kidding me? I would direct you to the plethora of forum topics not only here but also https://forums.2k.com/forumdisplay.php?794-Civilization-VI-General-Discussion, here https://steamcommunity.com/app/289070/discussions/?fp=2 and here https://www.reddit.com/r/civ/. This "biased observation" is shared by thousands of players and thousand many more who don't even visit the forums at all. So I would advise you AGAIN to go through the forum topics at least to find out what the vast majority of players are complaining about before reaching YOUR preconceived conclusion.

You noticed an odd behavior of the AI, cool.

When did it happen? is this a consistent or frequent bugged behavior? can be replicated? do you have a saved game to submit to the bug reports?

I knew some people would ask for this even though hundreds are reporting the same behavior therefore as i said in my last post ill be recording vids and pics of my last game and post them here.

 

[Gedemon]

I knew some people would ask for this even though hundreds are reporting the same behavior therefore as i said in my last post ill be recording vids and pics of my last game and post them here.

The AI logs would be very useful addition in those reports, as the observer bias is real.

 

[bitula]

Any modder happen to know whether there is an army versatile-composition check in the AI code? In the sense that the AI checks for a minimum ratio of ranged, melee and sieged units in the cluster before initiating an attack? Or AI just moves any unit available at the given moment? Or this would require seeing DLL source to answer? Was there in Civ 5? Or in VOX Populi? Or there aren’t even clusters (minimum army size requirements?) in 5/6/vox.

 

[Infixo]

@bitula Each operation type has a predefined team composition, basically what you described is how the game works. Required unit types are set via min and max values, plus land / sea ratio.
From my observations is mostly works, except for land / sea ratio which seems buggy.

There is also a check on team strength, min, max, starting and ongoing. It uses some ratios against the opponent, but I couldn’t crack the details here. This is important however because it governs the tactical decisions like getting a new unit or abandoning the attack.

 

[bitula]

@bitula Each operation type has a predefined team composition, basically what you described is how the game works. Required unit types are set via min and max values, plus land / sea ratio.
From my observations is mostly works, except for land / sea ratio which seems buggy.

There is also a check on team strength, min, max, starting and ongoing. It uses some ratios against the opponent, but I couldn’t crack the details here. This is important however because it governs the tactical decisions like getting a new unit or abandoning the attack.

Thx Infixo,

I just wanted some confirmation that the AI is not an inherent misconception or shallow or lazy implementation as one would conclude reading the so common “AI s*cks” type of threads but rather there exists a set of quite specific bugs or sometimes just observation bias mostly from those who don’t like RP elements.

 

[oSiyeza]

I consider this a really important factor for me, immersion wise, and I really like the way you stated this, Gedemon. I don't care personally care whether the AI is getting bonuses or, better yet, if the player is getting handicapped at the higher levels. Whatever it takes to allow the AI civs to develop generally in parallel with the human. "Keeping up with the Joneses" is a big part of "one more turn" for me, and if the AI isn't at risk of getting ahead of me, this is lost. Diplomacy's also more interesting, to me, if I need to worry about a two-vs-one attack, for example, which isn't just the AI's military competence, it's the AI's general ability to field and maintain an army (and ability to repair and replace after a lost war, which is another whole topic).

Regarding balance, one of the problems with the seventeen flavours of victory approach is that it rewards specialization. It's a fine balancing act to structure the victory conditions such that they encourage enough specialization to make them feel different without making whole other aspects of your economy irrelevant. I personally prefer an approach where there's only one victory condition, but many different ways to get there, with the best approach depending on your circumstances and those of the other civs in the game. I think it's easier on the AI, while still allowing players to try out different approaches with different leaders.

I think i expressed myself poorly here. The game already uses victory long term strategies. But it uses them badly and that is inmersion breaking. Never happened to you that an AI rushed your cities with appostles and then ten turns later suddenly forgot about religious game. Or AI being close to science victory and suddenly kind of stopping?

That is probably cause a change in the victory goal of the AI. In a healthy game, all players should be building an empire in a balanced way. Whyle at the same time, religious civs if have founded a religion playing a bit more the religious game and agresive leaders being a bit more expansionist if they can, while at the same time another civ is deciding that it needs to increase its science output cause it is falling down with regards to other civilizations and increase its cultural game to take advantage of the high tourism is getting. That is what a bit more focussed global strategy decission making would bring to the game.

If you still believe that is the wrong way to approach the AI, and that maybe FXS instead of making it more inmersive should abandon the current system so the AI does not try to pursue any victory goal. I would actually like to discuss what your alternatives or vision for the AI may be.

