Prince AI is pathetic

[mnf]

Could it possibly be due to too much focus on the agendas? In an effort to make the AI put more emphasis on their agenda, military gets neglected even while losing badly under an ongoing invasion?

 

[stealth_nsk]

Could it possibly be due to too much focus on the agendas? In an effort to make the AI put more emphasis on their agenda, military gets neglected even while losing badly under an ongoing invasion?

It's part of the bigger issues which is called "priority tweaking".

Overall I don't worry about things which could be fixed by altering priorities. What worries me a bit is what AI declares war without actually having any plan. Civ6 has a lot of possibilities for "raiding" war, which could be performed (or at least threaten to) even with small military. I'd like to see AI at least trying to do that if they declare war.

 

[cakes]

yeah it seems to me like issues of prioritisation too. i don't think a bad AI will kill the game for me (and most others who play on lower difficulties anyway, i.e. open to steam-rolling the computers since a lot of people find that fun) because production bonuses will help with prioritisation, and of course there is always the role-playing aspect of it as well. besides, for some reason i think they really are going to improve the AI in a huge way at least after release. it's more difficult to do this if you don't know how the game is going to be played but firaxis do seem to be listening to the fanbase even if they complain about nothing

 

[phryxolydian]

Could it possibly be due to too much focus on the agendas? In an effort to make the AI put more emphasis on their agenda, military gets neglected even while losing badly under an ongoing invasion?

Clearly not. Have you seen Germany's agenda? It sucks: "There are civilizations and there are city states so, let's have a civ that hates civs who interact with city states. Yup, that works!"

And, yeah, I agree about people being so optimistic to the point of blindness. I get it that the game is a few weeks from release but, come on! The game's been on development for how long exactly? They never noticed Civs and CS not upgrading their units? "Oh, yeah, we need to fix that. Don't worry, that's easy enough to do, we'll just do it a week before release. Sure, we could do it right now before people point it out to us but man, we gotta work more on these agendas!"

 

[Magil]

Yeah, that City-State unit spam is something they'll probably want to fix. Could cause a lot of traffic jams with 1 Upt.

I'll tell you what's going to cause 1UPT headaches.

Spoiler :

This. And yet there are still people who think we shouldn't be able to stack with enemy/neutral civilian units...

At least it proves the AI knows how to make one kind of unit

 

[footslogger]

I really thought that by releasing a version of the game 4 weeks before general release, Firaxis were showing their confidence in it. Now, after all the evidence of a not merely poor but imbecilic AI, I am left scratching my head wondering what Firaxis were thinking of. Did they think people would judge the AI to be quite good? My hopes for the game are now well and truly deflated. I'll get some fun out of trying out the new features on release but then I'll not be touching the game until it has had some big patches and even that might not be enough to make it playable for me. One encouraging sign - the main mover behind the Civ5 CP has said on the Reddit forum that he thinks something similar should be possible for Civ6. So that's it, basically I'm digging in for a minimum 6 months' wait before I play the game with any hope of lasting enjoyment.

 

[thunderbird_14]

It's not how game development works.

The more actively you change gameplay, the less advanced AI you may have and the less it follows the gameplay changes. Once the game is more or less stable and in the process of polishing - that's there hard work on AI begins.

The build is at least 1 month before release (maybe a bit more, depending on the development process) and a month of hard work on AI could fix significant flaws and surely could fix something as minor as build priorities IF there's no bigger problem hidden behind.

So, we'll have AI being improved by release, that's guaranteed. Whether it will be able to field as many military units as needed - possible, but we don't know.

But it's already less a month away, AI development is harder than it seems especially one that can adapt to changing conditions (i.e. not flavor-based)

 

[VainApocalypse]

I did not expect wonderful performance from the AI, but it is really really bad..
Doesn't aware at all of iminent invasion, not even after getting denounced and enemy units sighted, doesn't build any capable defenders. Ok, the AI doesn't have production bonus, but if the player can build many attacker units so could the AI, but apart from some warriors and chariots here and there I haven't seen a serious attempt to stop their own annihilation.
Recently I showed civ4 to my friend and started a game for him at warlord difficulty (below noble), and in that game the ai was quite capable.. here prince is a pushover..
Here is hoping that the ai gets optimized further with patches, but so far I'm dissapointed.

This is a non-issue. AI biases and behavior could be changed by swapping some numbers in a spreadsheet, something that's probably going to be done multiple times between whatever press release you're talking about and the final release.

 

[Al Bundy]

I must agree. Prince difficulty is very easy. AI doesn't counterattack, doesn't defend, doesn't upgrade units.... A complete mess.

