Whatever they do I hope the AI is substantially upgraded

I remember when my Militia used to spawn barbarian armies from goody huts! And they had to walk there using only 1 movement point per turn, uphill both ways!!

 

And then when you sent a battleship to take care of them they'd die to a spearman!

 

I usually play continent maps and don't know what changed so much, but in V it wasn't all that unusual to discover a continent completely subdued by a single civ. I've never seen that happen in VI and this can't be just because of 1UPT.

 

Why, you think, is that? Because the AI got better in attacking or because AI got worse in defending?

 

I've always wondered that when you create a new game for only certain victory conditions, does the AI still go for all, or does it narrow it's focus to the ones you chose?

 

Two things are at play, I think.

One is that the AI is programmed to make peace after capturing a city. It doesn't "go for the jugular" in Civ 6. That's understandable considering how silly (in my opinion) it is that the player can wipe out a whole civ in a single war rather than having to through a series of wars to slowly wittle them down. Flipped around, a player shouldn't be out of the game just because they lose a single war. At least, not in single player mode.

The second is that the AI does a reasonably good job of identifying high priority cities to attack and capture, such as capitals. But it does a poor job of identifying whether it will be able to hold them given the new Loyalty system. Combine that with the AI tendency to make peace after capturing one city, and you get the situation where the AI will often penetrate deep into an enemy empire, capturing their capital, then making peace, and having the capital loyalty flip back to it's original owner a few turns later. Ten turns later, those civs are back at war and the same cycle repeats.

 

The AI not going for the jugular is kinda required for the game to work. The way wars work in civ6, it is pretty easy to steamroll over an opponent. If the AI really did go for the jugular, the AI would probably steamroll the player 9/10 and the game would not be fun. By having the AI sue for peace, it gives the human player a second chance to get back into the game and they are not penalized with losing the whole game just for losing an early battle or two.

And I do agree with you that it is not realistic for any side to wipe out an entire civ in one war. I do think that warmongering should become more difficult as your empire grows. Keeping conquered cities should be more difficult where you need to keep more troops behind to secure conquered areas. War weariness in your cities should be a bit steeper too. It would also be great to have a concept of supply lines that would make it easier for the defender than for the attacker. Lastly, it should be possible to raise militia units quickly (1 turn) as a last ditch effort to hold back the attacker if you've lost a big chunk of your army. These changes would all make wars more interesting but I guess I am getting a bit off topic.

I do think in terms of the combat AI, the AI should be programmed to keep some units back in reserve so that it could counter attack better. Often, I defeat most of the AI's army in the first battle and then I can just steamroll over the rest of the civ and take everything.

 

What Trav'ling Canuck said sounds reasonable, but in addition the AI seems to have real trouble taking cities after walls are up. Someone here did some testing on how long it takes the AI to capture your city while passively waiting and sometimes they just didn't, IIRC. Don't know about the silliness, though. I guess it depends on what you think the war represents. A single war or just traveling to the battlefield can take a hundred years. Also, whether you see conquering as something similar as what the nazis did or more like conquest of the Americas.

I wouldn't even mind if some of the AIs just randomly let another AI player take them over to make the game more challenging. A more "immersive" version might be that a couple of the weaker civs formed a union to stop you, the menace, effectively becoming one larger civ (in game terms similar to one conquering another).

 

To me it seems, there is another pioint: AI tend to build wonders and stuff during war time instead of an army. Plus it takes pretty long to build a unit (at least in my prefered game speed it takes lot of time). So If you stand the first attacking/defending wave, then the resistance is quite broken/low.

 

In the Livestream today they seem to be addressing one issue which was why they weren't using Aircraft, and it was that Aircraft right now are pretty damn bad. If you are using them now for anything but Airlifting and getting Eureka moments, you are likely wasting hammers that could be better spent on something else that will let you do what you want to do faster and likely more efficiently. Cool for RP moments, not something you expect your AI opponent to do. They even talked a little about how in Civ 6 they went with Aircraft a bit with kid gloves to specifically avoid the scenario where the endgame became about spamming Aircraft like it was in Civ 5.



I've seen the AI use navies just fine now though. You do need to ramp up the difficulty and have an AI bent on getting a Domination victory to be potentially threatened by them though.

 

I found that whole conversation surprising because on the rare occasions when the AI has attacked me in the late game, I find that 3 Fighters in support of my cities are enough to destroy a huge land army. That's with the current combat strength of aircraft.

I'm okay with them being boosted, though. Aircraft are expensive to build. And of course, we know you can't adjust the production cost of a unit once you've set it's place in the tech tree.

