Is AI still brain dead?

[V. Soma]

Yes! I would like challenge but honestly my biggest beef is just how crappy it makes the game feel. I just watched Teddy approach an undefended Norse city with 6 units. Several turns later he hasn't even damaged the city and all his units are dead or dying. It's like a giant middle finger to anyone attempting get immersed in the game.

Yeah... I play TSL Earth as Germany and Vikings DoW me in classical age... AI has 4-5 longships and yet doesn't attack my coastal city on the Danish penninsula, though AI could attack on water 4-5 tiles!... (I have NO ship at all...)

 

[Xur]

Firaxis got 1 dude to program AI... and a building of "artists". Their priorities are just elsewhere. In my opinion Firaxis should have a team focused on AI - not just 1 guy... The AI never gets better than that of the ability of the programmer anyway, so why not expand that team.

 

[pgm123]

Yes let's hope so. Maybe if the community refused to buy more DLCs until the game was fixed 2K/Firaxis would take notice. Sadly, until that happens change is unlikely.

I assume Firaxis will just take people off Civ VI if no one is buying the DLC.

 

[Kruos]

No AI improvment = No buy for me.

And I am not alone to think that way, all the members of my game group are tired of this stupide AI, and no-one will buy the expansion if no AI improvments is done. We have so much games to play, with good AI.

 

[Magil]

Civilization AI has always been bad. It made up for this with its huge bonuses. This approach started to show cracks in Civ V with one unit per tile, when no amount of bonuses could make up for the fact that the AI would trip over itself and shuffle its units around, failing to accomplish anything noteworthy. Civ VI is continuing the trend, as now the district system prevents the AI from using its massive production boost and maintenance discounts to build everything in one city (if you build everything, eventually you'll build the right things). The game has evolved, but the AI has not evolved enough to keep up. The question becomes, do you cease adding features just because the AI is not keeping pace? AI is challenging issue, and honestly the best thing Firaxis could do for it is probably to give the modders all the tools they need to fix it. In the Civ IV days at least, Firaxis was not afraid to pick up the things modders were doing and add them to the game. As I recall, BtS's AI adjustments were largely based on adjustments modders had been doing in the years leading up to that expansion (this is what I hear anyway, please correct me if I am incorrect).

Frankly I think poor AI is bad but I'm glad it's the biggest problem in Civ VI rather than other things. Civ V shipped without research overflow (I still can't get over that, launch Civ V was absolutely terrible). I'm more annoyed by the UI design of Civ VI than the AI at this point. The UI could be miles better and this is a much easier problem to tackle than AI.

 

[Usama84]

civ4 AI is pretty decent... doesn't need to be perfect but they know how to defend their borders and launch attacks. diplomacy is based on stats so it's typically pretty logical in that regards. I'm flabbergasted when i hear members say "it's as good as any other game"... really??? i think if they got rid of 1upt that would fix 90% of the problems. i don't think the developers care though considering how financially successful this game was despite being unplayable for most civ fanatics.

 

[Magil]

civ4 AI is pretty decent... doesn't need to be perfect but they know how to defend their borders and launch attacks. diplomacy is based on stats so it's typically pretty logical in that regards.

In Civ IV the AI will constantly suicide large numbers of units against your fortified cities if you let them. They fight battles they have no hope of winning, over and over again. I love Civ IV but let's not sugarcoat.

Civ IV's diplomacy is much more logical, I'll agree with that point.

 

[Mr Jon of Cheam]

civ4 AI is pretty decent... doesn't need to be perfect but they know how to defend their borders and launch attacks. diplomacy is based on stats so it's typically pretty logical in that regards. I'm flabbergasted when i hear members say "it's as good as any other game"... really??? i think if they got rid of 1upt that would fix 90% of the problems. i don't think the developers care though considering how financially successful this game was despite being unplayable for most civ fanatics.

'Unplayable for most civ fanatics'.

Nothing like a bit of hyperbole to ruin your point!

1UPT is clearly a problem for the AI, which is a shame.

