Civ VII single player suffers from the same problems Civ VI did

[cabelhigh12]

Once again, the debate on whether AI should present competition or fun. ALL Civiization games have made the AI to be fun. They are speed bumps on your way to victory. They do not play to win. If you need that to have fun, you need to find a different game.

How is the AI not being able to use the new mechanics in Civ VII seen as a good thing in your eyes. I'm not saying the AI needs to be as competitive as a human -- all im saying is the AI needs to know how to actually PLAY THE GAME its playing.

The AI knew how to play the game extremely well in Civ4. This makes the game MORE fun not less, because it forces the player to sctually learn and exploit the deeper mechanics instead of coasting on by

 

[Shirotora Kenshin]

Don‘t want to write long texts in general, but i have to disagree with a few words with the threads title. I think Civ7 suffers from entirely other problems than Civ6 did. (and i love Civ6).

 

[UWHabs]

We do, you've just missed it because it looks near-indistinguishable from grassland (also unless you're also colorblind, this just validated for me that it isn't just my colorblindness that makes them hard to tell apart) and the actual yields don't matter.

Ftr I don't seriously mind that the terrain doesn't matter. I can see why Firaxis did it, I'm not convinced it's a good decision, but I also don't consider it a major problem.



I'd say both games have their issues. For me, adjacencies in VII are already a solved problem. The two tiles with the highest resource adjacency get production and science buildings, the two tiles with the highest coast/navigable river adjacency get gold and food buildings, and the two tiles with the highest natural wonder/mountain adjacency get culture and happiness buildings. Ageless buildings are used to get to those tiles. And wonders may complicate things slightly, of course. But I've also figured out already how to build wonders in relation to resources, wonders, coast and navigable river in order to maximize the number of relevant tiles that benefit from the wonder.

(side note: there are actually three gold buildings in the Exploration Era, so you might have to put one on a third tile - the last food and gold building only unlock right before Future Tech though, so unless you're ahead of the curve + play with a mod to extend the Age a bit more, you don't really encounter it)

Yeah, this is definitely the first civ game where you look at city placement options, and honestly, I don't even remotely consider the terrain. Not only because it's actually really hard to tell tundra and tropical apart, but yeah, it just seems like it doesn't really matter. Whether that's a good thing, I don't know. I do like the idea of the different biomes, but I do miss a little bit of a "oh there's a big patch of jungle, maybe I can get a campus there". Now, it's basically all about resources, and then stuff like rivers and mountains kind of ends up as a secondary concern.

But yeah, the simplified adjacency (everything is +1 instead of +0.5/+1/+2) should help the AI as well, and combined with the underlying terrain having less variation, even a dumb "always place a building on the best tile" and then worrying about where the population shifts if you build over a farm is probably going to at least simplify things a lot. Even wonders just giving you a flat bonus to everything, you don't have to do the civ 6 thing of trying to neatly pack your districts together, plan your entertainment complexes, figure out the 6 tile spread of factories, etc..

 

[King Flevance]

We do, you've just missed it because it looks near-indistinguishable from grassland (also unless you're also colorblind, this just validated for me that it isn't just my colorblindness that makes them hard to tell apart) and the actual yields don't matter.

Ftr I don't seriously mind that the terrain doesn't matter. I can see why Firaxis did it, I'm not convinced it's a good decision, but I also don't consider it a major problem.



I'd say both games have their issues. For me, adjacencies in VII are already a solved problem. The two tiles with the highest resource adjacency get production and science buildings, the two tiles with the highest coast/navigable river adjacency get gold and food buildings, and the two tiles with the highest natural wonder/mountain adjacency get culture and happiness buildings. Ageless buildings are used to get to those tiles. And wonders may complicate things slightly, of course. But I've also figured out already how to build wonders in relation to resources, wonders, coast and navigable river in order to maximize the number of relevant tiles that benefit from the wonder.

(side note: there are actually three gold buildings in the Exploration Era, so you might have to put one on a third tile - the last food and gold building only unlock right before Future Tech though, so unless you're ahead of the curve + play with a mod to extend the Age a bit more, you don't really encounter it)

See, this makes it sound easy on paper, and you are correct that this is the prime strategy.

