TotalBiscuit thinks the AI is good. and i'm now willing to believe it...

[Gali]

The numbers are different because the games use fundamentally different numbers of units. Filthy could walk into civ 4 build a respectable military and roll the Ai on Prince with minimal game knowledge or tactics. Civ 4's ai is good with Patches and those patches heavily incorporated improvements made by mods.

I have repeatedly stated that the AI in civ 6 will need mods to challenge elite players over time. Civ 5 AI was crippled by the lack of mod tools and unfinished game systems until Gods and Kings at least. The main thing I am interested in now for civ 6 is how good the mod tools are. But the AI in the previews has shown the ability to move and shoot and decent tactical moves. Barbarians are actually a concern now. The game systems seems to be in a much more finished state which will allow for the AI to be worked on more easily. Firaxis and modders had an entire expansion to learn how to code 1UPT AI. There are late game stacking systems to allow the AI to utilize its production advantages. Modders should be able to expand those if they are needed. Continued AI patches, bonuses and utilizing modders will be needed for the AI to keep up but the foundation is far closer to civ 4 than civ 5 it seems.

 

[jekke]

The AI is going to need time and modders to reach its potential. SO DID CIV 4. This thread is about the fact that the release state of AI is in a much stronger position than in civ 5. If modders managed to turn that Ai into a decent state they can do much more with civ 6. If Civ 6 has real mod tools on release and not after 2 years the Ai will improve much more quickly.

You know what is really puzzling? Granted, it took couple of years for the modders to bring Civ IV AI onto a reasonable level. And then AI in Civ V had the same problems, as if those modders efforts were gone absolutely unnoticed, and AI needed to be invented from scratch! Okay, said the modders (I extend my greatest respect to those people), we shall fix this once again, and created the community patch and other mods. Ai has improved, and regular players like me were anticipating civ VI to use at least some of the progress made over these years. And yet again, we hear that it is balance which needs adjustment, more testing and it will take years for the modders to make AI really AI and not just A. As an observer, I get the impression of reinventing the wheel every few years, while the resources are spent on drawing Cleopatra's boobs and researching an obscure name for Milford Sound.
Coming back to the topic, I am not saying that it is easy to fix current issues, and neither that AI should beat a first-time-civ-VI player in 100 turns on Prince. I am just surprised how easily people fall into the "you are in fact haters, look how many good things exist"-mode.

 

[Abraxis]

Totalbiscquit does not put his livestreams on youtube but there is an unofficial channel that does so with his permission as long as they wait a few days to give his subscribers first access and don't monetize the videos. Just search Totalbiscuit unofficial live stream.

Great! Thankyou so much for posting this!

 

[Ulthwithian]

jekke: Something I've read is that there can be legal liability reasons regarding the modders' AI balancing. That is, the developers may not be touching the mods like the Community Balance Patch (IIRC) and others because they are afraid they will get sued if they use code from it.

 

[Krajzen]

jekke: Something I've read is that there can be legal liability reasons regarding the modders' AI balancing. That is, the developers may not be touching the mods like the Community Balance Patch (IIRC) and others because they are afraid they will get sued if they use code from it.

They could ask.

 

[Josephias]

Well, if city has either walls or defending unit, it could defend quite well in Civ6. If the city has neither, I don't see why it should be a significant obstacle for a strong army.

Of course it also depends on how well AI could manage defenses, but patching holes in AI logic at the expense of gameplay doesn't look like a good practice.

It's a bit late and OT to reply, but: Well, buildings alone may provide valuable fire cover possibility, even if it is just civilians who are defending the city. Sure, it will not hurt much your units to be thrown stones and break down doors, but it takes time furing wich you may be countered
(however, now I think of it, this is the post-conquer unrest... so maybe it could be just considered the city stays at 0 HP post-conquer, until people is on your side, allowing you not to gain defensive benefits from the conquered city, easing recovery by defender)

So, yes, I agree lack of city garrisoning when war is coming should be considered bad AI.

 

[Ulthwithian]

They could ask.

I believe what they said is that the modder (you or others) have to release them first.

I know for a fact that other games (like Magic) deny developers the ability to even look at unofficial sources, because the policy protects them from being sued. If MaRo or other Magic designers can look at fan-made cards, they cannot prove that those cards were not the inspiration to cards that were made and thus the creator of them would deserve (or could legally be given) royalties.

 

[Xur]

I saw Pyrion Flax saying the AI was <snip>. I'm paraphrasing here: "the AI is <snip> again...". So I'm guessing that compared to the current civ5, the AI is pretty bad.

Moderator Action: Please remember that this site's inappropriate language rules remain in effect, notwithstanding the forum migration.
Please read the forum rules: http://forums.civfanatics.com/showthread.php?t=422889

 

[ExemplarVoss]

Similiarly, we've read articles complimenting Civ6 on how amazing is, written by people who've never played the game before. And if you've never played a civ game before, yet find it super fun, Firaxis did something right.