 

[Shorlin]

From recent civ6 games I've played I think the AI has improved a lot- decent city planning, some competition with science and culture victories, using planes, better unit movement (sometimes)

The one glaring problem is the AI doesn't pose any threat militarily and I just wish they would code the AI (warmongers at least) to make maybe 3 times as many units as it doesn't currently. I think that would go a long way to fixing it.

I miss the days of Civ4 or even 5 when you meet Shaka or another warmonger on your doorstep and go 'oh ****'

 

[Quintillus]

I'm sorry to hear the AI is still rather sorry. That's the #1 reason I haven't bought the game, and was one of the top reasons I didn't play 5 as much as III or IV. I like there to be at least a chance that I'll be caught off guard by a sneaky AI. And there are games where that happens nowadays. Paradox's AI isn't perfect, but I do have to follow Sun Tzu's maxim about victorious warriors winning first, and then going to war, and I've definitely wound up a target of an opportunisitc AI realizing I was a bit overextended from other conflicts. In Civ5, I could go to war first, and then seek to win, and reliably win, and it sounds like the same would apply in Civ6.

Whereas in III and IV, the AI could occasionally legitimately take cities, and if I was poorly prepared, win a war. I'm experienced enough in III now that losing at a fair game isn't going to happen except with the worst of start positions, but back in the day there were some back-and-forth wars. IV takes it up another notch in AI, and while I'm still unlikely to lose in a fair game, it doesn't take much of an AI advantage for there to be a real chance that I won't win.

Granted, Karpius is right that there are a lot of more casual players. I have some friends who play Paradox games with cheats so that they can essentially defeat whoever they please. To me that isn't fun - I want at least somewhat of a challenge - but it suits them, and likely many others.

Nonetheless, I am disappointed that Firaxis still only has one AI programmer. They had one AI programmer for Civ III (Soren Johnson, prior to his design work on Civ4), and now it's more than 15 years later, they're a considerably larger company, their Civ revenue must be way higher, I'm sure they have more artists, and they still only have one AI programmer for Civ. I don't expect an army of them, but I did hope they would have scaled it roughly proportionally with the success of the series, to at least a small team by now. Maybe with so many complaints about the AI, and the consequent impact on the reviews (especially user reviews), they'll finally do that for Civ VII.

 

[Kruos]

I'm sorry to hear the AI is still rather sorry. That's the #1 reason I haven't bought the game, and was one of the top reasons I didn't play 5 as much as III or IV. I like there to be at least a chance that I'll be caught off guard by a sneaky AI. And there are games where that happens nowadays. Paradox's AI isn't perfect, but I do have to follow Sun Tzu's maxim about victorious warriors winning first, and then going to war, and I've definitely wound up a target of an opportunisitc AI realizing I was a bit overextended from other conflicts. In Civ5, I could go to war first, and then seek to win, and reliably win, and it sounds like the same would apply in Civ6.

Whereas in III and IV, the AI could occasionally legitimately take cities, and if I was poorly prepared, win a war. I'm experienced enough in III now that losing at a fair game isn't going to happen except with the worst of start positions, but back in the day there were some back-and-forth wars. IV takes it up another notch in AI, and while I'm still unlikely to lose in a fair game, it doesn't take much of an AI advantage for there to be a real chance that I won't win.

Granted, Karpius is right that there are a lot of more casual players. I have some friends who play Paradox games with cheats so that they can essentially defeat whoever they please. To me that isn't fun - I want at least somewhat of a challenge - but it suits them, and likely many others.

Nonetheless, I am disappointed that Firaxis still only has one AI programmer. They had one AI programmer for Civ III (Soren Johnson, prior to his design work on Civ4), and now it's more than 15 years later, they're a considerably larger company, their Civ revenue must be way higher, I'm sure they have more artists, and they still only have one AI programmer for Civ. I don't expect an army of them, but I did hope they would have scaled it roughly proportionally with the success of the series, to at least a small team by now. Maybe with so many complaints about the AI, and the consequent impact on the reviews (especially user reviews), they'll finally do that for Civ VII.

