The AI on Sovereign Difficulty

[ggalindo001]

I will first say that I had been genuinely excited about Civ 7 until the last livestream. The last livestream really left me empty and feeling like the AI in many respects may not be as competitive as I would like, especially at a higher difficulty level.

Granted, my observations are based on a handful of turns, on what was likely a very carefully curated 'scenario' -- so I hope I am just in the wrong.

Observation 1 -- Xerxes -- He becomes Mongolia, has an ability to get "treasure points" on land and on his own continent, relocates the capital to within a stone's throw of Isabella's relocated capital -- and absolutely nothing. Ended up having a cog out there in the wild blue yonder that Carl declared war on to wipe out. I would have thought that a capable higher level AI, even as "behind" as Xerxes was, and becoming Mongolia, would mount -some- type of land challenge against Isabella. Even FXS_Sar was concerned when the capital was first relocated -- but to the best of my watch and review, he was a non-threat. Instead, we can just build Cogs look for treasure, and apparently not have to worry about a rather dangerous neighbor within easy striking distance, with supposedly agenda and incentive to come directly after us.

Observation 2 -- The whole "treasure point" mechanic -- which on the surface is nice, but it may be blinding the AI into one path where other paths (again, conquest) may be more appropriate or at least opportunistic. I got the feeling like the AI was running a programmed script without regard to the game scenario. There didn't seem any sense of urgency other than running the "game script" -- Confucius was just basically chasing the treasure ship -- got "mad" at a settlement on it's continent, but wasn't prepared to do anything about it other than chase the treasure ship.

Observation 3 -- In looking at a screenshot where it showed gold, beakers, etc. across the leaders -- it was all over the place -- which is good, with some civs higher on science, some higher on culture, etc. However, the game didn't appear to reflect this reality.

Observation 4 -- Not AI related, but to me, the mechanics feel very flat. It almost feels like golf -- I may be playing against 4 or 5 other civilizations, but I am playing my game and hoping my game is better than their game -- but with no real material risk of their games interfering negatively with my game. I want to feel like if Xerxes as Mongolia is a real threat to my ability to play my game, not that I just need to "do better" than them on treasure points.

Again, I am very much hoping this was a very scripted game and not truly reflective of the AI capabilities at this higher level. But given Civ 6, I have my serious doubts.

 

[AriochIV]

It's very early to be making these kinds of judgments, but... folks who are expecting really good AI in Civ7 are very likely to be disappointed, if history is any guide.

I appreciate that they have a larger AI team, and it would be nice if they improve the AI so it's not as bad as it has been in Civ5 and 6, but I don't expect "good" AI.

Also... higher difficulty level above normal almost never makes the AI smarter, just gives it more bonuses.

 

[Atlas627]

Keep in mind that the AI has to play the final game well, and they're still testing and rebalancing the game. The AI team has been putting in a lot of work to build a framework for playing the game, but it still needs to be refined...after they finish balancing the game and are on the road to release.

 

[bbbt]

I mean, these legacy paths are largely how the science victory works. Basically hit target A then B then C.

Frankly, if the AI is reliably able to do that at launch, that's not a bad start, at least compared to 5 and 6 at launch.

Having be able to wage war well and do disruptive actions would be great, but I won't hold my breath.

Unfortunately, I think 95% of the AI work for a game like this at launch, is simply having the AI recognizing and using systems period.

 

[Disgustipated]

I'm most worried the AI will behave too similarly to Civ 6. I want my Mongolia to be aggressive. I worry they are too reluctant to put truly warmongering civs/leaders into the game. Like older civs where aggressive was their trait. Agendas are fine for flavor, but I also feel that there needs to be an additional modifier to AI's personalities based on their leader/civ choice.

Most likely Mongolia was just exploring with that cog, which is fine. But as I said, some civs just need to be more inclined to go to war.

 

[Evolena]

The name of that third Persian settlement (Djedet? which became the Mongolian capital city) indicates that Xerxes had captured it from Egypt. From he's starting position behind a lake and mountains, that's not nothing, right?
Also, 2/3 AI got to the end of a Legacy path in Antiquity, and the last save around 60 turns into the Exploration Era showed that most AI were capable of making good progress there too, I think that's encouraging (of course with 4 empires concurring on a single rather small land mass, all cannot thrive and it's a shame that it happen to the warring one, but I thing the starting position was not favorable).

 

[Mr Jon of Cheam]

On the plus side, the AI appeared to be doing pretty well at keeping up on the Legacy Paths, but I must admit that I'm also a little concerned about the AI following these paths too relentlessly so that all leaders and civs feel the same.

If you play against a person and they're behind, they won't simply try to catch up by keeping to themselves, they will recognise the need to disrupt things and start planning accordingly. Hopefully the AI leaders will do this but in character; e.g., you want to feel that Machiavelli is plotting and using diplomacy as a weapon, whereas with Genghis, there should always be the threat of war.

It's a difficult thing to get right and I'm sceptical, but we shouldn't forget that this livestream was a marketing exercise. They had things that they wanted to show and marks to hit, they were never going to allow it to descend into an all out war on the homeland so that they didn't have time to show off the distant lands. I wouldn't read too much into it yet.

