Opinions on the "AI"

[Aristos]

Civilization VI is not a strategy game, nor simulation game or immersion game.

It is a dopamine loop system of addictively filling buckets and having a constant sense of futile accomplishment, where the player needs to be rewarded and cannot be punished.

It is afraid to provide drawbacks and negative modifiers, to the point of supposed "dark ages" having their own set of bonuses. The expansion is called Rise and Fall but you cannot fall. That would be unpleasant sensation.

Why invest in such things as capable AI? It's a difficult investment in a system which has no shiny buttons, sexy advertisements and worst of all: it is potentially detrimential to the casual pleasure.

Some part of the playerbase finds it alluring, to the point of insane praise give to how in civ "there are so many ways to win".

Yeah, because there is no way to lose. There is nothing at stake. There is no drama of challenge and joy of victory.

That's here we are: at a shallow, colorful pleasure generator; a board game where almost every move is a winning move and choices don't matter.

After all, games such as Darkest Dungeon, Dwarf Fortress, Dark Souls, Divinity: Original Sin or Kingdom Come: Deliverance have proven that challenge is detrimential to the satisfaction.

Also, programing good strategy game AI is impossible and Vox Populi is fake news hoax conspiracy of frustrated haters to put Firaxis in a bad light. Such mod never existed.

Sadly, you are right, and probably in more ways than you think right now (or maybe you are thinking the same as I am but not saying it?). I know some people may "feel" offended by this, but hopefully we can argue rationally without making this personal. The reality seems to be that we now live within the "Instant Gratification" era (Dark or Golden, your choice...). This core "attribute" permeates everything, from cars to fashion to attitudes to, of course, video games. In that sense, ironically, Firaxis can be "forgiven", as no one in their right mind can ask them to go "against the audience" and produce something that the majority will dislike. That is why we get "more shiny" and "less substance", not only with Civ post-Four, but with anything and everything else. Sad, but true.

In other words, we get what we deserve/want as an audience, even if not all of us fit that description of audience. It reminds me of an old, wise Spanish saying: "Cada pueblo tiene el gobierno que se merece" (Every society gets the government it deserves). The same is true for any market, or product: Every audience gets the product it deserves.

That is, from my point of view, why we get a parody of an AI.

 

[Victoria]

Every audience gets the product it deserves

One feels the slightly incorrect, to me it’s more the audience gets what is the popular vote.
It’s a global thing, money both walks and talks and to be fair the quality of games now takes investment, even the rare sparks of genius succumb.

I’ll take out of this what I can and many here do, you don’t have to play in anyway the thought police want you to play, I’ll play an immersive target and really enjoy the game and will find challenges along the way. I’ll play games comparing play on the same map as others to better myself with the primary aim of a lower turn win.
I love the fact we are all in the same audience.

 

[Aristos]

One feels the slightly incorrect, to me it’s more the audience gets what is the popular vote.

Agreed. Somewhat a democracy, say the Wallet Democracy, with all its goods and "bads", the "bads" being that there is no "qualified" vote... so, effectively, the "minority" gets whatever the majority votes for, be it a lesser video game or a Prime Mistake as leader of a country... (although in some strange cases of questionable systems, we get what the minority voted for)

If we could only get our hands on the dll code, but that also is far away, at least one expansion-one year away, if at all...

 

[hdbhdbvGFG541]

I don't think it is the audience.It is simply a problem that all the previous civ designers are gone.
Soren has his own game studio,Shafer is gone (which is a good thing ) ,Reynolds is working on mobile games and Beach was a never a lead but more of a fixer for expansions.
Most modern designers grew up on mobile and casual games and have no idea how to design complexity. I mean just look at beyond earth and rise and fall,a mishmash of good and bad ideas not clicking together.
Oh and it certainly doesn't help that they choose a random nobody for ai engineer position.

 

[kaspergm]

I’ll take out of this what I can and many here do, you don’t have to play in anyway the thought police want you to play, I’ll play an immersive target and really enjoy the game and will find challenges along the way. I’ll play games comparing play on the same map as others to better myself with the primary aim of a lower turn win.