However, founding cities with optimal placement for resources + fresh water impacts where the city center goes. This determines your "build path" and since you can't push districts to the other side of resources or around mountains you need to use ageless buildings to path there. However, if you are planning to have production tiles and there are only a few in this placement, carving through those tiles permanently destroying 1 of your only 4 production tiles is not ideal. Not every city will have mountains or coast/navigable rivers or enough production tiles. (Sometimes they only have 3-4 total farmable/fishable tiles.) But these usually have more production tiles and room for wonders if you dont decide to town them out.

I have expressed a desire to see something small additions to adjacencies but I think the system works well. It is easy to understand and offers variability based on map generation without handing someone an obvious advantage because mountains spawn in excess near them.
Adjacency bonuses are not always easy to just access to the ideal tile you want when you want it. Even using ageless buildings to expand. It is easy to make this system sound easier than it is to execute. Some cities are easy, yes. But I think many placement strategies often work in contrast to development strategies. And I don't think many people are acknowledging that, unless I am just getting maps that run counter to what everyone else is getting.

 

[ProMeTheus112]

I think the AI is much better than Civ6. Civ6 AI was suiciding everything in a war and very bad at tactics. Here they are a lot better, they're not amazing (they're AIs on random maps, they're not going to be amazing), but they're coherent and competent, surprisingly so to me. And yes now, BIG difference, they don't get extra settlers and units at the start, enabling you to go for just about any opening strategy that you want, which in turn enable you to go towards many more midgame strategies and etc. It's a much better strategy game that way than Civ6 was in higher difficulties imo, or at least it appears to me that way, I moved on from Civ6 after winning a immortal game (or two?).

Overall that was my comment and civ6 that AIs are too stupid in 4X games and you end up playing a repeated strategy of exploitation of their weaknesses which dumbs down the game, but other games where you play the environment have shown a different way. Even when the game isn't super hard to win, when they are well balanced they do open up a lot of possibilities and allow you to try and play and manage always better than the previous game, giving a real sense of progression.

So maybe Deity isn't super hard on Civ7 right now, but it's got a nice difficulty already, and has you solve a good good bunch of problems and make strategic decisions all along the course of the game, different each game. It's quite satisfying to me now, including the way AI behaves and attempts to pose a threat and play against your victory. I'll have to play modern age in deity some more to form my opinion more precisely, but I'm quite positive about civ7 gameplay atm, it's quite convincing.

I'd bring up Against the Storm as an example of a game that isn't very hard to win the games, but lets you make plenty of decisions and problem solving, always different, lot of room for improvement, with 20 difficulty levels.

 

[Formica]

To quote Sid Meier, from Sid Meier's Memoir! A Life in Computer Games:

"Highly realistic AI gets accused of cheating even more often than its dishonest brethren, because on some level, all players are unnerved by the idea that a computer could outsmart them. Part of the fun is learning the patterns of the AI and successfully predicting them, and when computers don't act like computers, the only psychologically safe assumption is that they must have accessed information they shouldn't have. AI isn't allowed to gamble, or behave randomly, or get lucky-even though humans do all of these things on a daily basis- not because we can't program it, but because experience tells us that players will get frustrated and quit, The same phenomenon doesn't happen when both opponents are humans, because they've already tempered their expectations for the possibility that the other guy is crazy. computers are too smart to be crazy, so if they start acting that way, we can't shake the suspicion that they know something we don't. Thus, from the designer's perspective, brilliant AI is usually not our highest priority."

Personally, I'd like to see the computer opponents in Civ 7 be more competent at using the new systems. And it would be cool if, instead of just adding combat bonuses, the computer opponents got better at the game as difficulty increased. But I also concur that the role of the computer opponent is not, necessarily, to try and beat the player. Instead, they are obstacles to be overcome on the way to victory.

I also wish that Firaxis would release the DLL source code for Civ 6 the way they did with 5. The resulting changes to the computer opponents in VP demonstrates that the community is capable of meeting that need if given the tools to do so.

 

[tedhebert]

All this adjacency talk simply reminds me how much this game misses map pins !

 

[universecreep]

WHAT is the point of introducing new and exciting systems if your opponents cant use them against you?

Civ VI was really bad for this. They introduced so many systems but never programmed the AI to use them effectively. V wasn't as bad.

I think Civ III and IV had the best AI; it seemed to "know" how to play the game and use all the available systems.

