Civ 6 too easy?

[Tychonoir]

I just recently got Civ 6 on Steam.
Is it my imagination, or is this incarnation easier than the others?
Don't get me wrong. Overall, there is a lot to like in this version, but the challenge is lacking.

I'm playing on Deity, and the AI is as bad as ever.
I know, I know, Civ is notorious for questionable AI, to the point that it seems to be one of its endearing features, but the situation now is ridiculous.

Essentially, the game consists of getting 6-8 cities as quickly as possible. And then... that's it. You've won the game, as there is absolutely nothing the AI can do to stop you now if you turtle up and go culture or science. It becomes rather pointless to play after the classical period, as the game is a foregone conclusion.

I used to think I was just really good. Now I see the AI is helpless.

This might be different if the AI could manage an invasion, but the warfare AI is the worst I've ever seen. I'm not talking merely sub-optimal, it will often make the very worst possible tactical moves.
I think the AI doing random things would actually fare better.

But you all probably know this.

But here's the thing: I'm pretty sure competent hex warfare AI was created in the 80s-90s. How has this knowledge been lost?
Given that, are there any chances that tactical AI can be modded?

Yeah yeah, strong AI is hard and expensive. How about just somewhat competent AI? Is that too much to ask?
So maybe they don't have the resources to build an 80s era AI. Fine. They could at least create a AI modding framework and let the community create it. I guess It'd be a bit of a cop-out (But a marketing opportunity!) But at least it would be better.

Sorry about the rant.
Probably gonna go play some more now.

 

[teks]

1UPT is whats really taxing the AI. Though it is pretty hilarious how good the AI was in the original Master of Orion and the older civ games in comparison to today.

 

[joncnunn]

But here's the thing: I'm pretty sure competent hex warfare AI was created in the 80s-90s. How has this knowledge been lost?

Pretty much all the hex warfare games of that era featured soft 2 units per tile; soft in that a 3rd unit could enter the hex as long as it was down to 2 units per tile before the movement phase was over. AI Path finding is a much easier problem when units can be temporarily stacked; the AI could be programed to always have an escort for artillery type units.

 

[Tychonoir]

1UPT is whats really taxing the AI.

I could swear Civ V was better at it than Civ 6. I could be wrong, but that was my impression. Though it was still abysmal.

AI Path finding is a much easier problem when units can be temporarily stacked.

Makes one wonder why Civ doesn't do this, then. Surely this problem was noted in Civ 5?

But things as simple as:
- Move all melee first, aim ranged and bombard units to sit behind the melee.
- Healthy units move first.
- Units closest to the enemy move first.
- Various combinations of the above.
Yes, these have plenty of weaknesses and problems, but it would be a huge step up.

Even if the AI just built far more ranged/bombard it would fare a lot better. As far as I can tell, melee is largely unnecessary and inferior to ranged/bombard, except to take a city.
I feel like there's supposed to be a rock paper scissors element: Ranged > Melee > Cavalry > Ranged
But in practice, ranged trumps all. Don't even get me started on ranged with the two attacks per turn promotion.

 

[unpossible251]

Too easy, yeah. Hey, is there an option / mod for changing difficulties mid game?

I often play emperor, won immortal once, never deity. So im no pro. But im sure everyone is familiar with the feeling that winners and losers are being shaped up in medieval, and by industrial you're so far ahead (or behind) that really youre just refilling and emptying buckets.

So, why can't we bump up difficulty mid game? Might round things out a bit. Especially if it will be years before the AI becomes patched into decency.

 

[Tychonoir]

I often play emperor, won immortal once, never deity.

I'm not feeling much of a difference on emperor, immortal, and deity in Civ 6.
Mainly, just how quick you have to be with early game expansion.

In the higher difficulties, the AI get extra settlers and bonus production, but then like to throw them at you with minimal escort, if any at all. So there's a free city. This seems to mitigate the difficulty significantly.
Declaring an ancient era war to capture a non-state city seems to be a waste of time in my experience, while declaring war to capture a settler is totally worth it.

Which brings to mind a question: I took a city-state in the ancient era (no warmonger penalty). When I meet Gandhi later, he thinks I'm a warmonger. How does he know? (Was still worth it, though. No one else seemed to care.)

 

[LukeHro]

Pretty much all the hex warfare games of that era featured soft 2 units per tile; soft in that a 3rd unit could enter the hex as long as it was down to 2 units per tile before the movement phase was over. AI Path finding is a much easier problem when units can be temporarily stacked; the AI could be programed to always have an escort for artillery type units.