I agree with this, I would love for the AI to be more capable, but its inabilities does not prevent me from enjoying the game. Sure, playing the game as a strategic challenge does not make much sense, because you can always role over the AI should you want to. But as an immersion/roleplaying game it gives me a lot of enjoyment, there is enough variation between the civs and the random map rolls that it offers excellent replayability, and there are enough different objectives in the game to make it feel generally fresh on replay. It would be an even better game if AI was more of a thread, but at the end of the day, I still prefer this to the "who has the bigger ... (stack)" days of Civ4.

 

[KyivanRusCivPlease]

I agree with this, I would love for the AI to be more capable, but its inabilities does not prevent me from enjoying the game. Sure, playing the game as a strategic challenge does not make much sense, because you can always role over the AI should you want to. But as an immersion/roleplaying game it gives me a lot of enjoyment, there is enough variation between the civs and the random map rolls that it offers excellent replayability, and there are enough different objectives in the game to make it feel generally fresh on replay. It would be an even better game if AI was more of a thread, but at the end of the day, I still prefer this to the "who has the bigger ... (stack)" days of Civ4.

What do you mean by roleplaying civ? Is it just a game with no intention to win?

 

[Victoria]

Win is just 5 targets made by the designers, you can make a new win in your mind without modding the game.
If you walk away from a game saying you have enjoyed it you have won.
If you get a hollow victory that did not feel like you enjoyed yourself I would not consider that winning.

 

[LobstaBoy]

Some earlier suggestions on how to increase the difficulty ( Deity, no restarts/save scumming, picking "bland" civs, larger maps)
all have merit . Of course multiplayer exists.
Still , your not wrong . The game is too easy and even doing all the above,if you have any skill at all you will eventually catch the AI.
Like others I try to give myself a real challenge by having goals.
Examples: win without declaring war, win faster by DOW EVERYONE the turn I meet them, build a specific wonder that doesn't do squat for my VC, conquer the world with fewer units, etc. Whatever works for you.
See, that's how I have fun, but mileage will vary.
You might hate all the above and feel this game just isn't challenging enough regardless and that is perfectly fine.
Meanwhile I'm dying to get home to salvage my conquest game that has bogged down due to this AI having the nerve to actually build units!!!

 

[acluewithout]

I mostly play to explore the mechanics of the game. Sometimes that is by creating some arbitrary goal, but mostly it's trying to achieve one of the existing victory types but deliberately using / not using some mechanic, or working along some specific strategy. I had a lot of fun recently playing Norway, and leaning into the early exploration with longboats. (On that note, I can confirm that this is not an effective strategy for a super fast science victory; and I can also confirm that it is a tonne of fun.)

At base, I think I play the game as if it was the strategy game I want it to be. So, I usually found a religion, or settle on other continents, even though I think the upside to doing these things is limited. Instead, I do these things because in a strategic game that plays out over the course of human history, I feel like that these are something which should matter. As I result, I generally play fairly efficiently, and have a good time doing so and end up with a fairly immersive empire, but only by playing within certain rules and objectives I set for myself. I hope that makes sense.

Which brings us to the AI. So, yeah, I wish the AI played better. I dislike it can't settle or build cities sensibly, and I dislike it can't build an army, get the troop mix right, or mobilise its forces. But I can live with it as is.

At some point though, the fact the AI isn't great actually does impact my ability to explore some mechanics. Air combat is a good example. You just can't play around with the defensive parts of air combat, because the AI never attacks you with planes. So, that part of the game is just a dead letter. It doesn't really bother me, because I'm not that interested in the air combat mechanics. But over time I feel like I'm finding more and more areas where the AI stops me really exploring what the game can do.

I don't know. I'm fine with the AI. I'm hoping R&F improves it one way or another, or it gets better with patches or mods. I'd been thinking the AI might do better with just some free units as eras progress or when certain events happen (ie rather than just at the start of the game), and see there is now a mod which sort of does that. Eventually I'll try that out. The AI is fine, I guess, and will probably get at least a bit better over time. But at some point, if it doesn't get better, I might get bored of this game sooner than I might have otherwise.

