[GS] Devs discuss AI, World Congress and other things

There’s a thread on Reddit where the FXS Devs discuss the AI and other issues. Link here. You can find some other quotes searching the Dev’s profile.

Some quotes below:

***​

“Both of these changes would fall into a camp of feature that I've tried to avoid with all of my work on Civ 6: they would be features that do absolutely nothing when used on a human player.

This is something I avoid for two reasons: 1, the obvious multiplayer implications of the game containing features that simply don't apply in that context. 2, the fact that it gives the player more tools than the AI has.

Your comment about Civ being singleplayer, with multiplayer slapped on top, is sadly kind of true. But it's something that we're trying to shy away from: we want multiplayer Civ to be as well-supported as singleplayer Civ. This is why more concrete effects have been added to systems that in base game were only for manipulating the AI: diplomatic visibility now affecting combat, alliances now having strong in-game effects beyond locking in your diplomatic state with the AI, etc.

And, for point 2, we're very aware of the feedback that we're constantly receiving about the level of play of the AI. A lot of people look fondly back on the difficulty of the AI in games prior to Civ 5, but with 5 and 6, we made a change to have the AI actually play the same game as the human player. Previously, the game would just magically grant units and techs and such to the AI players in order to keep them competitive. Now, we may sneak the AIs extra yields or have them start on better footing, but they're not then cheating with those resources.

This does however mean that the AI is having a harder time playing the game, because it's really difficult to make an AI that plays well enough (and poorly enough) for the range of people who want to play against it. This gets worse when systems are added to the game that exist solely for the player to exploit.

This is why I settled on the system coming in with the next patch: all players get to see what other players voted on last time Congress voted on this Resolution. It's data that anyone can use (granted that data is more useful against an AI than a human player, as the AI are more likely to vote the same way again), and it doesn't require players making and keeping promises of "I'm going to vote on this next Congress".

Incidentally, that last point is part of why the oft-requested "let me barter with other players to get them to vote the way I want" isn't being implemented (and is instead covered conceptually by trading Favor).

Thanks for the feedback!”

***​

“There's definitely an RP element to single player Civ, which is actually one of my personal most fun parts of the game. Roleplay versus competition, single player vs multiplayer... the game is played by a lot of people who play for different reasons, and it's our job to come up with new systems and effects that make the game better for everybody, whenever we possibly can. With the systems I've been developing, I've been trying to think about them from all angles of the playerbase, and land somewhere comfortably in the middle. I realize that this means that the system isn't the very best it could be for any one of those playstyles.

As for indirect clues in this case, that still 1) only works against AI (the game can't give you indirect clues as to how a human player will vote) and 2) doesn't actually add any new information to the game, instead making already-available information more obvious.

We've already discussed (1), but (2) is problematic for a few reasons, as well. Firstly, from strictly a production/codebase standpoint, getting that sort of information from the AI and making a new screen to show it off is a lot of work. Currently the AI makes choices for myriad reasons, and then throws away the reasoning and just returns what it wants to do, how much it wants to do it. It would require a lot of rewriting and new plumbing to expose that information. And, on the UI side, World Congress is already a very full UI, and I don't think that another system adding more gossip popups to the game would be well received.

Secondly, currently digging out that information, weighing it, and making a guess based off of it is an expression of skill. It's part of the strategy in World Congress. If we were to implement a system that basically tells the player "this is how people will vote", or that does that digging, weighing, and guessing for the player, then that takes out some of the skill in playing that part of the game.

Thirdly, it's really hard to make "hint" systems that don't feel like they're lying to the player. When making a hint system like this, you wouldn't want it to just give away the answer. You wouldn't want it to only say "Harald is keeping an eye on Nubia..." when Harald is in fact going to vote A against Amanitore. But players, when they see that line, are going to feel pressured by the game to make that assumption, and then when it turns out to not be true, they're going to be (a little) mad at the game for lying to them. It's much better for the player to feel like they made the wrong strategic choice (because they went out into the game map, saw that Amanitore had low loyalty, and assumed that Harald would want to use option A on her) than for them to feel like the game is misleading them (because a line popped up saying that Harald is likely to use A on Amanitore).

Hope that helps give some insight into why I've made certain tradeoffs in the system!”

***​

“If you take a look through my post history, you'll see something of a manifesto on my stance on displaying AI information to the player in World Congress, as well as a portion on the point you brought up in the comments section about making certain information more obvious to the player.