I may get you wrong or miss something but by your definition of "soft" isn't that what Civ6 does too (just that soft 1upt)? You may even jump even over an enemy as far as I could see. And that shouldn't be different from AI standpoint if it could or not complete it's turn in 2 steps.
I'm not here to argue about 1upt but I can't help but notice that the argument of that being the fault/cause is overused when talking about AI incompetence in Civ6. How is 1upt to blame that AI is not even trying? not using air units?
I can't imagine that AI it's like "I must take that city, i'll move this and this towards it, but hey, in 3 turns they'll both need that 1 tile so I'll abandon the whole plan and beg for pace even if I declared war 2 turns ago"
AI is not only about moving units, that's just a small part of it, is also about when to go to war, how many units to keep around when seeing the neighbors power (and it can cheat on that and I bet they have no problem doing that based on how difficulty levels are implemented in the game), etc and Civ6 fails at all of those.
The OP is right, the AI has no excuse to be that bad considering the 30 years history. Unless...and I'm trowing a wild one here...algorithms were patented by others. Yet still enough time to reinvent the wheel even better.

 

[teks]

1upt is overused because it is the main reason the AI went from a tough competitor in 1-4 to a total pushover in 5-6. Its a hard 1upt because you cant put 2 military units in a stack. Honestly the AI isnt the only entity struggling with this.

Ever try to move multiple units down a road?

This is a basic move. Basic logistics essential to positioning our units....and its practically impossible. Pathfinding will set them off into the wilderness. Bottlenecks leave units with wasted moves. Units make it to the destination in terrible formation. We have to babysit our units, and even then, even taking it upon ourselves to force the units to use a freaking road, we end up with wasted moves.

Maybe the AI has a good idea, and it tries to get its units in position, but its limited to pathfinder. Its units are scattered all over. Every strategic position it comes up with comes up with errors as something else is inevitably blocking the path. Half its units took an entirely different route and are already dead. The rest just helplessly wait in enemy territory for reinforcements that will never get in position in time.

Its prettt simple here. If WE can't get even use roads to quickly move units, the AI hasn't a prayer. If we can't get basic melee units to work right, because its impossible to position them properly in 2 moves, the AI hasn't a prayer. If we struggle to push units through a bottleneck...well you get the point.

Until they change 1upt, this is what we will have to put up with. And the devs have to put all their energy into the futile practice of teaching the AI to use this system, energy that could have went into other aspects of the game, like air, districts, and win conditions.

Eventually, just as I did in civ 5, I will return to civ 4, the last real civ game, before it all turned into an arcade game. The level of depth, the immersion, the strategy and tactics, it can't be beat. It just sucks because there are so many great ideas in the newer versions I would love to see, but they keep covering it in cartoon colors and nonesensical mechanics that don't work in a civ game.

 

[Cymsdale]

I was never really clear what problem 1UPT was trying to fix. Doomstacks? So what? Not worth neutering the entire AI to fix IMO.

Just make it 3UPT. No fancy rules about which units are the exception, just make every unit a unit and you can have any 3 in any tile. This even allows you to have interesting abilities on units like "will attack weakest unit in tile" (if say, the default was to always have the strongest one defend) or "will damage all defending units in tile simultaneously" (hmm, didn't Civ 4 already have something like this?)

 

[LukeHro]

Relay guys, this is not about 1UPT, 3UPT or nUPT. Despite you loving Civ4 and MUPT, that's your right and I don't want to express my preference regarding this, you need to see that AI has bigger issues then moving the troops on map. That should be basic problem for any AI regardless the frustration it may cause for a human player.
If they really tryed to make the MUPT AI just work with 1UPT that would be really the most idiotic thing to do but I give them a bit more credit then that. I just don't bite that. They tried and failed but that's a different thing. Is it harder to make an AI work with 1UPT then MUPT...maybe, most likely, but that's no excuse once the decision (bad or good) to make that change was taken. There's no room for "I told you you'll fail the AI if you make that change".
Maybe each of us had a different experience with our game, but I believe we all saw units moving mindless back and forth. But don't you do that too especially with no auto wake feature? Why can't an AI do that too? But once at war, one declared by it, there's no excuse to mindless movements.

Maybe the AI has a good idea, and it tries to get its units in position, but its limited to pathfinder.