Unsurprisingly, Soren Johnson was heavily involved in both games. Firaxis should think about partnering with Mohawk games to make DLC for the game.

 

[ThERat]

Civ VI was really bad for this. They introduced so many systems but never programmed the AI to use them effectively. V wasn't as bad.

I think Civ III and IV had the best AI; it seemed to "know" how to play the game and use all the available systems.

Unsurprisingly, Soren Johnson was heavily involved in both games. Firaxis should think about partnering with Mohawk games to make DLC for the game.

I don't think this is ever going to happen. We don't know why Soren left Firaxis but he has his own games going and they are doing well I guess.

 

[cdstephens]

The problem here is that AI is not really AI. It is just a preprogrammed rulebook. They dont learn, they dont innovate.

Long time ago, i made an AI for an online 2D action game. It was quite good until someone figured out how to exploit it. I spent weeks on improving it, making it invincible, and finally everyone complained it is too good and felt artificial

Making a good AI is really hard.

Machine learning might be a cool tool for future AI development. Design the game, then have the AI play against itself until you have something that functions alright.

 

[infiniterocks]

Hopefully one day we'll get adaptive AI, that knows how to properly counter play a player.

 

[Mr Jon of Cheam]

Imo, I think they "only" need to improve two things in VII and the AI will be perfectly good enough: 1) settling, 2) tactical combat.

AI is fine at snowballing yields and completing legacy paths, good enough at least, but their settling behaviour is downright absurd, and completely erratic. You see them sending settlers off to the far corners of the continent, taking a huge number of turns in the process and all to settle very poor locations, even when they have plenty of good land to settle near their capital. Then they still prioritise building settlers even when the good land is all gone, so you regularly see settlers milling about pointlessly and it just reminds you that the AI is being dumb again. Fixing this will go a long way to improving our perception of the AI.

Second thing is more difficult but even some basic things would help. I've yet to see an AI use a Commander effectively. This needs work. They don't need to behave like a tactical genius, but the amount of units I see wandering about alone is ludicrous. They simply need to be better at packing units into commanders to move them, and then deploying them as required, with a focus on 1) protecting the Commander, and 2) utilising the command radius. I also do not understand why the AI is always so busy! Why oh why do they shuffle units around every turn? There is no need to do this unless preparing for or fighting a war, but you watch them just moving things for no reason.

 

[Leyrann]

Imo, I think they "only" need to improve two things in VII and the AI will be perfectly good enough: 1) settling, 2) tactical combat.

AI is fine at snowballing yields and completing legacy paths, good enough at least, but their settling behaviour is downright absurd, and completely erratic. You see them sending settlers off to the far corners of the continent, taking a huge number of turns in the process and all to settle very poor locations, even when they have plenty of good land to settle near their capital. Then they still prioritise building settlers even when the good land is all gone, so you regularly see settlers milling about pointlessly and it just reminds you that the AI is being dumb again. Fixing this will go a long way to improving our perception of the AI.

The Artificially Intelligent mod already takes care of the settling part. And I believe it's currently aiming to fix the commanders.

 

[Horatio Thorne]

This might be a controversial take, but I'm not fully convinced that the AI is actually an opponent in civ games. They are more of a game mechanic than an opponent or other player. Their main role is to put some military threat, fake competition for regions to expand peacefully, providing regions to expand militarily, and competition in some areas (wonders, legacy milestones). The player is not really playing against them as equals. Looking back, I can see this in all civ games, and while it might have been more balanced between game mechanic and opponent in the earliest games (1 & 2), the scaled tipped in favor of game mechanic with the huge bonuses etc. and an AI that didn't really play the same game as the player form 3 onwards. Since 6, this was even more true and now, with 7 and it's system of some leaders not even competing despite being on the map, it is clearer than ever: there are no opponents in single player civ games.

That doesn't really excuse the bad AI, but for me personally, it explains why a competitive AI (or a real opponent) isn't a priority. It's also how I see the AI paradox games, and playing these for many years, I reached this conclusion that the AI-played nations aren't meant as opponents to begin with.

But yes, some comparable games actually have AI-played entities that are full opponents, that play the same game by the same rules, and feel competitive, e.g., Ozymandias. But that game is of course much, much simpler.