 

[VainApocalypse]

But it's already less a month away, AI development is harder than it seems especially one that can adapt to changing conditions (i.e. not flavor-based)

The hardness really depends on what's the issue. If it's simply production biases, those things can probably be changed in an afternoon. However, if the AI has the wrong heuristics altogether and they're not merely poorly weighted, then there's a problem. We can't tell which is the case.

 

[one179]

I am really disappointed about AI, combat AI, ed promise improved and it is even worse than in the CIV . Why try to play when you just make three archers and conquer the entire map ...

 

[thunderbird_14]

The hardness really depends on what's the issue. If it's simply production biases, those things can probably be changed in an afternoon. However, if the AI has the wrong heuristics altogether and they're not merely poorly weighted, then there's a problem. We can't tell which is the case.

Could be, but it does still presents challenges to scenarios wherein a specific condition came.

 

[CaiusDrewart]

Given the deplorable state Civ V combat AI was left in--as of the latest version of BNW, the AI still randomly embarks units into danger and still cannot move and shoot with ranged units, to say nothing of the AI's more general incompetence--I see no reason to expect Civ VI's AI to somehow improve between now and release. Would it be possible? Probably. But Firaxis doesn't prioritize AI.

Like others, I'm not worried about the AI not building enough units. I'm sure we'll see the same unit spam on Deity that we've become accustomed to.

 

[Septentrion]

Given the deplorable state Civ V combat AI was left in--as of the latest version of BNW, the AI still randomly embarks units into danger and still cannot move and shoot with ranged units, to say nothing of the AI's more general incompetence--I see no reason to expect Civ VI's AI to somehow improve between now and release. Would it be possible? Probably. But Firaxis doesn't prioritize AI.

Like others, I'm not worried about the AI not building enough units. I'm sure we'll see the same unit spam on Deity that we've become accustomed to.

I agree, with everything here. It's simply not a game without an competent AI. You're never outsmarted after a certain point.

 

[aguliondew]

In civ 5 King is the first difficult that the AI has a possible of winning. Above King difficult I have to be more focused on a victory type. Prince has always been a difficult for casual play. I just hope their are not balance issues like in civ:BE. The preview I have seen so far shows a lot of variety when it come to gameplay. Doubt we will get a preview of the updated AI before the game is released. I still do not see any major issues other than eurekas but that works both ways.

 

[criZp]

London falling to a single viking longship:

 

[Silverdawn]

London falling to a single viking longship:

They are close to taking London, but they don't seem to have done so by the end of the video.

 

[Evs]

It's happening all over again by the looks of it. I'm getting the feeling of Dejavu....

If you think the AI will be much better on release, remember the AI upon release of CIV5 ?

The game was unplayable for me for months. There was patch after patch after patch to improve the appalling AI so don't think Firaxis will not release the game without the AI being competent.

It's the only reason I haven't pre ordered since I DO remember CIV5 release and won't spend my money until I'm sure this game will be playable from an AI point.

There was one reassuring video from the Devs very soon after it was announced, which Ed Beach says they have these new computers in the lobby running AI games 24/7 and it's a completely new AI coding. Then we see the AI on the you tubers videos this week...... Settlers wandering round the entire island on Marbozirs vids, Germany declares war on him with no troops and immediately throws a builder out of his capital in range of Marbozirs troops for him to walk in and capture whilst Germany has no troops at all around his Capital to even defend let alone attack !

The list of AI flawes go on and on with these early release vids. Regardless of whether this build is up to date or weeks old they have had years to work on this AI.

The Devs had years to resolve the AI issues on CIV5 so any hopes of this new AI being anywhere near competent within the next 3 weeks are gone. I never had hope to begin with this time around though.

If you think the AI can't be programmed to be anywhere near competent and this is the best they can do then play CIV5 with the community patch (forget the name). This was done by a few guys in the spare time, unpaid and the AI is so much better and challenging. Their bonuses are also reduced on the higher difficulties further proving it CAN be done.

No Pre order from me after watching these you tube vids....

 

[Golehm]

There was one reassuring video from the Devs very soon after it was announced, which Ed Beach says they have these new computers in the lobby running AI games 24/7 and it's a completely new AI coding.

Exactly. I was wondering why they won't release anything like that, video where only AI plays, if they are watching it all the time.

 

[Haggbart]

To be clear, the main problem here isn't the AI per se, but that the AI don't build and upgrade military units. That should be fixable.

But I'm exited where prince level is on the difficulty ladder. It can make a huge difference if it's nr 3/4 out of 6, or nr 3/4 out of 8. AFAIK we don't know that yet.