[GS] Yet (Not) Another AI Thread

[oSiyeza]

It's always been like this. It's just a lot easier to develop- you'd have to change many versions, debug them all, you'd need QA testers who could actually handle deity (most of the devs play prince-king) etc.

I disagree, this is actually how you do it. You have the same AI for all difficulties, but you just disable some branches of the behaviours for lower dificulties.

That is, in higher difficulties the AI gets to use more advanced behaviours. But you not using a different AI, you are using the same one and you only need to QA the specific branches that only advanced AI uses.

This is not so hard to implement, and a lot of games use already this strategy. You can add AI bonus on top of that, and even just increase chances of some agressive behaviors or some counters to player strategies.

And there you got it, an AI that gets better with higher difficulties and with only a fraction of the cost of a new AI.

 

[AsH2]

I disagree, this is actually how you do it. You have the same AI for all difficulties, but you just disable some branches of the behaviours for lower dificulties.

That is, in higher difficulties the AI gets to use more advanced behaviours. But you not using a different AI, you are using the same one and you only need to QA the specific branches that only advanced AI uses.

This is not so hard to implement, and a lot of games use already this strategy. You can add AI bonus on top of that, and even just increase chances of some agressive behaviors or some counters to player strategies.

And there you got it, an AI that gets better with higher difficulties and with only a fraction of the cost of a new AI.

I agree. It's like they (devs) took the lazy way and just throw in AI from 5 without adapting it to the fundamental changes in 6. Eventually they patched in expansive behaviours and added the loyalty mechanism to (better) control it. They should have seen that and made it happen before release.

 

[Gedemon]

I disagree, this is actually how you do it.

But AFAIK no-one has ever done a civ AI which was good enough to make this method viable, even modders with unlimited development time.

Yes, the AI can be better in civ6, is already better than 3 years ago and will be even better in a few years, especially if the source code is released.

But good enough ?

And if it was, wouldn't be better for new players to experiment a "good" AI, to learn the game, with the good old bonuses mechanism to compensate under prince level?

 

[Kaan Boztepe]

most people want easy games and the rush they get from "winning". good AI or ridicolous bonus systems are really only for people who want to challenge themselves who are so in the minority that no gaming company could justify costs to improve it to a challenging level to their investors due to lacking returns in monetary terms. Modders are my only hope.

 

[oSiyeza]

But AFAIK no-one has ever done a civ AI which was good enough to make this method viable, even modders with unlimited development time.

Oh, I think you just missunderstood what I was trying to say. Having more advanced AI behaviours for different difficulties does not require nor warrants to have a good AI.

To do this, you just need to take into account the difficulty level in the AI codding. I just pointed that to code the AI like this, with the kind of AI of civ, does require extra resources, of course. But it would cost only a fraction more of the resources of the current AI.

Fxs did not do it, cause simply they preferred a slightly simpler cheaper aproach. And maybe the additional resources required were not even the reason for that decission. They could have considered and discarded the option to make Civ VI more similar to previous Civ games, expecting that more changes would be disliked by the fanbase.

In my opinion, and this is only speculation. They did just put the AI just so low on the list of dev priorities, that the people in charge ended with no time or room for innovation.

I simply was saying that the point: "Making a harder AI for higher difficulties is insanely expensive and reasonable players should not expect this to happen" does not hold ground at all.

most people want easy games and the rush they get from "winning". good AI or ridicolous bonus systems are really only for people who want to challenge themselves who are so in the minority that no gaming company could justify costs to improve it to a challenging level to their investors due to lacking returns in monetary terms. Modders are my only hope.

This is in my opinion not true at all. You see, Fxs made a bet on how the budget would be assigned to the game, trying on one hand to match the budget distribution that would benefit the sales the most. And on the other hand, and despite all their flaws, I think money was not the only concern. As to me is obvious that a lot of care and love has been put into the game. We all know that Fxs is quite far to the worst standards the industry has to offer.