Why would you link so loose the two components in your implementation? If they did that it was just stupid and could be fixed by allowing more interaction between the components (AI can interrogate pathfinder where each unit will be at EOT). They failed that maybe, but not because it can't be done with 1UPT.

 

[Leathaface]

Essentially, the game consists of getting 6-8 cities as quickly as possible. And then... that's it. You've won the game, as there is absolutely nothing the AI can do to stop you now if you turtle up and go culture or science. It becomes rather pointless to play after the classical period, as the game is a foregone conclusion.

It should be difficult for you to get an 8 city empire IMO. I agree that if you get 8 cities and don't get invaded you should win, but you should have to fight to get 8 cities.

Plus you should have to invest a good amount of your time into Military incase you get invaded, so you can't easily put all your time into Science/Culture making victory easier. However the AI is bad at invading and capturing cities with walls.

 

[isau]

Have you tried mods? I don't want to plug my mods too much but if the base game is boring you, maybe try my 'Combined Tweaks' mod in Hardcore mode, paired with the AI+ mod. It's still not necessarily difficult, but on Emperor the AI has given me a serious run for my money.

FWIW there are two or three key things I think are "off" about the balance in Vanilla:

• You get free Culture or Science from Population, which makes it very easy to ignore Campuses and Theaters, since you are under no pressure to build them. This in turn means the optimal district in most cases is the Commercial hub.
• Trade routes become overpowered as the game progresses, with too many potential bonuses, leading again to the Commercial district being the most important, since it is ultimately the source of a huge portion of your yields.
• City walls are too strong. (I don't necessarily recommend this fix for the core game, but in the mod I play, Walls require an Encampment with a Barracks or Stable. I like that change because it creates empires with soft spots, and again emphasizes the importance of districts/makes it harder to just spam Commercial districts.)

With Civ V it was kind of the same thing. The core game was just ok IMO. I did really enjoy the mods tho, particularly Vox Populi. Civ 6 has some excellent mechanics and some fairly poor balance. But I do feel this is a great game once tinkered with, just like V was.

 

[teks]

Relay guys, this is not about 1UPT, 3UPT or nUPT. Despite you loving Civ4 and MUPT, that's your right and I don't want to express my preference regarding this, you need to see that AI has bigger issues then moving the troops on map. That should be basic problem for any AI regardless the frustration it may cause for a human player.
If they really tryed to make the MUPT AI just work with 1UPT that would be really the most idiotic thing to do but I give them a bit more credit then that. I just don't bite that. They tried and failed but that's a different thing. Is it harder to make an AI work with 1UPT then MUPT...maybe, most likely, but that's no excuse once the decision (bad or good) to make that change was taken. There's no room for "I told you you'll fail the AI if you make that change".
Maybe each of us had a different experience with our game, but I believe we all saw units moving mindless back and forth. But don't you do that too especially with no auto wake feature? Why can't an AI do that too? But once at war, one declared by it, there's no excuse to mindless movements.

Why would you link so loose the two components in your implementation? If they did that it was just stupid and could be fixed by allowing more interaction between the components (AI can interrogate pathfinder where each unit will be at EOT). They failed that maybe, but not because it can't be done with 1UPT.

We bring up civ 4 because the AI was good. Its not a preference thing. Its a fact that the AI in civ 4 could wage war effectively. You can choose to believe that 1upt had nothing to do with the drastic loss in competitive AI between 4 and 5, but I have strong doubts that the company lost their notes, or forgot how to program between iterations of the game. Personally I'm intererested to hear exactly why the AI got so much worse between 4 and 5 if this has nothing to do with 1upt.

 

[Tychonoir]

It should be difficult for you to get an 8 city empire IMO.

I'm not saying it's possible to get every game, and it depends on the game settings. With a small map and the max number of empires (10), yes, it can be a bit of a mad dash. I just tried a normal map with the normal 8 empires, and I have 10 cities with a space reserved for 11, and a settler headed to a probable 12. And I only had to steal 1 settler.

This is partly only possible because you don't really need to escort your settlers much, and you can defend your early overextended empire with relatively few archers because, again, the AI is incompetent at invasions. Furthermore, if you have a moderate sized military the AI is reluctant to declare war in the first place, and doesn't seem to factor in that your scattered army might be 10 turns away.

