So they still haven't fixed the AI with regard to war.

[Zinowolf]

So now we have a pretty complete game with the addition of GS and the game still remains incredibly easy to beat. First off, this has nothing to do with difficulty level. The AI behaves the same whether you play on easy or deity. All the difficulty level does is give them mega buffs. It does NOT change how the game plays against you.

The strategy that I have used since launch day is build between 4-6 bombers, some melee units and it's game over for the AI. This strat is still as relevant as ever in GS. The AI simply does nothing to counteract a bombing campaign. I was able to take every single one of my rivals cities with this method and did ONE of them build anti aircraft, defend themselves, try an offensive mission against me? No they did not. The game is still laughably PISS weak in this area and as far as I am concerned , it's simply unfinished/broken. So if a rival is running away with tech or culture, you know what to do. I guarantee you simply cannot lose.

Will they EVER fix this?

 

[Equilin]

first, how is their tech? they cannot build AA if they haven't got the tech for it yet, and you can't blame them for that
second, late game production is really high the ai probably can't do it, and they cant effectively rush buy either

 

[Nathaniel VII]

You're right, they haven't. I upped the difficulty again for my second game because I just steamrolled the entire map with Hungary in the previous one, despite having played at a higher difficulty than I used to. And while the AI has a better starting position thanks to the bonus stuff it gets from the difficulty, it just can't do anything with it.

When I got boxed in by two civs I declared a badly prepared surprise war against Scotland to the north, with just three warriors and three archers. They used their horde of warriors to kill half of my "army", but then my first horsemen came online and I took their first city.

From here on out all they produced were catapults, probably because the script says "recapture lost city". Not that the city I took had any walls, mind you. The script doesn't check for that. Couple turns later I'm standing in front of Scotlands capital, waiting for my own siege units, because the AI finally thought to build some walls so I can't just storm the place.

It is so boring. And all higher difficulty does is give the AI an early advantage, which soon becomes your advantage because they just can't defend themselves.

The bomber issue you mentioned is true as well. I took an entire continent's worth of cities in my game as Hungary with one carrier and two bombers.

So I just end up clicking through the ages, waiting for an accidental culture win to happen. Maybe it's time to uninstall Civ VI for a while and try IV or V with Vox Populi...

 

[sa1vy]

I content myself with trying to achieve quickest wins. The AI will never be able to compete in terms of strategy. I compete against myself then, improving my play to get faster and faster wins.

Try increasing the difficulty you play at, or change the time you go to attack. By the time you get bombers, the game could be over several different ways. Try archer rush, war carts, knights... There's a lot of other military strategies that you might find interesting.

 

[Victoria]

The strategy that I have used since launch day is build between 4-6 bombers, some melee units and it's game over

Well Dom GOTM games always end before bombers so the issue is more than just that.
This seems to be an air units complaint more than anything.
Both AI cleverness and Air unit issues have been covered many times already, maybe rename the thread to AI is still bad in the air as well as everywhere else. But do not expect a good long debate as they have been done already.
Bottom line is yes it is broken and many people think it cannot be fixed, certainly without investment it seems an issue. And generally the game is about making money, the huge investments in art teams are likely where your AI has hidden itself.

 

[Krajzen]

You know, I too want AI to be more competent, I really do, but I also really dislike using the word "fix" in regards to strategy games AI. Because it implies to problem is just a binary simple dysfunction, a bug, a technical issue, imbalanced feature, and can be simply "repaired" to a "optimal" state.

AI will never be fixed because it is not something fixable like a leaking tap, there is no clear finish line line herr. It can made slowly, continuously better by a ton of coding and testing, and definition of what makes "better" AI may vary. Personally I'd enjoy this game being sort of hardcore, when surviving to modern era is achievement in itself, but most of people want to have sandbox fun and win reliably in most games, and Firaxis made more players and money than ever with appealing to casual masses of players of all ages and genders, who can reliably win their very first game of civ. On top of that, they want AI to not be ruthless chess math optimizer, but to roleplay certain leaders and cultures at the cost of "rational" behaviors. On top of that a recent thread on this forum has proven that devs program this AI to play the game the way they think works best, but very experienced players actually know the high difficulty level game better than devs themselves

It's too complicated issue to simply fix it, it can be only managed worse or better with the current very primitive gaming AI that is incapable of processing any deep strategy game in general.