I actually disagree with this, rather intensely. What this means to me is they don't have enough experience with the game to find or understand the problem areas. With any new game, it is entirely possible to spend the first 10 hours in blissful ignorance, but with time and experience the flaws start becoming understable.

To me it doesn't mean Firaxis did something right (except possibly their PR department knowing some suckers), just that those people (and the average 'general audience' youtuber like TB, Crendor or Mathas, as opposed to the specialized strategy youtubers) don't have the skillset to properly evaluate the game.

Don't get me wrong, I think the game is in a much better release state than the last several offerings from Firaxis (civ 5, beyond earth and starships, or even the still buggy x-coms), but the flaws in the AI strategy, missionary spam, and the fast science and slow production are big problems that very much need to be addressed.

 

[Rathelm]

I really think people are overreacting here. In Quill's Rome playthrough the AI was so passive during midgame I cannot imagine that being the case in the released version, especially on harder difficulties. His production was so weak because he focused so heavy on science that any aggressive civ that focused on encampments and industrial zones would have wiped him out.

 

[Stilgar08]

Don't get me wrong, I think the game is in a much better release state than the last several offerings from Firaxis (civ 5, beyond earth and starships, or even the still buggy x-coms)...

Huh? I had XCom quite early and don't think it was buggy at all! XCom 2 as well.
Couldn't argue about BE or starships since I never played them... I'd say starships (and even BE to a lesser extent) were more of a 2nd /3rd priority release and not comparable with CiVI, the 25th anniversary edition of THE strategy flagship they have (no pressure, Ed!! ). OTOH CiV was obviously flawed and buggy like hell in vanilla... :hmmh:

I don't have the time to watch each and every playthrough but I see a lot of small bugs, inconsistencies in UI or small missing Infos from what I'm watching.. . This gives me hope this release is further away from the release version many here might fear...
It is build 1.011.xx or something alike , right? (on my mobile right now so I cannot check). Is there a rule of thump which version likely might be the release version? (2.02.x or 1.22.y or something like that) Or is it impossible to guess that?

 

[ExemplarVoss]

Huh? I had XCom quite early and don't think it was buggy at all! XCom 2 as well.

I was thinking particularly of the weird animation glitches- where they shoot through walls, break nonexistent windows (or not break existent ones) and things of that nature. That still happens, despite all the time that has passed. That and the horrible optimization, especially for 2.

 

[Larsenex]

I agree with Madjinn on some of his points. So many calls saying this or that is easy to fix and the reality is we know nothing of the code and ergo would know nothing on how easy to fix.

One comment was 'just a few lines of code to fix' which I thought was kinda funny.

I do feel from looking at the various Utube and streams that this thread is valid. I think at release we will have a game and an Ai that is better than the one we saw at release for Civ V.

 

[Stilgar08]

I was thinking particularly of the weird animation glitches- where they shoot through walls, break nonexistent windows (or not break existent ones) and things of that nature. That still happens, despite all the time that has passed. That and the horrible optimization, especially for 2.

If those would be the problems with CiVI this thread would be meaningless and I'd happily embrace them. ;p I'll wait and see (nothin we can Do otherwise, IMO)

I do feel from looking at the various Utube and streams that this thread is valid. I think at release we will have a game and an Ai that is better than the one we saw at release for Civ V.

Fair enough.

 

[kaspergm]

Well, if city has either walls or defending unit, it could defend quite well in Civ6. If the city has neither, I don't see why it should be a significant obstacle for a strong army.

Of course it also depends on how well AI could manage defenses, but patching holes in AI logic at the expense of gameplay doesn't look like a good practice.

I don't quite know what to read into your wording "could" here, but problem as I see it here is that no, they don't really defend well, even with walls and units in it. For instance, in Quill's Rome game, Madrid was a pretty big city with walls and a unit inside it when he attacked it, yet it was a complete walkover. And the fact that it came down to pretty much a coin toss whether the Yogscast team was able to capture London with one single Longship speaks volumes, imo.

I'm certainly not asking for a return to Civ5 balance, but problem is we've gone from cities being both tough and very damaging to being neither tough nor very damaging. What I say is I think the better balance is somewhere between these two poles, i.e. tough but not very damaging.

 

[Opera]

You know what is really puzzling? Granted, it took couple of years for the modders to bring Civ IV AI onto a reasonable level. And then AI in Civ V had the same problems, as if those modders efforts were gone absolutely unnoticed, and AI needed to be invented from scratch! Okay, said the modders (I extend my greatest respect to those people), we shall fix this once again, and created the community patch and other mods. Ai has improved, and regular players like me were anticipating civ VI to use at least some of the progress made over these years. And yet again, we hear that it is balance which needs adjustment, more testing and it will take years for the modders to make AI really AI and not just A. As an observer, I get the impression of reinventing the wheel every few years, while the resources are spent on drawing Cleopatra's boobs and researching an obscure name for Milford Sound.
Coming back to the topic, I am not saying that it is easy to fix current issues, and neither that AI should beat a first-time-civ-VI player in 100 turns on Prince. I am just surprised how easily people fall into the "you are in fact haters, look how many good things exist"-mode.