 

lmao at AI apologists for this game.But hey a good chunk of people are perfectly fine with broken products being released because the expansions will magically solve everything. Right, this is why Civ 5's AI was "fixed" by modders making an entirely new game. Or maybe I'll get some nonsense about "programming very AI is very hard". Next thing you'll know, a restaurant will serve me a burnt steak, and instead people will tell me I shouldn't complain because 10% of it is edible and being a chef is very very hard, and you couldn't possibly know. And you know, that other person thinks it tastes okay!

Almost as bad as "I can beat the game by whacking the keyboard with a random body part, why is it so easy? And why does nobody take me seriously?"

But this is why we'll never see improvements. Although to be fair, difficulty in Civ has by and large been just piling bonuses on AI anyways, so "challenge" is w/e. However, considering the game cannot even physically win besides time victory even with so many bonuses, well, that's just bad even for this franchise. When it comes down to it, this is by and far the easiest Civ game even with the biggest bonuses given to the AI.... I mean deity bonuses in former games are now on difficulty 6.

BTW, I still love this game. I just don't love stupid arguments that make the AI look smart by comparison.

 

I'm not so sure that it would cost a lot.I'm also not sure I'd use neural network for most of the computer caracterization of a situation as someone suggested but I could be opening a bland expert debate. They would however need logs from games played. If only they had access to a community of collaborative successfull players spending some hours on the game that could share said logs .... oh wait ?!?

 

Now, I am curious what an AI created by feeding civfanatics games into a neural net would be like. I bet it would create a pretty good AI. The problem would not be a lack of game logs clearly. When I said it would not be cost effective, the issue would be hiring someone or taking someone else off task, just to do machine learning to improve the AI when the benefits would probably not be worth it. As I have said before, while the civfanatics player base would love a strong AI, most civ6 players don't care. If the choice is spending a chunk of their budget on hiring graphic artists to create cool graphics for the game that all players will love or hiring a machine learning engineer to create a super AI to make the hardcore civ6 player happy, which do you think gives Firaxis the most bang for their buck? The answer is the graphic artists.

 

I just yesterday won an Immortal Standard Continents game, SV, with Genghis, and with FIVE cities. FIVE. A huge Poundmaker with 20+ cities that lead the whole game in Science, could not avoid my spies and consolidate his SV. Yes, I know what I am doing and especially with spies, but this shouldn't be possible in a game that has basically re-invented ICS.

Honestly, that statement yesterday sounded, to me, like a lame excuse for the mediocre Air AI. They think we are all stupid, and would believe such a nonsense. "Yeah, we downplayed the air forces on purpose, because they were overpowered in civ 5" blablabla... complete nonsense! In civ 5, at least the AI knew how to use air units to some extent, and if you go Artillery while the AI goes Flight, you are in for a fight.

We are not stupid, FXS. Don't treat us as such. Better to remain silent if you have nothing to say about AI improvements, or else you risk showcasing yourselves as fools.

 

The loyalty system could easily slow down conquest, requiring multiple wars to wear down a Civ. You’d just need captured cities to have a high negative loyalty and for there then to be some bandwidth limit on how many non-loyal captured cities you could absorb at one time.

So, instead of negative loyalty being improved by capturing more cities, it would actually make it worse. You’d be forced to take little bites of a few cities at a time, settle these down, and then go again.

You could even have spy missions or world Congress resolutions allowing other Civs to stir up trouble in these cities to make your expansion harder and harder.

Civ has all the peices. FXS just need to make them work together.

Instead we get Rockbands.

 

I doubt it’ll ever be improved to a point where real hard core strategists will be happy with but I do hope that after the expansion they take some time to make improvements

In Firaxis defence, it’s definitely not as horrible as when the game launched. They’ve made significant improvements.
But definitely not enough
At least the AI settlers are guarded by a unit now
Mostly

 

I don't think anyone is being treated as being stupid here. But I do think if regardless of what they say, you're not going to believe it anyway? Then it doesn't matter what they say, does it

To be clear, you have the absolute right to not believe them. It's just that bit ironic that you're complaining about the accuracy of what they do say, then. It's entirely possible they dumbed-down the usage of aircraft to drive the AI in other areas / priorities.

 

I think difficulty levels shouldn't be about granting some bonus to AI but rather its ability to take optimal decisions. But we are faaaaaaaar from this in Civ6.

 

Wrong. It matters. I did not denounce the accuracy of their statement, but the complete nonsensical nature of it. You have the "right" to not see it, but to me it's pretty clear. Watch the video again, listen to the wording, and if you really want to go into details, watch the body language. I can clearly see them saying exactly that: "We purposely made a mediocre AI because the air units were too powerful in civ 5, and we did not want them to be dominant in civ 6". How is that not complete nonsense to you???