There is also this mantra Civ Reddit mindlessly repeats that "high dif levels should make AI play better instead of giving bonuses". If those people seriously imagine any game having entirely separate AI for each of several difficulty levels, they have no idea how this all (AI, game dev) work in practice. This is simply impossible, beyond making low diff levels AI intentionally unable to use mechanics and unlocking them later, and highest lvl AI beinf insane warmonger that wants to murder you at ant opportunity - but that would be not necessarily clever behavior, and would ruin immersion aspects.

 

[Tomice]

There is another recent thread here with extensive AI testing. It seems that if you just leave the AI's to beat on each other, they will eventually have armies full of Jet bombers and GDR's.
The AI is however very bad at the following:
1) tactical troop movement
2) upgrading troops, resource management
3) deciding to actually capture a city

Regarding 1, there's little that can be done. A competent human will always be better at complex 1upt maneuvers, the AI can only compete in games with relatively simple rules (e.g. Chess). Returning to stacked troops might help for civ 7.

Regarding 2, the AI seems not very capable at securing sufficient strategic resources, which results in leftover warriors and chariots. It might be possible to improve this with patches, but it's a problem with game design first and foremost.

Regarding 3, look at the other thread for theories. It seems that amenities might play a role (AI's may be hesitant to take cities if they lack amenities).

 

[kb27787]

I'm pretty sure I saw one aerodrome from deity AI at around very late game (210 turns or so). So, yes I'm sure they will eventually get planes to fight your planes if they get the tech.

No sense to complain about AI not fighting planes if you play settler or so and then discover AI takes 300+ turns to reach flight...

 

[Pistol90]

AI is still somewhat capable of fighting at the begining of the game .... but after Medieval era its criminal , you could do whatever you want --- even if you lose cities early , you can easly win them back later.

 

[Aussie_Lurker]

Funny you should mention that, because my recent City-State emergency is saying something quite different. All the Civs-& even some of the City-States that they controlled-who took part in the Emergency actually sent decent, "modern" (for their tech level) units along to fight Norway-the target of the Emergency. Meanwhile, though it was confined to an Inland Sea, I saw The Maori & Norway participating in some brutal naval battles using Caravels. I only won the Emergency because I had the help of 3 other Civs & the troops I levied from Bologna. Otherwise I may have struggled......& this is only at Prince Level.

I was also pleased to see the AI both leveraging CS units & bringing along escorted Siege Weapons-though my Muskets & Cavalry took back Kabul before the Siege Weapons could be brought to bear.

 

[Aussie_Lurker]

There is another recent thread here with extensive AI testing. It seems that if you just leave the AI's to beat on each other, they will eventually have armies full of Jet bombers and GDR's.
The AI is however very bad at the following:
1) tactical troop movement
2) upgrading troops, resource management
3) deciding to actually capture a city

Regarding 1, there's little that can be done. A competent human will always be better at complex 1upt maneuvers, the AI can only compete in games with relatively simple rules (e.g. Chess). Returning to stacked troops might help for civ 7.

Regarding 2, the AI seems not very capable at securing sufficient strategic resources, which results in leftover warriors and chariots. It might be possible to improve this with patches, but it's a problem with game design first and foremost.

Regarding 3, look at the other thread for theories. It seems that amenities might play a role (AI's may be hesitant to take cities if they lack amenities).

I am not seeing issue #2. Though slightly hodge-podge, I have noticed that all my AI neighbours have upgraded their units to the most advanced they can build. This is particularly true of Naval Units, which have all been rapidly upgraded to Frigates & Caravels. There are very few Ancient/Classical Era ships left.

 

[Krajzen]

I want AI to do one thing well and that is my priority: when it should take a city, send a big force and take it in fast decisive assault. That's like half of my AI concerns right there. No dancing in city range, suicidal attacks, indecisiveness, no: when AI enters city range, defenders die and city defenses go down.