Also, I think it is simply not true that the average player does not care for the AI or that Firaxis thinks that the AI is not worth investing in. Why? Cause most players of Civ play still in single player, and most of the bad critics of the game have to do with the AI and the UI. Fxs knows this and they know good reviews are better for sales than bad ones. Also they have continuouslly showed a will to improve both aspects of the game to the point they are the two single things that were touched in every single one of the patches the game has received till now.

These kind of comments also are totally not helping anyone. They do a big disservice to players like me (and I think the majority of civ fans) that have no interest in "challenging ourselves" and do not enjoy Deity, but want a better AI and a fun challenge. And tells firaxis we are not the fans they should care about.

We should expect, demand and ask for a better AI if that is what we want. And, as far as I know, this is the only single issue the entire community here agrees upon.

Not having expended more resources in the AI of the game was simply a mistake, and we as fans need to remind Firaxis that we expect and deserve better.

 

[Kaan Boztepe]

I would love to have a good AI. I agree with you and the others on this forum that we should ask for and get better AI but I also believe it is a disservice to discuss why we did not get it. the first step to solve any problem is to idetify the problem.
Only 42% of players have actually won a game ( including settler difficulty ) only whereas only 5% did finish a deity game ( including in that 5% are also i
imagine are those that did the 1st turn score trick to get the achievement ) .

 

[oSiyeza]

I would love to have a good AI. I agree with you and the others on this forum that we should ask for and get better AI but I also believe it is a disservice to discuss why we did not get it. the first step to solve any problem is to idetify the problem.
Only 42% of players have actually won a game ( including settler difficulty ) only whereas only 5% did finish a deity game ( including in that 5% are also i
imagine are those that did the 1st turn score trick to get the achievement ) .

Let me play the devils advocate here for a while.

What if a significant % of the players have not finished a game cause they simply got bored before the actual end? I know in most of my games, once the only thing left to do is end turns, I just start another game.

That number actually shows that a big problem for the game is that it fails to engage a lot of players in a long term. Which, given the XP, DLC, nature of the game and it being a franchise. It must be a big concern for Firaxis.

Therefore, you can use those same numbers to argue that a stronguer AI should be a very high priority for FXS right now....

And this is why we should not put on the players (and therefore on ourselves) the reasons why the AI is not better. And even less argue that this is the right decission from Fxs point of view.

 

[Kaan Boztepe]

i think firaxis tried "to solve" that problem by introducing the hall of fame feature. much cheaper and handled with just an UI upgrade. psychologically speaking of course.

 

[GrumboMumbo]

I completely agree with what @oSiyeza is saying here.

I also want to add that even lower level players playing at the easier difficulties notice bad AI and then get frustrated with the game as a whole. Obviously they aren't going to notice everything, but in my experience one of the most obvious problems with the AI is how they communicate with you. The AI communicates in a very blunt, often Schizophrenic way. Now, I am not expecting very advanced behaviour here, but in my opinion it needs to be more subtle.

Harald of Norway:
I like boats
I like boats
I like boats
You don't have enough boats
I like boats
I like boats

Tamar:
I like walls
I like walls
I like walls
You have nice walls....

I would like somebody to come up with a compelling argument explaining why this kind of dialogue is fun for a player of any experience level. Surely even a new player plays one game of this and goes, 'Wow this game is stupid' time to uninstall...

 

[Lord Shadow]

I simply was saying that the point: "Making a harder AI for higher difficulties is insanely expensive and reasonable players should not expect this to happen" does not hold ground at all.

It holds plenty of ground, because it doesn't make any sense to make a better AI for the benefit of the less than 15-10% of the playerbase which plays on the higher difficulties.

AI needs to be better across the board. Every difficulty level and player segment would benefit from it. This doesn't mean the utopian end of AI "cheats" (which is a ridiculous proposition to begin with), however.