I did want to address a different piece of feedback here, specifically the desire to influence which Resolutions are available at each World Congress. There is no plan to move in that direction for a couple of reasons:

1) It would be a balance nightmare. Currently, the various Resolutions aren't particularly balanced to each other: is it as nice to have the option to vote on a potential Unit Strength bonus as to vote on whether or not Units can be bought with faith (again)? Well it doesn't matter currently, because the game picks what you get to vote on, and you can decide to save your Favor if you don't like what's offered. In the current system we can have that sort of disparity, but if players were voting on or proposing Resolutions, we'd see the powerful ones pop up every time.

2) It would be an AI nightmare. You've certainly seen the cases where all the AI decide that Option A targeting City Center is definitely the right option, and there's not much that you as one player can do about it. Imagine if there were a layer on top of that where they all decided, every 30 turns, that they really really cared about Urban Development. You, as one human player, would have very little say against a block of votes like that.

Yes, I realize that both of those reasons for not doing this could be maliciously boiled down to "cuz Firaxis is lazy," but trust me that when making decisions where we're putting our manhours, we're doing our best to give our players the most value we can. I believe that the time that would be spent on balancing the Resolutions relative to each other and building the AI to select them in a fun-for-the-player manner and doing all the testing to make sure we've hit both of those goals, on top of the UI flow and various other ancillary work for that, is better spent on other systems.”

[Emphasis added]

Feel free to discuss below (play nice please).

But before anyone comments on the details, can I just add that I really appreciate the Devs taking time to explain things like this. I genuinely love hearing about the process of the game’s design and it makes the game experience much better knowing why things work in a particular way.

[edit: corrected quotes.]

 

It seems like you quoted the second thing the devs said twice versus quoting both things he said.

 

Fixed.

 

Sadly it's what it's probably going to push me away from the franchise... All these abstract systems and bonuses, hard rules which prohibit something that makes no sense to not allow, all this because it can be an "exploit" in MP. The lack of resources towards things that are SP only (throne room/palace, demographics, etc). Anyway, just venting a bit my personal frustration on the direction the franchise is going.

 

Can you be more specific? IE: They added combat bonus for diplomatic visibility, which adds utility to MP and an extra layer to SP. Exploits in MP are just as exploitable in SP if not moreso.

 

I think what he means, which is a concern to me at least, is that the game will be designed in a way which is less optimal / fun for SP because it has to work in multiplayer too. The change you mentioned is not of this category. A spy mission to get insight into the AI voting is an example FXS gave themselves of a feature that wont happen because of MP. Im not particular concerned with that specific feature, but I wonder what nice features gets axed before we see them because 'they wouldnt work in MP'.

I wouldnt say this risks to drive me away currently, but it is still a concern.

 

Eh, that was one of many reasons. There was also the conundrum of how do you ‘hint’ without just telling the player outright, and what happens if the ‘hint’ is wrong, and also the perspective that they would be giving the player predictive information on the AI that the AI would not have on the player (which fundamentally already exists in core gameplay elements that are easy for a player to abuse, they don’t want to add more).

I’m also in the camp of playing both modes, so I can better appreciate the value of features that apply to both.

 

Well, I found what @FXS_draabe said fairly reasonable.

In particular, I think the comments about having to decide where best to put man-hours are very fair.

Anyway. Some further thoughts for what it's worth...

AI: I thought the comments about Civ IV v Civ V and VI pretty interesting and make sense. Although, you know, I'd be okay if FXS did throw the AI some free techs or units to keep up too.

Good to know FXS have heard people about the AI. It seems to me they are trying to improve things and, hey, you know it is getting better.

SP v MP, RPG v Game: seems a reasonable position to me. I think it makes sense FXS focus on creating (I guess) a "base game", and then if people want the game to lean into MP more, or RPG or whatever, then that's more the role of Mods.

I do ultimately think MP needs its own rules, because the balance requirements are so different. But I think that really has to come from modding, because those rules need to be so responsive to MP players.

World Congress [0]: I much prefer having random resolutions anyway. I think random resolutions makes for better gameplay than players picking resolutions - it’s a good mix of random (what resolutions you get) and player controlled (how you vote and how you spend favour), so you’re forced to adapt and be opportunistic. It also allows for asymetrical resolutions - ie some are more powerful than others.

I also think the ability to call Emergencies adds enough player control into the mix to feel like you do get enough of a say about resolutions overall.