 

[oSiyeza]

Biased observation??? Are your kidding me? I would direct you to the plethora of forum topics not only here but also https://forums.2k.com/forumdisplay.php?794-Civilization-VI-General-Discussion, here https://steamcommunity.com/app/289070/discussions/?fp=2 and here https://www.reddit.com/r/civ/. This "biased observation" is shared by thousands of players and thousand many more who don't even visit the forums at all. So I would advise you AGAIN to go through the forum topics at least to find out what the vast majority of players are complaining about before reaching YOUR preconceived conclusion.

Yes, Biases are very often shared by millions of people, you only need to look at the time everybody knew the Earth was flat (some people still does) or current distribution of religious beliefs.

One of the things that makes bias very likely is being emotionally attached to a subject.

The problem often is that what is obvious from the outside is very hard to reach from a different perspective. For example when you grated that the AI builds planes, to me that just confirmed on your own terms that you know the AI has improven. Since that was an issue for a long time in the forums, and Im sure you know it. Building planes is better than not building them, using navies is better than not using them.. and so on. To me that is the very definition of improvement, no matter how pointless all these changes are to you.

It really should be very easy for us to be on the same team here, since we both agree on almost everything. The AI being lackluster, both thinking that is one of the major problems of the game both wanting it to be improved and so on. However, here we are in the middle of an emotional discussion about how much change for good is the minimum requirement to be called an improvement. That is exactly what happens when people is biased, and it does not really help either of us.

 

[ashendashin]

Thx Infixo,

I just wanted some confirmation that the AI is not an inherent misconception or shallow or lazy implementation as one would conclude reading the so common “AI s*cks” type of threads but rather there exists a set of quite specific bugs or sometimes just observation bias mostly from those who don’t like RP elements.

That's quite the conclusion you got there(I think @Infixo himself said that the oft claimed 'RP' part isn't a thing. Correct me if I'm wrong). Care to explain your reasoning as to how you concluded that the presence of tactical assessment means that not only that particular area is either fully or efficiently utilized (with the exception of specific bugs), but also representative of the overall state of the AI? The concepts are all well and good, but then comes the implementation. 'mostly works' doesn't necessarily mean that it works well, but it's not as if everything needs to abandoned, I think.

 

[bitula]

That's quite the conclusion you got there(I think @Infixo himself said that the oft claimed 'RP' part isn't a thing. Correct me if I'm wrong). Care to explain your reasoning as to how you concluded that the presence of tactical assessment means that not only that particular area is either fully or efficiently utilized (with the exception of specific bugs), but also representative of the overall state of the AI? The concepts are all well and good, but then comes the implementation. 'mostly works' doesn't necessarily mean that it works well, but it's not as if everything needs to abandoned, I think.

I am addressing generic statements of “AI s*ck’s, braindead etc.” I see no proof for that. And in general universal statements about AI are useless, because they are subjective. Now, the inability to handle units as a group (or army) is by far the most common complaint. Now it turns out there IS code for that. So now it can’t be said that the AI is lackluster in this domain. Seems no one cares however, people will again-and-again complain that this behavior should be added (as feature) instead of posting a screenshot with some specific bugs, like look, despite team composition parameter set to this and that look here is an army with only catapults. Or despite there is a min army size, here a lonely archer attacks a city etc, etc.

 

[darko82]

I am addressing generic statements of “AI s*ck’s, braindead etc.” I see no proof for that. And in general universal statements about AI are useless, because they are subjective. Now, the inability to handle units as a group (or army) is by far the most common complaint. Now it turns out there IS code for that. So now it can’t be said that the AI is lackluster in this domain. Seems no one cares however, people will again-and-again complain that this behavior should be added (as feature) instead of posting a screenshot with some specific bugs, like look, despite team composition parameter set to this and that look here is an army with only catapults. Or despite there is a min army size, here a lonely archer attacks a city etc, etc.

People don't need to post screenshots for something that every child is aware of. Come on! Let's just not try to be blind about obvious things. And no, it does not only concern the "tactical AI". This one is the most visible one, simple as that. And proof is visible in every game... man... not sure what you are talking about?

 

[bitula]

People don't need to post screenshots for something that every child is aware of. Come on! Let's just not try to be blind about obvious things. And no, it does not only concern the "tactical AI". This one is the most visible one, simple as that. And proof is visible in every game... man... not sure what you are talking about?

What is your standpoint? I don’t get it. You think that AI is braindead because of lazy/lackluster, badly thought out implementation/bad design? Or because of series of bugs? Just telling you than in the latter case you can’t call it braindead. That was the whole point of what I was saying. Or you think it is the first? Well in that case you have no proof for that, it’s just in your imagination. That was the second main point I was talking about. I heard never any proof (or discussion) regarding major design flaw in LUA scripts or DLL code.