However, as I keep saying, improving what exists to a meaningful extent requires a disproportionate amount of resources for the percentage of the playerbase which would truly appreciate it. It's not a very sound investment, and it goes somewhat against the Sid Meier school of thought, which sees AI as a kind of supporting character for a protagonist player's entertainment, as opposed to something resembling a real antagonist. More an illusion of challenge than an actual challenge, unfortunately for many of us.

 

[oSiyeza]

It holds plenty of ground, because it doesn't make any sense to make a better AI for the benefit of the less than 15-10% of the playerbase which plays on the higher difficulties.

Oh please, give me a break. .

You cannot simply argue that implementing a better difficulty system does not make sense cause not a lot of players use the current badly implemented difficulty system.

These reasons are totally backwards, and again. Why so many people feels entiled to say that Fxs, should indeed not improve the game. Cause the players (we) are not gonna care anyway.?

I will say it again, we know that here most people cares about it, because we say it constantly. We know the reviews of the game care about it cause they usually speak of these points.

Why then insist in repeating that Fxs should indeed not care? Why keep undermining our goals? And why do it in the name of the players when we know as players that this is not what we think and also when we know nothing about why other players are not so engaged with the game as we are? Maybe it is cause they find it boring due to the lack of a difficulty progression and a decent AI? Naaah players dont care about AI and difficulty and therefore, Fxs should not care either...

However, as I keep saying, improving what exists to a meaningful extent requires a disproportionate amount of resources for the percentage of the playerbase which would truly appreciate it.

Yeah, saying Fxs should not spend money to give us what we want cause we are not the customers Fxs should care about, really helps too. Furthermore, this is completely false.

You dont know how much investment this requires. I explained this should only be a fraction of the current AI cost, that is probably less than 1% of the cost of the game.

Second, you dont know how many players a better game would attract. This is insane, why not then say: oh you know, how many players really enjoy nowadays the Civ soundtrack?, today an average fan will probably only care about dubstep, why bother? How many civ players know the real name of the cities of old civilizations? Why bother? How many players actually finish a game with diplomatic or cultural victories? Why bother? ...

Definitely the message we want to make across here is Fxs should add less content, and care only about the numbers... (Sarcasm alert)... It looks to me that this is the point of many people here, though.

I really dont get it. Why people here keeps doing this? Why keep validating a game design that is the opposite of what we want with reasons that are obviously wrong? Could this be madness?

If it is, I hope is only Christmas madness.

 

[Gedemon]

Oh, I think you just missunderstood what I was trying to say. Having more advanced AI behaviours for different difficulties does not require nor warrants to have a good AI.

To do this, you just need to take into account the difficulty level in the AI codding. I just pointed that to code the AI like this, with the kind of AI of civ, does require extra resources, of course. But it would cost only a fraction more of the resources of the current AI.

Fxs did not do it, cause simply they preferred a slightly simpler cheaper aproach. And maybe the additional resources required were not even the reason for that decission. They could have considered and discarded the option to make Civ VI more similar to previous Civ games, expecting that more changes would be disliked by the fanbase.

In my opinion, and this is only speculation. They did just put the AI just so low on the list of dev priorities, that the people in charge ended with no time or room for innovation.

I simply was saying that the point: "Making a harder AI for higher difficulties is insanely expensive and reasonable players should not expect this to happen" does not hold ground at all.

The point is that making a harder AI even without trying to make it modular is time consuming thus expensive.

Oh please, give me a break. .

You cannot simply argue that implementing a better difficulty system does not make sense cause not a lot of players use the current badly implemented difficulty system.

These reasons are totally backwards, and again. Why so many people feels entiled to say that Fxs, should indeed not improve the game. Cause the players (we) are not gonna care anyway.?

I will say it again, we know that here most people cares about it, because we say it constantly. We know the reviews of the game care about it cause they usually speak of these points.

Why then insist in repeating that Fxs should indeed not care? Why keep undermining our goals? And why do it in the name of the players when we know as players that this is not what we think and also when we know nothing about why other players are not so engaged with the game as we are? Maybe it is cause they find it boring due to the lack of a difficulty progression and a decent AI? Naaah players dont care about AI and difficulty and therefore, Fxs should not care either...