 

[JacquesBquick]

There are certainly some hyperbole i think in this discussion, but overall I do think its a pertinent conversation. Particularly what I key in on are two points:
1) choices feel meaningless - I agree that many times i don't feel like I have tough choices to make. a majority of social policies are still garbo in 90% of situations, or at best "i'm going for culture victory i will use the culture card" type of deal. building wise you have had only two choices between museums and barracks/stables. Government plaza certainly adds here, but it still feels underdeveloped, and falls into the same bucket as social policies, these don't feel like meaningful choices. 7 governors to choose from? what choice, really. free food and production from magnus > free culture and science from pingala > everyone else. part of the problem here is also that the game, still, is unbalanced in terms of production relative to everything else. I'm no game designer, but production is still king and its not close. I get in a mathematical game like civ its impossible to avoid one thing being the most valuable, but more could be done to close the gap.

2) the AI doesn't play over half of the game's content - this is more to the area of how the thread first started. I am not asking for Deep Blue AI here, but I am asking for the AI to play each area of the game. They don't. when i see GPPs at +3 per turn when i'm at +48 in mid to late game, i see an AI that just doesn't play the game. When i see harald praising my strong navy of 2 battleships, i know the AI is not playing the game. When I see vast stretches of farms and one district in a capital or one of the AI's early cities I know they aren't playing the game. When i see obvious adjacency bonuses overlooked by the AI I know they aren't playing the game. The AI doesn't understand that food matters most in the early turns only and that production is king after. The AI doesn't understand how to leverage AoE industrial or entertainment districts. The AI doesn't understand how to use a navy or air force in even a mediocre manner. The AI doesn't understand how to position units to effectively take a city . The AI doesn't know how to handle walls. The AI doesn't know how to move units. The AI doesn't know how to collect or theme great works. The AI agendas are next to impossible to manage, and there are some clear and obvious gaps in many agendas and their effect on regular play of the game (Alexander giving you a thumbs down after you declare peace WITH HIM? glitches like that really ruin the experience). If just one or two things were wrong with the AI, It wouldn't be so bad, but with so many things wrong with it, winning on a high difficulty doesn't feel like an achievement, it feels like an inevitability. for the record, I do not do reloads or quicksaves or anything unless I do something incredibly boneheaded like fat finger a district placement. If i lose a wonder i lose it. If i lose a lot of units to my bad positioning, i take the L. But once you're in the medieval era, that just doesn't happen.

 

[Victoria]

the AI doesn't play over half of the game's content

I am fairly sure this is down to neither incompetence or accident.
It is primarily money motivated, patience and the game will get harder.

 

[Stringer1313]

How about these easy fix for the AI, that doesn't require impossible Watson-like programming:

- Make more elaborate and challenging victory conditions -- while this will make it even more unlikely that AI will achieve victory, it will at least make it harder for you to achieve victory, which is all that matters. Perhaps when going for culture victory, when you have achieve 50% of the tourist goals, you have to decide whether to gun for becoming the culture capital of the world, which you have to do to achieve victory, but if you do it, then all your unit production is reduced by 75%, all unit maintenance costs skyrocket (or skyrocket if you have too many, forcing you to delete units), or something like that? And everyone in the world knows when you've made that decision, and will know that you are ripe to be attacked.

- (Plus the era-difficulty-scaling that everyone suggested).

 

[Victoria]

I always liked the idea of adapting the game difficulty to the person playing, that’s just what the adaptive difficulty mod by @FearSunn has done. Still fairly new, give it a go.

 

[Eliminator_Sr]

The best game studios put major effort into even the smallest components of the game. It's about overall experience and IMO it's a very bad decision by Firaxis to neglect such a critical component of the game for so long. The AI has gotten better yes but in such small increments that it's barely noticeable. The early game is still somewhat challenging so I think they are close to having the balance right there, but once you get enough cities down the AI is laughably bad. Is it so much to ask that the AI at least remain somewhat competitive on the highest difficulty levels with the insane bonues they get? I've been playing civ since the original release and this is hands down the worst it has ever been. I remember from civ 1 that the AI would sometimes have battleships before they hit 1 AD on emperor. At this point I have given up on Firaxis to address this so I think it's going to be up to modding community.

I am not asking for Deep Blue AI here, but I am asking for the AI to play each area of the game. They don't.

This is exactly right. There is no semblance of a threat from the AI once you hit the medieval era and AI behavior seems to follow no coherent strategy at all except maybe to intentionally suck which is the only explanation that makes any sense to me now.

 

[Aristos]

2) the AI doesn't play over half of the game's content - this is more to the area of how the thread first started.