'More' doesnt systematicly means 'Better'. Espescially for complex subject like the Civ6 AI, which seems to based on behaviors trees. I think the talent of AI guy and his knowledge of the game matter a lot. At least I hope he has some support and the testers help him improve his behaviors trees by setting up good test scenarios and giving him good feedback. But not sure. From what I have read in the AI mods forum here, there are still lot of blurred area with strange things like the operations which stop suddenly sometimes, definetly not enough test have been done in this case.

Regarding the AI comparison with Civ3&4 I agree with you 100% : what people seek is not an unbeatable AI, it is just an AI who will give sense to the game, by at least being able to punish the player error and threat him credibly, like it was the case in the previous pre-1UPT games.

 

[PYITE]

The one glaring problem is the AI doesn't pose any threat militarily and I just wish they would code the AI (warmongers at least) to make maybe 3 times as many units as it doesn't currently. I think that would go a long way to fixing it.

I miss the days of Civ4 or even 5 when you meet Shaka or another warmonger on your doorstep and go 'oh ****'

Give this mod a try, AI will build more units and be more aggressive.

https://forums.civfanatics.com/resources/civ-flavor-deity.26561/

http://steamcommunity.com/sharedfiles/filedetails/?id=1206813450&searchtext=

 

[pgm123]

Give this mod a try, AI will build more units and be more aggressive.

https://forums.civfanatics.com/resources/civ-flavor-deity.26561/

http://steamcommunity.com/sharedfiles/filedetails/?id=1206813450&searchtext=

I can't figure out what that does, exactly.

 

[PYITE]

I can't figure out what that does, exactly.

Changes the build preferences for each individual leader so they play in a style that suits their strengths. This is most evident in the districts that they choose to build. For example Tomyris will build encampments, commercial districts and entertainment districts when amenities are needed. Only religious civs and cultural religious will build holy sites!

They all build more military units and are generally a larger threat for invasion; building walls isn’t just to get overflow. When the AI is attacked they are much more likely to trigger the defensive operations which make them build walls and units for defense.

If beating Diety is a forgone conclusion and has become a bit repetitive and boring for you, then give it a try.

 

[ThunderLizard2]

Regarding the AI comparison with Civ3&4 I agree with you 100% : what people seek is not an unbeatable AI, it is just an AI who will give sense to the game, by at least being able to punish the player error and threat him credibly, like it was the case in the previous pre-1UPT games.

Agree that the Civ IV BTS AI was not unbeatable but put up a good challenge. With 1 UPT in Civ 5 I never lost a city unless I just left the totally undefended. I don't think programming an AI to deal with 1 UPT is feasible.

Nonetheless, I am disappointed that Firaxis still only has one AI programmer.

That's crazy given the complexity of the game. Just dealing with all the new features like districts would take one programmer.

 

[Thormodr]

The AI is roughing up Yogscast in this video. So, not totally brain dead. Lol.

 

[Magil]

The AI is roughing up Yogscast in this video. So, not totally brain dead. Lol.

The AI can absolutely take cities if they don't have walls. They tend to steamroll city-states in high difficulty games. Things get a lot iffier if the city has walls though. Also the Yogscast is playing pretty poorly in that video

 

[Thormodr]

The AI can absolutely take cities if they don't have walls. They tend to steamroll city-states in high difficulty games. Things get a lot iffier if the city has walls though. Also the Yogscast is playing pretty poorly in that video

Certainly they were playing poorly. However, it was on Prince level and the AI did what had to be done. Impressed with the AI there.

 

[Magma_Dragoon]

AI still moves their starting settler off fresh water to get into a position where it gets +1 for a holy site from trees and then chops the trees

 

[magha77]

From recent civ6 games I've played I think the AI has improved a lot- decent city planning, some competition with science and culture victories, using planes, better unit movement (sometimes)

The one glaring problem is the AI doesn't pose any threat militarily and I just wish they would code the AI (warmongers at least) to make maybe 3 times as many units as it doesn't currently. I think that would go a long way to fixing it.

I miss the days of Civ4 or even 5 when you meet Shaka or another warmonger on your doorstep and go 'oh ****'

I don’t see how the number of units would change how the AI doesn’t use them. I’ve had so many games were the AI could have taken a city but just stop for some reason. It’ll pillage or attack units but not take cities.