Yeah, saying Fxs should not spend money to give us what we want cause we are not the customers Fxs should care about, really helps too. Furthermore, this is completely false.

You dont know how much investment this requires. I explained this should only be a fraction of the current AI cost, that is probably less than 1% of the cost of the game.

Second, you dont know how many players a better game would attract. This is insane, why not then say: oh you know, how many players really enjoy nowadays the Civ soundtrack?, today an average fan will probably only care about dubstep, why bother? How many civ players know the real name of the cities of old civilizations? Why bother? How many players actually finish a game with diplomatic or cultural victories? Why bother? ...

Definitely the message we want to make across here is Fxs should add less content, and care only about the numbers... (Sarcasm alert)... It looks to me that this is the point of many people here, though.

I really dont get it. Why people here keeps doing this? Why keep validating a game design that is the opposite of what we want with reasons that are obviously wrong? Could this be madness?

If it is, I hope is only Christmas madness.

A lot of people want a better AI. Some of them are even ready to pay more and wait more to get it. But it doesn't mean that it's worth the development time for Firaxis.

That's what people keep saying in those threads: "don't expect it from them, because X or Y".

They're not saying they shouldn't do it, they're explaining why they don't and won't.

Or maybe the survey they've made this year has indeed shown them that the player base have a higher interest in a better AI than new shiny (or beardy) leaders, but I doubt it, and IMO the best chance to get a better AI is modders and that's one of the reasons that we're waiting for the source code.

 

[oSiyeza]

They're not saying they shouldn't do it, they're explaining why they don't and won't.

I think that if the hardcore fan base that wants a feature is also saying that is not a profitable feature and 99% of the players dont want that feature and therefore they expect to never have it... To me that is exactly the same as giving Fxs the msg "you should not do it".

Remember that this is the place where Fxs looks for what their most loyal fans want and demand. If we openly state that is not worth it for firaxis to give us what we want, do you think that they are going to care at all?

This is precisely the moment where, instead of demanding a better game we give ourselves the excuse to firaxis to never bother.

In what other place are the customers that need something, (that certainly is not too expensive for other companies to add), tell the company they buy from, that this something is too expensive to add just for them?

We dont know how much an AI costs, or how much money Fxs has, or how many people want this, or how much profit Fxs makes from the game, or how much profit in the long run investing in an AI will make for them. But still are telling firaxis that is not profitable for them to do what we want. It is pretty obvious to me that this thinking is the most reliable way of not getting anything.

Fxs now does not even need an excuse to deliver a poor AI, their most critical fans create that excuse for them. They know that if we dont espect a good AI, certainly no one else will. Yes, in my opinion these msgs are the same as screaming to them "dont bother, we expect a poor product, and will buy it anyway, so you dont have to put effort for us, we will fix it for free later anyway"

PS: I just realized that im a Civfanatics, warlord, so it only fits that i feel combatant about this!.

 

[Sostratus]

A lot of people want a better AI. Some of them are even ready to pay more and wait more to get it. But it doesn't mean that it's worth the development time for Firaxis.

That's what people keep saying in those threads: "don't expect it from them, because X or Y".

They're not saying they shouldn't do it, they're explaining why they don't and won't.

In what other place are the customers that need something, (that certainly is not too expensive for other companies to add), tell the company they buy from, that this something is too expensive to add just for them?

FXS is welcome to prove the idea wrong. No one would be upset if they announced work on this. They aren't totally oblivious, and they certainly know their hardcore fans would be happy about AI work. But I don't know if you are fully appreciating this:

The point is that making a harder AI even without trying to make it modular is time consuming thus expensive.