I usually get 4 slingers pretty early, and spend the 120g to upgrade them to archers because slingers are quicker to produce. These 4 can defend your empire for a quite awhile. Typically, I'll place 1 in a city close to my rivals. If they start to be angry, I'll move a second there. 2 Archers can defend a city for a long time, if not indefinitely. Certainly long enough to bring reinforcements. If Egypt starts to make noise about how weak I am, I'll build another 1 or 2 when I can.

Oddly enough, the AI seems to position scouts as if they are looking for weakness, but then doesn't take advantage.

Could be buggy, seeing as diplomatic relations sometimes still seem a bit random.

For example, Spain and India both declare war. That's strange, Gandhi doesn't hate me at all. Spain must have suckered him into a joint war. I kill a scout and a apostle from Spain, and that's the only troops I ever see. A few turns later, and Spain wants peace. Gandhi absolutely refuses to make peace no matter what I offer. So I check our relations, they are only at -2. I try again every few turns, and nothing. Still don't see his invasion force. Finally, I decide to march 3 crossbows towards him, and I get to kill a few units. Still can't make peace, but now relations are down another 3 or 4 because I'm moving troops near his border. Kill a couple more units. Now he wants peace and is offering all kinds of gold and resources for it. Gee guys, thanks for the war.

I don't think there's any way to characterize this kind of behavior as anything other than a bug.

Have you tried mods?

Not yet, but it looks like it's going to be required.

 

[joncnunn]

I may get you wrong or miss something but by your definition of "soft" isn't that what Civ6 does too (just that soft 1upt)? You may even jump even over an enemy as far as I could see. And that shouldn't be different from AI standpoint if it could or not complete it's turn in 2 steps.

Not quite: In these games you could use the numeric keypad to move a unit into a tile which already had two units without it auto-moving an existing unit out. (Civ V & VI both auto swap the units) Some of them may have required you immediately remove one of those three units now stacked before anything else, but you had the choice on which one.
So applying Civ V/VI rule to the AI, if where it wants a given unit to go is beyond one turn's movement and occupied by another unit that has movement points, it will cause that other unit to move backwards one tile.

 

[greygamer]

I wonder if the combat AI might be better if ranged units had 1 range and maybe lower attack values. Siege units could keep 2 range.
As regards the other issues with the AI, it has difficulty reading the map which means it needs more of a meta approach, unfortunately players would quickly learn how to exploit this as well.

An AI that adapts would be the ideal but turn lengths would be drastically increased as a result.

 

[Gehennas]

I am not sure that "easier" is correct word here. If you look at the minor decisions you have to make, they are certainly more difficult:

1. Your production is greatly limited, so you just can't build everything you want unless you are backwards in both science and culture
2. The weird cost scaling makes everying even harder, because you need to build both settlers and districts as early as you can
3. City placement is now more important because of the districts and wonders.
4. Eurekas/inspirations require you to plan ahead a lot. Also, they require a lot of micromanagement
5. Goverment cards replacement need some forward thinking too.
6. Barbarians are really a problem
7. Much slower roads require more effort in defending your borders

But yes, you are correct that there is no challenge in the game, because the AI is not just terrible. It seems that we have a bunch of simple scripts instead of a complex AI.

AI player does not have any goal when declaring war. It just does this from time to time if relations are low enough.
AI is not able to evaluate your threat level. It just gives you everything aside from great works and cities if it looses war by some kind of K/D ratio.
AI does not seem to compete for any victory type. Instead, it just builds more campuses/holy sites/theatre districts but nothing more.
Coincidently, AI is not able to evaluate your overall threat level as a competitor for a certain victory type and react accordingly.
Instead we have an "agenda" which is obviously a lazy solution and emulates some sort of competition for... fleet size, for example?

Combat AI is a completely different story: it is certainly not simple, because even straightforward barbarian-like "attack the most vulnerable enemy unit nearby or pillage otherwise" would be much better than what we see at this moment.

About 1upt: it would be really hard to program AI that can compete with human (like in every other game). But AI that is able to outflank less protected units, hold formation, focus wounded targets and at least shoot at enemy units is quite doable.

Unfortunately, looking at the huge bonuses AI gets on higher difficulties, I am sure that devs knew that AI is hilariously bad and just didn't care about it at the moment of release. Thus I doubt that it will become better in the future.

 

[LukeHro]

Not quite: In these games you could use the numeric keypad to move a unit into a tile which already had two units without it auto-moving an existing unit out. (Civ V & VI both auto swap the units) Some of them may have required you immediately remove one of those three units now stacked before anything else, but you had the choice on which one.
So applying Civ V/VI rule to the AI, if where it wants a given unit to go is beyond one turn's movement and occupied by another unit that has movement points, it will cause that other unit to move backwards one tile.