 

[Siptah]

On the plus side, the AI appeared to be doing pretty well at keeping up on the Legacy Paths, but I must admit that I'm also a little concerned about the AI following these paths too relentlessly so that all leaders and civs feel the same.

I fear the main goal of the AI is to be competitive in the legacy paths and victory conditions. That way, the devs can say that the AI is good in playing the game (as it racks up score) and competitive with the player (and if this is by getting free progress/having smaller goals on high difficulty levels). Despite being still terrible on the map with city and unit management.

Hopefully the AI leaders will do this but in character; e.g., you want to feel that Machiavelli is plotting and using diplomacy as a weapon, whereas with Genghis, there should always be the threat of war.

While it makes leaders predictable, I also long for this to be a thing. Give leaders personalities and weights, not just agendas.

 

[Civinator]

I fear the main goal of the AI is to be competitive in the legacy paths and victory conditions. That way, the devs can say that the AI is good in playing the game (as it racks up score) and competitive with the player (and if this is by getting free progress/having smaller goals on high difficulty levels). Despite being still terrible on the map with city and unit management.

Yes, the old tactics of the civ series, to give the AI big bonuses compared to the human player, seems to be very real.

 

[Prester John 2]

I'm most worried the AI will behave too similarly to Civ 6. I want my Mongolia to be aggressive. I worry they are too reluctant to put truly warmongering civs/leaders into the game. Like older civs where aggressive was their trait. Agendas are fine for flavor, but I also feel that there needs to be an additional modifier to AI's personalities based on their leader/civ choice.

Most likely Mongolia was just exploring with that cog, which is fine. But as I said, some civs just need to be more inclined to go to war.

I guess the problem with militaristic civs that are building up a large army is that if they don't use that army to actually conquer other cities it effectively puts them at a disadvantage compared to civs that focus on science, wonders, economy or culture and disregard the military. Armies are a sinkhole investment unless you gain new land and ressources with it. If you have two militaristic civs next to each other who fight that becomes a zero sum game. Turtling science civs are much better equipped to pose a challenge. But even they won't disrupt a linear or exponential progress a human player can achieve. In Civ6 you don't even need to outright conquer a rival. It suffices to plunder all their districts and improvements.

 

[Boris Gudenuf]

I guess the problem with militaristic civs that are building up a large army is that if they don't use that army to actually conquer other cities it effectively puts them at a disadvantage compared to civs that focus on science, wonders, economy or culture and disregard the military. Armies are a sinkhole investment unless you gain new land and ressources with it. If you have two militaristic civs next to each other who fight that becomes a zero sum game. Turtling science civs are much better equipped to pose a challenge. But even they won't disrupt a linear or exponential progress a human player can achieve. In Civ6 you don't even need to outright conquer a rival. It suffices to plunder all their districts and improvements.

Machiavelli's great quote applies in Civ as well as Rennaissnce Italy, and it applies to Civ armies as well as Italian mercenaries:

"Gold may not get you a good army, but a good army can always get you gold."

Use it or lose. There is not, to my knowledge, any victory condition of "has the biggest army". You have to use the troops to get what you want or need.

And that doesn't even mean a really large military force, either. Another quote, from Maurice of Saxony (probably one of trhe best French commanders after Napoleoni, so he knew what he was talking about):

"It is not the big armies that win wars, it is the good ones."

 

[Disgustipated]

I guess the problem with militaristic civs that are building up a large army is that if they don't use that army to actually conquer other cities it effectively puts them at a disadvantage compared to civs that focus on science, wonders, economy or culture and disregard the military. Armies are a sinkhole investment unless you gain new land and ressources with it. If you have two militaristic civs next to each other who fight that becomes a zero sum game. Turtling science civs are much better equipped to pose a challenge. But even they won't disrupt a linear or exponential progress a human player can achieve. In Civ6 you don't even need to outright conquer a rival. It suffices to plunder all their districts and improvements.

True enough, but I was thinking in terms of variety. It would be boring if all civs focused on science and culture. It helps to have an aggressive civ spice things up. It asks the question, do we want all civs to try to win? Or to sometimes act irrationally to spice things up.

And I was thinking in Civ 6 how Alexander would be mad at you for not being at war, when he wasn't at war himself. LOL It was weird seeing civs like Macedon, Zulu, and Mongols not actually do any conquering.

 

[Josephias]

If you play against a person and they're behind, they won't simply try to catch up by keeping to themselves, they will recognise the need to disrupt things and start planning accordingly.

I feel (specially after Civ VI) that the AI is indeed playing to conservative and “by the book”, even when their chances are minimal.

i feel they are missing something like a “rage mode” where, when they fall to much behind victory, they change their objective to disrupting as much as possible the players they have most grievances with / worse relationship. - hopefully, as you say, in character.

I underestand they may be wary of such a mode as it may frustrate some players, but I think for others this will spice things up. Maybe this mode coud be made toogleable.

Think of it as the biggest threat comes from these that have nothing to loose.