To make the AI "better" is an open ended goal that is really hard to plan out. The project manager is faced with sinking time+cash into something that may produce absolutely nothing. He has to justify to his boss why he's investing resources into something that is "good enough" already. Your exact arguments in this thread would also apply to questions like why don't they make unique building graphic sets for every civ, or redo all the art assets in 4k. After all, it's not like they couldn't, surely it would improve the game in some regard- and in those cases they actually know they could do it and how long it would take.

The monetization pathway is also murky, since the attitude of most people is that AI improvements are like bug fixes and should be patched in for free, so then FXS would need to sell it bundled with other content, at which point- why not just do the content standalone and save the cost.
We've seen in the past that they will make improvements to the AI, and they are often specific - make the AI use air units, make them not leave settlers unguarded, etc. Those kinds of things one can pretty clearly envision the planning process for. A Project manager could approve that without feeling uneasy.
More than the engineering work to make "better AI" happen, game companies need someone on the business side to figure out how to make better AI = more profits. Then this will happen automatically.

 

[Lord Shadow]

A lot of people want a better AI. Some of them are even ready to pay more and wait more to get it. But it doesn't mean that it's worth the development time for Firaxis.

That's what people keep saying in those threads: "don't expect it from them, because X or Y".

They're not saying they shouldn't do it, they're explaining why they don't and won't.

Or maybe the survey they've made this year has indeed shown them that the player base have a higher interest in a better AI than new shiny (or beardy) leaders, but I doubt it, and IMO the best chance to get a better AI is modders and that's one of the reasons that we're waiting for the source code.

Pretty much. Nobody is saying Firaxis shouldn't produce a better AI, but after Civ5, CivBE and Civ6, forgive me @oSiyeza if I'm not exactly hopeful. I'd be delighted if they were to prove us wrong, but they don't need us to "openly state" anything to be encouraged or dissuaded.

They can see the state of AI by themselves, and they're aware of their own resources. The fact they haven't done anything about it is telling.

I'd rather they released the source code so that people with plenty of free time and interest can tackle the issue in a far deeper fashion than they ever would.

 

[darko82]

Defending developers who did NOT create a functioning AI is simply stupid. There are many games that lost because of it. I can give you a few examples if you want.

Defending developers who did CREATED nonfunctioning AI is also stupid.

I want the AI not the best, but the one that can make any challenge or the one that does not totally break immersion so much.

CiV IV might have a duller AI, but I enjoyed it more on higher difficulties. I did not have so much problem with the AI as long as I saw it COULD ACTUALLY DESTROY ME. IT POSED A REAL THREAT etc.

So for me, the stupidness of the AI is not an actual problem, it seems. I enjoyed Stalin in Civ I who was cheating and was much overpowered more than CIV VI stupid AI who looks like a disabled person and any child can beat it. Most players do not want to see that they play against NO ONE. Whatever they choose and do in the game won't matter (choices don't matter etc.) It's the best thing to bury any game!

And that is what I have personally experienced in Civ VI so far. Choices do NOT matter because of the stupid AI. I can but I do not have to do this or that. No matter. Whatever you do, you win because the AI is so stupid, no challenge. After a few playthoughs, you just lose interest in the game. Is this even AAA strategy game that is worth mentioning these days? I am not sure anymore!

Every CHOICE SHOULD MATTER. That is what makes every game interesting.

And I personally think that CIV VI already lost because:

CIV VI TRIES TO BE SOMETHING IT IS REALLY NOT!

Maybe something people like me HOPED FOR!

Not this time, sorry.

 

[cvb]

I'd rather they released the source code so that people with plenty of free time and interest can tackle the issue in a far deeper fashion than they ever would.

I like this way of looking at it. Enthusiasts finetuning their factory produced hobby.

I did not have so much problem with the AI as long as I saw it COULD ACTUALLY DESTROY ME. IT POSED A REAL THREAT etc.

Yes! That is the essence: we all want to win (far) most of the games we play - (a problem for multiplayer mode!!).
But for the victory to be valuable, there must be ALWAYS some fear to lose. At least a tiny REAL possibility something might go wrong. We know, we will win most games - but still, THIS TIME, isn't the threat XXX very dangerous? Will we lose this time??

