Battle Royale exposes the weakness of AI.

[SupremacyKing2]

The AI focuses on agendas a lot. And in today's AI battle, Anton and Pete said that the AI no longer looks at who is close to a victory condition when determining who to attack. So, I get the impression that the devs purposely designed the AI to role-play rather than win.

 

[Slayan]

Because its supposed to be a STRATEGY game.

Strategy game doesn't mean that all the AI players have to take over all the city's

 

[rpgalon]

The AI just looked like it didn't want to win AND did nothing to avoid losing, it was just there with no interaction between them.

Actually the only part of the game that looks to have AI interaction is Religion, maybe they wasted too much lines of code making the AI play that stuff.

 

[Magil]

Notice how the Aztecs had tons of builder sitting around for basically the whole game? Imagine if they were a Civ with a unique ability to utilize builders in a way no other Civ could, giving them a way to easily make use of any extra builders they had lying around...

Firstly, comparing BtW Civ4 AI with vanilla pre-release AI is ... I don't know. That's 'skewed' perspectives if I ever saw one.

The AI doesn't have to play a substantially different game in VI than V. One of the main selling points of VI is that it isn't releasing being gutted of the features of its predecessor, as V was (remember how vanilla V launched with no religion, no espionage...)--so is it so much to expect the AI to be able to play the game a little?

More to the point, they made the stream very uninteresting because the AI just sat on its hands for most of it. The only things it did worth talking about are the blatantly stupid plays. Otherwise, it was just generally passive and non-interactive. I'm honestly confused how this could be considered acceptable by the developers of the game.

I don't care about the AI wanting to win/not wanting to win. That's irrelevant. But I would like to see the AI try to accomplish various goals. Conquest is absolutely one of those goals. Why are leaders like Gorgo and Montezuma so passive? They should be some of the biggest warmongers out there.

 

[Lucius_]

Notice how the Aztecs had tons of builder sitting around for basically the whole game? Imagine if they were a Civ with a unique ability to utilize builders in a way no other Civ could, giving them a way to easily make use of any extra builders they had lying around...

Not only that, but repairing the pillaged tiles in the conquered Russian cities wouldn't have hurt.

All in all, I see it as a mixed bag. Some of the tech progression I thought was impressive, but the win time seemed really late to me, even for King difficulty. I would have expected around t400 win. Like others here, I'm pretty disappointed with the AI's ability to war. Rome should have been able to take cities from Japan with that army. The camera goes so fast, it's hard to see all the details of what happened there, but to me it seemed like legions should still been relevant and Japan's army was not significant. Also, the Aztecs seemed to take awhile to take Russia's cities. It had the units, and the tech advancement, and could have swarmed those satellite cities and taken them in a few turns, but that war just seemed to go on forever with nothing happening.

 

[Panixs]

Notice how the Aztecs had tons of builder sitting around for basically the whole game? Imagine if they were a Civ with a unique ability to utilize builders in a way no other Civ could, giving them a way to easily make use of any extra builders they had lying around...



The AI doesn't have to play a substantially different game in VI than V. One of the main selling points of VI is that it isn't releasing being gutted of the features of its predecessor, as V was (remember how vanilla V launched with no religion, no espionage...)--so is it so much to expect the AI to be able to play the game a little?

More to the point, they made the stream very uninteresting because the AI just sat on its hands for most of it. The only things it did worth talking about are the blatantly stupid plays. Otherwise, it was just generally passive and non-interactive. I'm honestly confused how this could be considered acceptable by the developers of the game.

I don't care about the AI wanting to win/not wanting to win. That's irrelevant. But I would make to see the AI try to accomplish various goals. Conquest is absolutely one of those goals. Why are leaders like Gorgo and Montezuma so passive? They should be some of the biggest warmongers out there.

I agree if I had had a guess at around turn 50 I would have throught that Monty would take out Brazil and possibly Russia and Gorgo should have attacked Spain or Rome. Also an AI delevoper should not be happy with the fact that when two civs declare a joint war no one attacks because they dont want to be first and peace gets issued 10 turns later, he even called it as soon as the joint war decleration happened

 

[ExemplarVoss]

Strategy game doesn't mean that all the AI players have to take over all the city's

But they should be able to found cities at good city site near their capital, not burn 20 turns walking to the other end of the continent before they build a second city.
Nor should they not attack after declaring war.
Nor should they build settlers then not use them, or fail to repair pillaged tiles, or immediately give cities back in the once in a blue moon scenario when they do take them, nor build spaceports but then unaccountably fail to build the projects to achieve the space victory, or spam mech. infantry instead of doing so when they refuse to go to war in the first place.

The only thing I was impressed with was Brazil's megafarm... but I have no idea if it had sufficient production to go along with that.

Actually the only part of the game that looks to have AI interaction is Religion, maybe they wasted too much lines of code making the AI play that stuff.

Sadly, like the domination victory, that seemed impossible for the AI to win, especially with the two continents. They just spammed back and forth, and some civs got converted, then converted back, and converted again. It seemed like an endless cycle with no way of winning... except not to play.