That said. As I've said before, I think the current resolutions are too narrow - buff or nerf this particular Civ, or ban or buff this particular luxury, or buff or nerf this particular unit or building or spy mission. Resolutions should be wider; more like buff or nerf all Civs with a particular government, or ban or buff “animal products” (furs, ivory, whales) or “conflict stones” (diamonds, gypsum), or buff or nerf classes of units, buildings, and spy missions (eg Defensive Militias (Anti-Cav, Ranged, Scouts), Military Infrastructure (Harbour and Encampment buildings, Walls), Industrial Sabotage (rocketry, Industrial Zone and Dam missions)).

I think that would make balancing the pros and cons of voting for a particular resolutions feel more interesting and nuanced, make resolutions more impactful and game changing, and make resolutions feel more real / historical / believable.

I also think targeting particular government types rather than particular Civs would make for really really interesting decisions - first, because if you vote to target a particular Civ (ie by targeting its current Government) there would be the potential for other Civs to get targeted as collateral damage (including maybe allies), and second because everyone would then have to decide whether to swap out of that government and into something less optimal (but avoid the nerf or gain a bonus) and or keep their government but suffer a negative.

[0] Full disclosure, I'm paraphrasing here what I posted on Reddit in reply to Devs.

 

I guess they are mimicking a human player getting bored sometime around 1AD.

On a side note: Sid Meier is on record having said that people would not be playing games like Civ if they had friends (in regards to why he didn't make Civ 1 multiplayer).

 

I think one of the main problems of the 4x genre is that they are mulitplayer games played primarily in single player.

The fact that the player and the AI is symmetrical is one of the reasons the AI struggles so much.

Makes me wonder what a 4x game that builds asymmetry with the AI into its core would be like. You would have so much more freedom in how to design the mechanics and have a much larger variety of different ways to provide challenge to the player

 

Yes. If I wanted multiplayer games I would try to be friends with people.

 

Regarding SP vs MP, me (and my friends) play only/mostly multiplayer games.

It's simply more fun, singleplayer civ is just very boring compared to multiplayer.
Although that also has to do with the -predictable and exploitable- AI, of course.

I'm actually quit sad that multiplayer is just an afterthought in Civ. And it is noticeable in many ways.
A good example is saving a game configuration in the lobby, this only came with the last patch I believe, while it was in Civ for a very long time (granted not at launch).
Or something like banning certain civs from games, which isn't possible at the moment.

I wish multiplayer would get some more love from Firaxis.

 

I would be happy if they try to balance it just a little bit though. Like the luxury amenity one being completely meaningless.

 

To address Firaxis’ defense re: AI deficiencies that AI before Civ V would get free units etc—is it true as such that the Civ IV AI cheated? Because I don’t recall seeing or noticing that at all. Meanwhile in a Civ VI game the AI do get extra settlers on higher difficulties...

 

Perhaps a more global one, like Amenities applying to just two cities per amenity or six per amenity, which would still be impactful and could be used to curb warring civs or allow high amenity civs to expand rapidly.

The one that I find problematic is the diplo points vote. There is no incentive in either SP or MP to vote anything other than against the leader. I think the devs are trying to replicate the world leader vote from CIV V but I'm not sure it works as is.

 

I'm just thrilled that the devs took some time to communicate these perspectives with us.

Their comms have gotten so much better over the past year or so.

 

From what I can remember, there were more things that the difficulty affected in Civ IV. I think the difficulties were set up the same, that the AI would get stronger bonuses for every difficulty level you went up. But I think there were simply more things that it affected in IV. For example, going up in difficulty level would result in the AI having additional starting techs. And I think there were also some things that would affect the player adversely, so as you went up in difficulty it would increase the chance of getting barbarians pop from goodie huts or similar things

Please take this with a grain of salt, this is only from distant memory so I may have dreamt it all. I'm sure modders have the info if you google it maybe.

Even so, I think the main difference in difficulty levels are how the games are set up. Civ IV is an easier game for the AI to play. Civ VI is harder with its non-stacking, the use of ranged units, city walls ranged attacks, district placements, wonder placements. All that stuff. There are a lot of things to keep up with and you can really tell how the AI sorta loses its teeth as you go further into a game.
I think that's the real issue.

EDIT: And yeah, the AI in Civ IV was never as advanced like *some* people claim. It did a lot of stupid ****. Like a lot of it. But again, the way the game was built, the rules it had and so forth, made it so that the AI could be a threat.

 

This whole point answers a HUGE amount.
All those civ IV player posts whinging about a bad AI while all the time their version was cheating it’s pants off
It does bring up a very important fact, that 2 players in civ VI playing the same game could have quite different levels of cheating used against them which in a way invalidates their competition games. Intriguing.
This one clarification really limits their argument about a bad AI compared to VI.