Better AIplayers will of course help with that. The rest can be done by frontloaded bonuses ("starting parameters"), but also by balancing the later game ("cheating"):

I enjoyed Stalin in Civ I who was cheating and was much overpowered more than CIV VI stupid AI who looks like a disabled person and any child can beat it.

 

[bitula]

I dont think Civ6 AI is stupid. It's just a bias that AI bugs (or game bugs affecting AI) are far more memorable, than normal or good behaviour. Unfortunately there are still lots of bugs in the game which also affects AI. Also AI is not difficult to implement, just time consuming. Yes, its really time consuming. I dont think AI can be programmed by more than one person, so there is no point or possibility to further invest in it, just to let that single person work on it (refactor etc), which maybe is not the case currently, cause he is doing something else.

 

[Lord Shadow]

Defending developers who did NOT create a functioning AI is simply stupid. There are many games that lost because of it. I can give you a few examples if you want.

Defending developers who did CREATED nonfunctioning AI is also stupid.

I want the AI not the best, but the one that can make any challenge or the one that does not totally break immersion so much.

CiV IV might have a duller AI, but I enjoyed it more on higher difficulties. I did not have so much problem with the AI as long as I saw it COULD ACTUALLY DESTROY ME. IT POSED A REAL THREAT etc.

So for me, the stupidness of the AI is not an actual problem, it seems. I enjoyed Stalin in Civ I who was cheating and was much overpowered more than CIV VI stupid AI who looks like a disabled person and any child can beat it. Most players do not want to see that they play against NO ONE. Whatever they choose and do in the game won't matter (choices don't matter etc.) It's the best thing to bury any game!

And that is what I have personally experienced in Civ VI so far. Choices do NOT matter because of the stupid AI. I can but I do not have to do this or that. No matter. Whatever you do, you win because the AI is so stupid, no challenge. After a few playthoughs, you just lose interest in the game. Is this even AAA strategy game that is worth mentioning these days? I am not sure anymore!

Every CHOICE SHOULD MATTER. That is what makes every game interesting.

And I personally think that CIV VI already lost because:

CIV VI TRIES TO BE SOMETHING IT IS REALLY NOT!

Maybe something people like me HOPED FOR!

Not this time, sorry.

Okay, you went full hyperbole. I'd say that while flawed, the Civ6 AI is far from "nonfunctioning".

It does need an extra kick, definitely, but then it also has a much more complex environment to deal with compared to Civ4, and especially Civ1.

 

[darko82]

It does need an extra kick, definitely, but then it also has a much more complex environment to deal with compared to Civ4, and especially Civ1.

As many people on plentiful threads have said about it:

We do not need new additions (complex things) if the AI cannot cope with them. (this also concerns other games out there, generally). Thus, they add things to the game without even fixing the old bugs, issues, problems, complaints etc. I know and understand - it's a business model. But if it's a business, why not to invest into the AI? Improve the quality, reputation of the company, Civ franchise?

Wondering around units letting themselves killed is a malfunction and huge immersion breaking, fun spoiling, for instance. Malfunction of planes or ships, too.

The game fundaments (rules) are against it.

Vox Populi might not be perfect, but that kind of level (even much lower than that) would satisfy most people here. And there would not be hundreds of threads about it.

And I will repeat.

- Every choice should matter (be important), especially in a strategy game.

- On higher difficulties, you should feel that you can be destroyed. Feel that the AI (opponent) is a real threat. Thus, every choice should matter if you want to win.

That would be a perfect CIV VI game to me because I do like CIV VI in general. Even more than CIV V.

And why would you go "complex" if it is not working as supposed to? I don't know if it is a good path for developres. I doubt it.

I think it's better to go simpler and make sure that things work as supposed to.

With this approach, who knows, CIV I REMAKE in 2020 could become the best CIV in the history