It's not like they're the same games, you can't simply copy & paste code.

(Also, how obscure is it if it's right there in the Wikipedia article, really?)

 

[REDfeller]

At harder difficulties they need to ensure AI builds a decent military. I think the game will be fine with average players (just a gut feel) but they gotta do something about human players steamrolling with a giant military.

Watching FilthyRobot CIV (5) multiplayer streams, he checks the demographics screen almost EVERY turn, and admits that all he is really looking at is "Amount of soldiers" and where he ranks with other players.
I think we have all seen the CiV strange tactic of Prince AI refusing to build an adequate standing army (or react to DOW etc).

If he drops in power he builds a unit to "keep up with the Jones's", so that when other players look at the stats, he doesn't look weak or ripe for attack.
Surely its not that hard for the AI to be programmed to check these stats and based on "AI personality", ie: America has 110% of average, India has 70%? (random examples) make the same binary decision?

Am I behind in soldiers with average of top 3 civs? NO - carry on, YES, build a unit.

In a management/industry equivalent day to day we all have KPI's. This screen is just your own live KPI performance standing in the game.

You make you own judgement and check back on the scouted area or look through the eyes of a spy . . .

Are the Zulus building a wonder? Which one?Are they focused on production?Do the English have a lot of farms?Can they build or have built a large navy?Better check it out.

The info is there for the player and AI alike, all this talk of not being able to programme an equivelant human AI. Don't need to.

Programme simple process map decisions based on AI "personality and KPI priority".

Am I behind in production? YES - build a mine and work it, NO - carry on.
Am I behind in culture? YES - build culture building/work tile, NO carry on.

Maybe I will get flamed for this as I am not a programmer, but it doesn't seem to be rocket science for AI to have modifiers to keep up with a human in decision making.

Unit placement/usage is another kettle of fish though . . . . . .

 

[Ovarwa]

One of the features of all Civ games is that a city requires multiple turns to create something. Another feature of all civ games is that a city only gets to build one thing at a time.

Taken together, these present human players with interesting and difficult choices. But they also do the same for the AI.

Easy enough to say "The AI shall build units to maintain a specified power level." But what if it is in the middle of building a Wonder, or satisfying some other priority? Does it abort what it is building to build a military unit, and then abort the unit next turn when 2 civs kill some of the other's units? What if it neglects science in order to build those units, and finds itself building an endless stream of warriors to cope with your tanks? What if you have 2 AI civs that both want to be #1, and do nothing but build units, striving to beat the other in the standings, while otherwise utterly neglecting their civ?

Not so simple.

 

[oakdragon]

Could be a direct side effect of roads no longer costing maintenance + trade units auto building them. I could envision times in which the fastest way for a barb to get a captured civilian back to the camp is via a road in neutral territory. Which of course sends it near a players territory.
In addition edge case of the "home barb camp" has already been destroyed, but a barb unit from it had already captured a civilian and is now trying to get it to what is now the closet barb camp.

I was just noticing in ep. 7 of Marbozir's Let's Play Greece, a barb camp spawned surrounded by his cities with two melee units and a scout, and there was apparently nowhere else the scout could have come from. Hmm.

 

[Evs]

I'm not too worried because I feel in the end I will be playing a modded version of Civ 6, because modders inevitably are better at balancing the emerging metagame than developers are, for a number of reasons. The Community Balance Patch is noticeably harder than unmodded Civ V BNW and has crushed me a few times. (I still remember jumping when Erik the Viking showed up on my screen and declared war on me even as he was busy taking out the last of Ethiopia).

The game will be unbalanced at release. I simply expect this. The game has too many components to reasonably expect not to find some exploits. This has happened with every civ game, and in fact every strategy game I've played in the past 10 years. The final product will be something different. Hopefully we'll get more balance patches than we got with BNW (I'l confess the final state of balance in that game was highly disappointing). But as long as mechanics are solid I know I'll have a game I'll be spending 100s of hours on.

This is precisely why I AM worried.

Why should we have to wait another 3 years for modders to make the AI anywhere near decent again with a community patch ?

I initially thought it was the 1UPT concept which made the AI just incapable of being able to war against a human player and it just wasn't possible to program the AI to compete against a human, no matter what the difficulty. The community patch changed that so surely the Devs must have seen what they did to help program the CIV6 AI ??

To the OP I dont recall anyone expecting an AI which is more intelligent than an experienced human player, that's what difficulties are for. But surely after seeing what the community can do we can expect an AI that doesn't send unguarded settlers around the island, doesn't send builders out after declaring war on a human to be instantly captured, doesn't declare war with no army at all, doesn't send 1 unit as their attack force, does upgrade its troops etc etc ?

Also where are these "epic AI battles" Ed Beach claimed the Firaxis teams were watching in May ? I haven't seen anything epic about the AI so far.