Anyway, I still think 1UPT would be much lesser pain for AI if all units had only 1 range. The existence of fragile ranged - strongstrong dichotomy on so cluttered map with so unforgiving movement rules and so many complicated factors are nightmare to program AI for. It is one thing to program AI "keep ranged in the back, melee in the front" when game is like heroes of m&m and battles happen on separate chess-like map with at most 7 units on one side and at most 7 on the opoosite side. Its entirely different matter to make AI keep this rule in extremely chaotic civ map environment, with dozens of units attacking each other in countless combinations and from all directions. Civ combat, movement and terrain system make it essentially impossible for AI to keep any sort of simple formation like it can do in Total War games "spear at front, bows at back, cavalry at sides". It has to costantly juggle a **** ton od units and unit types and objectives, and all this is simply to overwhelming for it to not repeatedly end up in with archer on the frontline and infantry standing in the ranged of artillery. And keep in mind that it also has to manage peaceful layer, that there can be 40 AIs in one game (CS and civs), that turn processing times need to be short so AI cannot be too good, that it has to adhere to certain immersive fun behaviors that are often irrational from realpolitik perspectice. That the game is so overloaded with bonuses and systems that human power can chain insane combos to be super strong in very narrow aspect, and AI is not meta enough to grasp such combos. That AI cannot be flexible enough to truly change its behavior depending on civ its playing at the given game. That civ6 district minigame is also hard for it to manage.

AI could be better in civ5 and civ6, but huge part of the problem how extremely hard is to program AI for such game mechanics at the current technological level.

The simplest way to make this game much harder would be to simplify its mechanics so they'd be easier to grasp for AI. Doomstacks, building stacks, less and more shallow mechanics, more plain boring factions with passive universal bonuses. What a terrible Catch 22 dilemma: either boring or easy game.

 

[Gengsikan]

This! My biggest problem with CIV 6. AI is stupid. If you survive to first 2 eras game is done. If you survive a warrior horde from AI, you easily conquer them with 1 warrior and 3 archers. In the beginning they only care about monuments and spread cities. Playing in deity I just just conquered 5 cities from Georgia with 3 archers 1 warrior and 2 spearman. No matter that I attacked its capital, as it was on the coast AI continued to build boats instead of defenses.

It does not matter if the AI starts with numerous advantages at the start if it does not know what to do with these features. Well... GS has not changed anything for me apart from new mechanics. After 2 games I lost interest because I just use the same formula every time, if you win one or two AI cities, the game is over and it's a stroll to victory.

 

[Deggial]

Anecdotal evidence:
Yesterday, I invaded Mali that was technologically quite far ahead.
It had a fighter deployed and a bomber stationed in an airdrome.

And - very much to my surprise (and horrified delight!) - it used both of them to attack my ground troupes! The poor targets were obliterated, as they were not enforced to corps (let alone armies) yet.

Fortunately (for me) respectively unfortunately (for the state of the AI), they only attacked once each.

This might have to do with the fact that I conquered the city with the airdrome this turn and the bomber got stuck there (not being attackable until I got a biplane myself).
The fighter on the other hand was deployed in my cultural borders as well now - something that is not possible under the current rules, iirc. It was "sleeping", too, after his first attack. It healed itself though, after I attacked it with multiple city strikes from my newly conquered nearby cities.

Conclusion: Progress does seem to exist.
It doesn't seem to go far enough, though and there are still bugs keeping the aircrafts from being re-deployed from now occupied places (or, in case of the bomber, being destroyed like other units in conquered encampments).

 

[Council 13]

The AI is unbelievably bad. Well below Civ V standards. Its disheartening that there has been pretty much no progress since initial release in this area.

 

[Takfloyd]

I forward settled an AI on Deity on a different continent with an undefended city recently. The city had a combat strength of 10 and no walls.

Not long after, it moved a Pikeman and a Warrior to my fledgling border and declared war. I had no way to reinforce the city from anywhere else, and my strongest available unit for construction was the Archer.

• Enemy Turn 1: After declaring war, the units do not attack the city. They just stand outside of it.
• My Turn 1: I purchase an Archer in the city. I immediately regret this, as I see the city's strength only increase to 15, the melee strength of the Archer, when I was assuming it would be 25, its ranged strength. I wish I instead had bought a Spearman.
• Enemy Turn 2: The Pikeman and Warrior both attack the city, leaving it in the red. One more attack from the Pikeman will take the city.
• My Turn 2: I move the Archer north outside the city opposite of the Pikeman, and fire on the Pikeman for negligible damage. I now buy a Spearman in the city, raising its strength to 25, which still wouldn't be enough to resist the 41 strength Pikeman.
• Enemy Turn 3: Instead of attacking and taking the city, the Pikeman moves one step to the north towards the Archer, onto a forest tile. The AI does nothing else. My city heals up to yellow HP.
• My Turn 3: I attack the Pikeman with the Archer again, move my Spearman out of the city and buy another Archer. The AI will still take the city if it attacks with the Pikeman and Warrior.
• Enemy Turn 4: The Pikeman activates a promotion instead of taking the city, regaining its HP. The Warrior moves away. My city keeps healing.
• My Turn 4: I can now attack the Pikeman with 2 Archers while keeping my Spearman in the city for defense. The AI's attack has failed spectacularly.