Culture actually seems the inevitable win, as it is just a matter of accruing a numeric threshold. Any computer can do that- addition and subtraction is pretty much the basis of everything. Science is also possible, but it would require the AI to focus on that and be coded to build the next step in the chain. It looked like each project was disassociated and they'd have to reconsider the project from its 'natural' place in the priority tree for each of the 5 steps (not including building the spaceports in the first place, let alone doing what a player would do and build 3), which means long stretches of not pursuing the victory condition.

Also an AI delevoper should not be happy with the fact that when two civs declare a joint war no one attacks because they dont want to be first and peace gets issued 10 turns later, he even called it as soon as the joint war decleration happened

Out of everything, that gets me the most. They're perfectly content with how terrible it is, and can anticipate it being terrible. Have some professional pride in public, guys.

 

[oakdragon]

Personally, I didn't think the AI was that bad. I don't agree that domination victories are the be-all and end-all, so I don't find it annoying when AI's don't go for them. I found the developers' explanations of what was behind their moves usually made a lot of sense and it was hard to really tell what was happening sometimes with the icons turned off, so some things are just guesswork on our part. Other than the joint war issue, the one thing that bothered me really was that the AI's still don't seem to know they can disband units when they no longer need them. That could've freed up a lot of gold for other things.

 

[ToothedBomb]

Notice how the Aztecs had tons of builder sitting around for basically the whole game? Imagine if they were a Civ with a unique ability to utilize builders in a way no other Civ could, giving them a way to easily make use of any extra builders they had lying around...

Aztecs do have the ability to rush districts with builders, right? Doesn't that count?

 

[Magil]

Aztecs do have the ability to rush districts with builders, right? Doesn't that count?

Sorry, I was being sarcastic. Doesn't always translate well over the Internet.

 

[ToothedBomb]

Sorry, I was being sarcastic. Doesn't always translate well over the Internet.

Lol, I guess I'm almost as dumb as CiVI's AI

 

[nokmirt]

blah blah blah.

really tho I see a better AI.

Why does the AI have to take city after city to be "Great" in your eyes

It does not need to be Alexander the Great, but it cannot hardly take and hold a city period. That's why.

 

[darko82]

It's ridiculous how fast they get to the modern age. And the AI is a total mess. It can't play. And it's so passive. It cannot take a city. Incredibly weak!

War in this game does not make any sense.

Culture victory? If a human played this game, he would have won culturally long before the year 2000.

There are no true opponents in this game. No challenge at all. You play against yourself, not against the enemy.

Even diplomacy makes no sense if the AI is so weak.

 

[mzprox]

That AI guy should be replaced. There are no excuse for some of the behaviors.

 

[Xyriach]

If the AI available in a video game could play as well as a human, it does not point to impressive AI, but to an unimpressive human.

 

[narmox]

That is pretty much my take away from it. They DO know. They just find it acceptable to be awful at their jobs.

More than likely, it's the bean-counters who find it acceptable to release a game on a date that was announced months before and have the devs patch the game afterwards.

Also imagine the outrage if the game was delayed.. Lose-lose situation at this point for the devs. I sympathize.

 

[itkovian1128]

I had not counted on AI which will match a man but a point which would be interesting playing as in the Europa Universalis 4.

There, AI is able to take advantage of weakness to attack, no mess, and the world gives the impression of enjoying a consistent player.

I see here that the AI how to build an army is not able to attack when your advantage.

Skipping flowers like an army of unused workers at the Aztec army catapults with the Russians, or the army of warriors in Brazil.

Castrating game with elements of conquest in theory, nothing aggressive civilizations are able to plan the war. So I write it again war is part of civilization as diplomacy, culture, science or religion.


These things unfortunately offend and spoil fun of the game.



One can only hope that it can be repaired or that during the Battle Royale something jammed in the script ...

 

[Frank327]

More than likely, it's the bean-counters who find it acceptable to release a game on a date that was announced months before and have the devs patch the game afterwards.

Also imagine the outrage if the game was delayed.. Lose-lose situation at this point for the devs. I sympathize.

That's the impression I got as well, especially because they mention very often how much work was done in the past month. Seems like they're in too much of a rush to release it. On top of that there was a bit of a contrast between the first half hour of the stream from 2k and the rest, which was done by the devs. Maybe they're not on the same wavelength.

But for whatever reason, the AI is really shoddy right now. I'd totally accept an AI that is lacking in one or two areas, because it suggests that it can be fixed in a patch. In the stream though, too much went wrong. Expansion choices were poor, workers were poorly managed, complete lack of decisive aggression and capability to win wars, lack of self defense and I even thought the city development was lacklustre. There were too many river grasslands next to city centres undeveloped into the information era and in other cities too much focus on one tile improvement over everything else.

 

[Zhahz]

Forming an opinion based on a single AI vs AI "royale" when that's not even how civ is played exposes the weakness of flawed debating.

 

[Frank327]

Forming an opinion based on a single AI vs AI "royale" when that's not even how civ is played exposes the weakness of flawed debating.

We're not just judging from one game because these same issues were present in games played on the pre-release version.

AI vs AI is definately a big part of how the game is played because if I'm on the other continent, the ohter civs are stuck playing AI vs AI for a very long time. And most of their diplo relations will be AI vs AI and on top of that many issues weren't caused by the fact that there was only AI on the map. I'm not gonna repair spain's tiles in my games for example.