That's (yet I didn't notice AI actually getting a unit displaced like that) would just be a bug in the AI, a corner-case not dealt with if you want. I'm not saying the AI is ok with handling 1upt just that is not all to blame as I tend to see on this forum.

We bring up civ 4 because the AI was good. Its not a preference thing. Its a fact that the AI in civ 4 could wage war effectively. You can choose to believe that 1upt had nothing to do with the drastic loss in competitive AI between 4 and 5, but I have strong doubts that the company lost their notes, or forgot how to program between iterations of the game. Personally I'm intererested to hear exactly why the AI got so much worse between 4 and 5 if this has nothing to do with 1upt.

Imo it has as much to do with 1upt as with every other features introduced by civ5 and civ6. I repeat: I'm not saying the AI is ok with handling 1upt just that is not all to blame as I tend to see on this forum. Moving units with 1upt or not is just a small part of what an AI is responsible of and imo maybe one of the easiest part to implement. In civ6 even if it did that well AI players are already behind for it to matter. Let's take a few example from my experience with the game and other complaints I noticed on this forum:
-"AI won't upgrade units": with what? It doesn't ever seem to have money for it anyway as far as I could see in trading screen and it doesn't know how to make them: it doesn't clear villages only if it steps on them by mistake, it won't clear barbarian camps if there's no unit on home for it to attack (I seen that a lot times in my game...barbs just spawning on the same old camps giving AI a hard time), can't maintain trade routes with all those mentioned barbarians (once I appeared and cleaned up the second continent AI started to spawn trade routes), I just harvested their settlers from barbarians once I moved to their continent.
- AI can't handle city state and especially their quests to get envoys unless it completes them by chance maybe (or does it gets those actually?). AI actually attacked me and was about to win in my first game because my lack of units but I reloaded, send a trade route to a close CS which had that quest, got it vassal and payed for the units and with what my few units I obliterated it's whole army...it gave me it's only city beside it's capital for peace. A few turns later I worked on just accumulating some capital, got 2 CS near it's capital as vassals and with just 2 units of mine and the CS troops I eradicated Greece.
- AI can't handle eureka mini quests, and those bonuses are too much for player: how else would they be half way behind you by the time you get planes ~1900s?
- AI can't handle their own personalities: imo they are too scripted to follow those totally ignoring context
- AI can't handle districts: from what I seen on my game it may be okish for an individual city but definitely not for the sum of the city it had.
- AI can't handle GP and wonders in the actual context, again only strictly in line with their personalities and few ones that ridiculously look like scripted like going for Great Libray.
- AI spawns religious units: while a huge annoyance for the player it's focus on those harms it's other directions, plus the whole religious thing is poorly balanced compared to other game features.
All those are part of what I would call an AI and by failing at them the part of moving of units with or without 1upt is just the end result. All those are changes compared to civ4 which AI can't handle yet every one seems to blame the one change named "1upt". AI won't stand a chance at the wargame tactic part unless those are solved regardless it's 1upt or not. I rest my case.
PS: is there a way (sort of console command maybe) to just give AI a ton of money, and a dozen of envoys midgame just to see how it will handle a war with those at it's disposal?

 

[Gehennas]

- AI can't handle their own personalities: imo they are too scripted to follow those totally ignoring context

My favourite AI is Qin Shi Huang: he hates everybody because others build wonders and Qin just doesn't. Don't know why but this happens every game.

 

[WileyWilson]

Not yet, but it looks like it's going to be required.

I've been playing since release and have a modded version which is more challenging on Immortal / Deity. It's not as good as modded Civ 5 (early days yet, and no mod kit available).

Here are my "essential" mods:

• AI Siege Help - See the discussion on this mod as you can remove AI "full movement" penalty from all siege weapons
• AI+V9 = changes vanilla AI behaviour
• Better Trade Screen - Improved trade screen
• More Lenses - Essential for spotting barb camps and unimproved tiles
• Research Reminder - Triggers Civic/Science pop-up when you hit 50% (40% for china)
• Smoother Difficulty - Less starting AI bonus on higher levels, but increased bonuses throughout rest of game
• Unit Report Screen - Significantly improved "Reports" UI

I've used all of these mods together for as long as they've been out, without any issues.

Good luck. Let us know how you get on.