 

[ashendashin]

I heard never any proof (or discussion) regarding major design flaw in LUA scripts or DLL code.

DLL ain't open in Civ6. Scripting bugs bordered on silly spelling errors that went ignored for quite a while.

Here's something that I can be bothered to repeat again; many responses to Civ6 AI complaints said it was the same case for the whole series. With this in mind, one can compare the quality, if not the exact technicality, to Civ5. The AI does, in fact, contain major flaws and even missing functionality in Civ5, such as making defensive pacts. To go over the entirety of why and how such poor functionality came to be is pointless to users who can only see that it clearly sucks.

I think the vast majority of fixes and improvements to the AI in VP were summed up as such in changelogs. You get things like improved AI building selection or better embarked unit management, but the technicality of such things is largely driven by active development, not entirely reported bugs in a static system. So even with the assumption that all the necessary functionality is there for Civ6, that does not in any way prevent one from claiming the product of those systems braindead/lacking. I guess what you should be doing in a case that's as poor as Civ AI is to show examples of it working well, beyond not crashing.

 

[darko82]

What is your standpoint? I don’t get it. You think that AI is braindead because of lazy/lackluster, badly thought out implementation/bad design? Or because of series of bugs? Just telling you than in the latter case you can’t call it braindead. That was the whole point of what I was saying. Or you think it is the first? Well in that case you have no proof for that, it’s just in your imagination. That was the second main point I was talking about. I heard never any proof (or discussion) regarding major design flaw in LUA scripts or DLL code.

You know, we could talk about those bugs when the game was released. How much time is that? How much has been done to fix any bugs in this respect? I cannot talk in terms of "bugs" because of that. Most likely those are the bugs that will never get fixed by FIriaxis. They could get fixed by the community if they let the modders to do so... Their policy really hurts. For me, they should be ashamed because of that. That's all.

 

[tab]

People don't need to post screenshots for something that every child is aware of. Come on! Let's just not try to be blind about obvious things. And no, it does not only concern the "tactical AI". This one is the most visible one, simple as that. And proof is visible in every game... man... not sure what you are talking about?

This. Some people just wont accept that the building is on fire cause they dont see any smoke. We need to band together on this issue and let firaxis and all the other developers know that selling unfinished game is not OK. This DLC, micro transaction bull**** is ruining the gaming industry and to top it all off we have people who are defending this behavior.

 

[bitula]

DLL ain't open in Civ6. Scripting bugs bordered on silly spelling errors that went ignored for quite a while.

Here's something that I can be bothered to repeat again; many responses to Civ6 AI complaints said it was the same case for the whole series. With this in mind, one can compare the quality, if not the exact technicality, to Civ5. The AI does, in fact, contain major flaws and even missing functionality in Civ5, such as making defensive pacts. To go over the entirety of why and how such poor functionality came to be is pointless to users who can only see that it clearly sucks.

I think the vast majority of fixes and improvements to the AI in VP were summed up as such in changelogs. You get things like improved AI building selection or better embarked unit management, but the technicality of such things is largely driven by active development, not entirely reported bugs in a static system. So even with the assumption that all the necessary functionality is there for Civ6, that does not in any way prevent one from claiming the product of those systems braindead/lacking. I guess what you should be doing in a case that's as poor as Civ AI is to show examples of it working well, beyond not crashing.

Well exactly. As I mentioned before, there are cases when AI performs exceptionally good. Like I’ve seen early conquest of my city with close to perfect coordination of AI movements and attacks. Some just tend to ignore all these moments, since seemingly a single bug makes some people so angry that they tend to forget the parts which they otherwise enjoyed and then make broad statements about the catastrophic state of AI, which latter part is the problem (the conclusion), the former part is understandable. Again, you have no proof that the AI is braindead, only a modder can have oppinion here, since AI is a piece of code and cannot be judged by your observations of things happening on your monitor until it is not bug free. Also one cant make too much conclusion from LUA alone, so until the release of DLL AI overall quality cannot be judged. And there is no such thing as defensive pact in the game, so I don’t get it why should AI bother with it.

 

[tab]

so until the release of DLL AI overall quality cannot be judged.

What? So the actual execution of the code is not an indication of its actual quality?

Assuming it was a single bug that caused the AI to go from Geogre Patton in Ancient era to kamikaze Patton in all other eras, then why hasn't this bug been fixed after so many patches and DLCs. This is much bigger than a single bug which is exacerbated by the non-seriousness/unwillingness of the developers toward this issue.