Agreed. The AI Civs are more like NPCs. I begin every game fully expecting to “win”. The question is how I’m going to win, and if I can win more efficiently than my last game. I also set myself the task of completing all legacy paths in the first two ages and getting as close as possible in the final age. I set myself challenges and my success is not measured by technically winning the game or by beating the AI. I’m playing against myself.

As the OP points out, anyone can learn how to beat the AI even on deity, but it requires playing in boring repetitive ways with no room for creativity. If you enjoy that then great, but I don’t think that was ever the intent of the game designers really.

If you really want a competitive challenge then that’s what multiplayer is for, and single player is practice.

 

[Mr Jon of Cheam]

The Artificially Intelligent mod already takes care of the settling part. And I believe it's currently aiming to fix the commanders.

I'm glad to hear that! I'm a lowly console peasant unfortunately, but at least settling behaviour is listed in the FXS roadmap, so I have hope for that one at least.

 

[ProMeTheus112]

As the OP points out, anyone can learn how to beat the AI even on deity, but it requires playing in boring repetitive ways with no room for creativity. If you enjoy that then great, but I don’t think that was ever the intent of the game designers really.

Isn't this exactly what separates Civ7 from the previous titles? (at least Civ5 and Civ6 for sure in my mind). In Civ5 or 6, you are absolutely BOUND to a completely aggressive opening strategy and follow ups on high difficulties. However Civ7 is not at all like that, because the AI doesn't start with extra settlers and units (and also I think does play a lot better). Because of that, you can go for a ton of different openings on Deity, at least on a standard size map. My first Deity game I tried going for a low army start focusing on building many wonders, and it worked perfect! (this is was impossible in previous titles I think probably even Civ4). I did engage into war, but not to take towns from opponents, rather to for example protect a city state that I was friending that he was attacking, or to take advantage of neighbour being in a bad spot at war with someone else trying to just pick off units and pillage him a bit. Most of the game I played like this not really trying to take towns but still engaging or defending in wars and I was in a good spot in modern. I thought it was great and so much better than past titles, now I can see playing tons of different openings and strats on deity.

 

[Horatio Thorne]

Isn't this exactly what separates Civ7 from the previous titles? (at least Civ5 and Civ6 for sure in my mind). In Civ5 or 6, you are absolutely BOUND to a completely aggressive opening strategy and follow ups on high difficulties. However Civ7 is not at all like that, because the AI doesn't start with extra settlers and units (and also I think does play a lot better). Because of that, you can go for a ton of different openings on Deity, at least on a standard size map. My first Deity game I tried going for a low army start focusing on building many wonders, and it worked perfect! (this is was impossible in previous titles I think probably even Civ4). I did engage into war, but not to take towns from opponents, rather to for example protect a city state that I was friending that he was attacking, or to take advantage of neighbour being in a bad spot at war with someone else trying to just pick off units and pillage him a bit. Most of the game I played like this not really trying to take towns but still engaging or defending in wars and I was in a good spot in modern. I thought it was great and so much better than past titles, now I can see playing tons of different openings and strats on deity.

That’s good to hear. I haven’t played deity myself because that’s just not how I like to play, but from what others had said and from previous Civ titles I assumed you still had to go full military right from the start. Maybe I will try deity eventually in that case.

 

[stealth_nsk]

That’s good to hear. I haven’t played deity myself because that’s just not how I like to play, but from what others had said and from previous Civ titles I assumed you still had to go full military right from the start. Maybe I will try deity eventually in that case.

I managed to go through antiquity on deity without a single war, although I was crushed in exploration.

I believe one of the things I did wrong is not managing relations with leaders enough. I was too focused on getting city-states, but on Deity you need to focus on maintaining good relations, including saving some influence to reject denounces and sanctions.

 

[Noble Zarkon]

The AI knew how to play the game extremely well in Civ4.

It really didn't - watch Sullla's AI Survivor Series and you'll see that! However it was a much simpler game in many respects (No 1UPT, everything gets built in the City Centre) so it could take it's bonuses build a huge stack and wipe you out or race through the Tech Tree and launch a spacecraft. I still play a lot of Civ4 and love it but no the AI wasn't great.

 

[Leyrann]

including saving some influence to reject denounces and sanctions.

I don't really find that worth it to be honest.

If they really want to pick a fight so bad, I'll let them. I'll build a few more units to be ready and any declarations of war will be responded to by taking one or more settlements before accepting peace.