A few turns later: The AI starts begging me for peace, offering gold per turn and luxuries.

A few more turns later: I start taking the AI's walled cities with my Archers, despite the AI having several Coursers at this point. The AI only keeps Archers in his cities, which as previously mentioned means their strength is extremely low. If he had just put Pikemen or Coursers in the cities it would have been impossible for me to take them.

Pathetic.

 

[Tech Osen]

Well Dom GOTM games always end before bombers so the issue is more than just that.
This seems to be an air units complaint more than anything.
Both AI cleverness and Air unit issues have been covered many times already, maybe rename the thread to AI is still bad in the air as well as everywhere else. But do not expect a good long debate as they have been done already.
Bottom line is yes it is broken and many people think it cannot be fixed, certainly without investment it seems an issue. And generally the game is about making money, the huge investments in art teams are likely where your AI has hidden itself.

Getting the AI to use airplanes might be a complicated matter, and I can understand that that might take quite a bit of resources. The easy counter to player used planes however is simply building AA guns, and I have yet to see an AI build those. It can build and use other support units just fine but never AA. And I can't imagine that that would be all that hard to fix.

 

[gunnergoz]

The AI is definitely a mixed bag. I see some improvement, yes. But in one recent game I saw the AI build two warships in a two tile lake far from any possible ocean access. Complete waste of AI resources. What would it take to add code testing ("tile adjacent to seacoast=true") for warship builds?

 

[Aristos]

If those people seriously imagine any game having entirely separate AI for each of several difficulty levels, they have no idea how this all (AI, game dev) work in practice. This is simply impossible, beyond making low diff levels AI intentionally unable to use mechanics and unlocking them later, and highest lvl AI beinf insane warmonger that wants to murder you at ant opportunity - but that would be not necessarily clever behavior, and would ruin immersion aspects.

No it's not impossible. Brad Wardell coded exactly that into Galciv 2 and it worked very well. He coded the AI algos so that only at the highest difficulty levels all full algorithms for the AI would be active, while gradually decreasing the chance of each algo to trigger the lower the diff level. You could even "feel" the effect of that during gameplay, and it was the opposite of immersion ruining.

I.e.: you build up in the border with a hostile (or target) civ; if the AI civ had "vision" on your build up, on higher diff levels they would come to you and tell you something in the line of "I know what you are doing", and then immediately start their own build up in the area. In lower diff levels, sometimes they would notice, and react, and sometimes they would not. At the lowest level, they would not react most of the time.

Many more examples like that. It can be done, it has been done, and it can be very immersive (and effective, GC2's AI is still considered the best of the series today, even with its flaws, and of course it is eons ahead of civ 6's).

The AI is definitely a mixed bag. I see some improvement, yes. But in one recent game I saw the AI build two warships in a two tile lake far from any possible ocean access. Complete waste of AI resources. What would it take to add code testing ("tile adjacent to seacoast=true") for warship builds?

That's nothing. Yesterday I saw a Flood Barrier on a two tile Wonder lake.

 

[Sostratus]

I wonder how much of these military issues are actually economic issues in disguise.
Especially around amenities, producing more units when they lose them, brutally bad card slotting, that sort of thing. If the AI has growth bonuses it will constantly be running into amenity problems. Other than adjacency, we would expect that this huge emphasis on just having infrastructure would make yield management pretty easy since the AI only needs to place lots of campuses etc.

Their district placement is also bad, though; I wonder if there is a way to improve it. I often raze captured cities because of the poor district placement. Part of it is planning, but perhaps a simple model of "flat land=farm, hill=mine, resource=improve resource" With a tweak for environmentalist civs to use lumber mills. Then you'd know how the terrain will be improved in advance and can,for example, plan IZs.

Doomstacks, building stacks,

Stack building may be appealing for Civ4 vets, but dropped into civ6 as is would probably result in stacks dancing in your city strike range instead! The addition of the City as a static unit and ranged units in general, as well as the corps system has, IMO, made stacks "obsolete." Stacks' AI upside is force concentration- simply having much earlier corps/armies in the game would also give them that advantage (because they also have very high unit production bonuses.)