That is, indeed, the core problem. This fact creates a sort of hollow middle ground; at lower difficulties, the smaller bonuses may extend the AI game in order to allow for some role playing, but there is no opposition, fake or real, at all. At higher difficulties, we get some fake opposition in the form of huge bonuses for the AI, but as it cannot or does not know how to use a good portion of the systems, offensively nor defensively, the game also becomes trivial after a start of catching up. Worse, it's impossible to role play with a "no victory condition" mentality either at higher difficulties, because the huge bonuses will make some AI stumble across a random victory if you let it. Thus, you don't let it, but then that AI does not offer any resistance (again, my example of the good use of super-spies against which I have never seen any AI in Civ6 resist with success... I can just solve any and all problems to win by using well promoted spies).

The bottom line is, the game is fun in some aspects, but not challenging, and any victory at any difficulty level becomes as trivial as almost expected. And no, it's not only a matter of money (I am sure all the shiny things are more expensive than a good AI).

 

[kaspergm]

What do you mean by roleplaying civ? Is it just a game with no intention to win?

No I play to win, I just don't necessarily play to win as fast as possible. Most notably, you can pretty much always get the fastest win by going all-out domination. Thus, in order to have a meaningful (non-domination) game, you need to accept that you don't necessarily DoW the AI, even if this would be the (most) opportune move*. I like the empire building aspect of it. Making my cities grow, developing districts, building wonders (I *love* building wonders). I do play to win, and when the time comes, I'll try to win as fast as possible - I will buy great people to boost space production etc. - but I'm not playing to win at the lowest turn number possible, that has no value for me.

* Just to clarify on this: This is not saying that I never go to war. I do go to war, particularly early game, if I want the lands of a near neighbor, or if there's another good reason. But if I have a stable diplomatic relationship with an AI, I will not flip on them and DoW them just to win as fast as possible.

 

[Bitterman]

I always play with Smoother Difficulty mod enabled and/or in multiplayer.

AI right now is very flawed, but then again I don't remember Civ V AI being that much better honestly. I have the certainty that it will eventually get better. And even if this is not the case we'll always have the mods.

There is also the thing (at least for me) that this game is not so much about winning, but about just having fun and a good time. Roleplaying is a big thing for many people.

 

[Unconquered Sun]

Any determined player with a grade school education should be able to *beat* the game.

Sure. The ubiquity of grade school degrees nowadays is the reason Civ 6 seems so easy.

Compare to years ago, in the waning days of Civ 4, only a handful of people had grade school education. That led to screens like this.

Go back a couple more years to the time of Civ 4:Warlords, and it's a veritable Dark Age. Grade school education wasn't yet invented!

 

[orasis]

Yes the A.I is bad, I use mods.

That being said... I was playing 4 (I HATE 4 - too much strategic resources. Why even have a 'balanced' option at all? Give everyone even more??) and I must say, hate aside, the A.I is much, much better in that game without any mods.

My problem is that I do not understand the problem... Barbs are pretty clever and pose a pretty good threat throughout most of the game. So, why not just use that 'method' with war-time 'real civ' AI. I don't get it.

 

[acluewithout]

I don't think I really understand the problem either. I don't have any background in computers.

Part of the issue may be that the AI doesn't build enough units or the right units. My guess is that, in part, it's because the AI can't anticipate future needs.

For example. I was tooling around with Japan recently. Emperor, continents, peaceful game. I was surrounded by mountains and city states. But I had one fairly open border with Russia, who really didn't like me - I'd actively managed my relationship by selling luxes, but I had built close to him so not so friendly. Peter had a tonne of horses. Looked like he was mobilising. I was wanting to play around with spears anyway, so when things started going south, I put two spears into my border city along with my three archers and build walls. Spears and archers were ready to be upgraded when needed. (...of course, no rush ever came...)

My point is, the AI can't do things like that. I'm not sure it's reasonable to expect it should.

I think it would help if the AI got free units in certain situations (both offensive and defensive), instead of just at the start of the game. The trick would be to give the AI units which reflect what it should have been building in the first place, had it been actually planning ahead, so that the bonus units don't feel like a boondoggle. Of course, if the AI was smarter, that wouldn't be necessary. But I think extra units etc. would be easier to implement.

The other issue is where the AI settles, what districts it builds, what improvements it makes, and where it places certain situational wonders. It's more of an eyesore than anything else.