 

[Stoic]

but seeing the livestreams, is it true that the difficulty levels have changed and they are no longer called immortal imperator etc? It seems that firafix wanted to make too much of a revolution in the game, also considering the gameplay changes.

 

[ggalindo001]

The name of that third Persian settlement (Djedet? which became the Mongolian capital city) indicates that Xerxes had captured it from Egypt. From he's starting position behind a lake and mountains, that's not nothing, right?
Also, 2/3 AI got to the end of a Legacy path in Antiquity, and the last save around 60 turns into the Exploration Era showed that most AI were capable of making good progress there too, I think that's encouraging (of course with 4 empires concurring on a single rather small land mass, all cannot thrive and it's a shame that it happen to the warring one, but I thing the starting position was not favorable).

I went back and looked, it is Djedet -- Ed Beach said that Xerxes "dropped one of this settlers way down there" at the 43 minute mark, however, I think you are right that this was a captured city from Egypt. And that would be very promising if this was reality.

I also agree that is it is way too early, and that we shouldn't get our hopes up -- I just would hope that the AI would somewhat mirror the advantages and strengths of the various player. Mongolia, with "incentives and bonuses" to capture settlements on their home continent, really should be something feared, not ignored.

I also agree this may have been a "marketing exercise" -- in which case things were scripted/toned down.

I'm not expecting combat and tactical AI quality of Old World (tho it is an aspirational goal) -- I just want to feel like I am not playing a solo game with 5 other "solo" players. That is just too much Civ 6 for me.

 

[bbbt]

i feel they are missing something like a “rage mode” where, when they fall to much behind victory, they change their objective to disrupting as much as possible the players they have most grievances with / worse relationship. - hopefully, as you say, in character.

I underestand they may be wary of such a mode as it may frustrate some players, but I think for others this will spice things up. Maybe this mode coud be made toogleable.

They made a (poorly done) stab at this in 6 with a negative modifier if the player was ahead ("They hate us because we are winning.") but it didn't do much and was '4th wall' breaking.

I agree 100% on a toggle/slider - essentially "AI plays like a competitive player" on one end and "AI roleplays on the other". You think it could be done with a better implementation of that sort of modifier like in 6.

Since the essentially question for that is basically - If you and Gilgamesh have been allies since the ancient era, should he turn on you and start nuking you because you are getting close to victory? I suspect a lot of people do not want that.

 

[JNR13]

but seeing the livestreams, is it true that the difficulty levels have changed and they are no longer called immortal imperator etc? It seems that firafix wanted to make too much of a revolution in the game, also considering the gameplay changes.

Might just be the gendered difficulties getting more neutral terms - King, Prince, Emperor.

 

[ggalindo001]

They made a (poorly done) stab at this in 6 with a negative modifier if the player was ahead ("They hate us because we are winning.") but it didn't do much and was '4th wall' breaking.

I agree 100% on a toggle/slider - essentially "AI plays like a competitive player" on one end and "AI roleplays on the other". You think it could be done with a better implementation of that sort of modifier like in 6.

Since the essentially question for that is basically - If you and Gilgamesh have been allies since the ancient era, should he turn on you and start nuking you because you are getting close to victory? I suspect a lot of people do not want that.

I like the idea on a slider -- but maybe multiple sliders. I want a more competitive player and more likely roleplay.

To address -- in my mind, if I have been allies with Gilgamesh since the ancient era, backstabbing for no good reason wouldn't sit well -- however, and similar to the livestream, if I do not have good relations with Xerxes, and they are Mongolia, and they have a roleplay to take settlements on our shared home continent -- I want them to be someone feared.

Also on the Gilgamesh theme -- if we have been allies, but I become a democracy and they become communist, that should create "tension" in the 3rd age. Hopefully that mechanic will exist.

 

[Nigel_Tufnel2]

Playing a VI game now, had a near-death Horseman next to a capital. Next turn, AI shoots at an Archer with an Archer in the city, and 2 of his horsies swap places, neither of them finishing off the easy picking Horseman, nor finishing the easy to reach Archer.

Sure, the AI may deliver troops to the front quicker and at one time with Commanders - but if the actual tactical battles keep using VI or VI-like code, it will be a total joke once again. I saw on the livestream that the foreign ships sailed away from danger. Definitely tells us nothing...

 

[Trav'ling Canuck]

Playing a VI game now, had a near-death Horseman next to a capital. Next turn, AI shoots at an Archer with an Archer in the city, and 2 of his horsies swap places, neither of them finishing off the easy picking Horseman, nor finishing the easy to reach Archer.

Sure, the AI may deliver troops to the front quicker and at one time with Commanders - but if the actual tactical battles keep using VI or VI-like code, it will be a total joke once again. I saw on the livestream that the foreign ships sailed away from danger. Definitely tells us nothing...

Per one of Ed's early interviews on Civ 7, they haven't done much with the tactical AI. He felt it was in a pretty good place at the end of Civ 6 and didn't want the team spending time on it when there were higher priority things for them to work on. Ed's take was that fixing movement to the front via the new commanders was all that was needed to improve the AI's combat abilities. Per your anecdote above, I have my doubts about that, but time will tell.