What AI combat bonus would be necessary to make Domination Victory hard?

[HF22]

+2/3 per level may make the game challenging but it also means no fights until you cover the gap with better units. Not a fan of the idea.

No need to cover the gap with better units. Basic use of ranged units etc would still cover it. You might just lose a unit now & again, which would still be an improvement.

 

[Acken]

Then you skew even farther the need to rely on range vs melee imbalance.

I fail to see what we would be getting out of it.

 

[HF22]

Then you skew even farther the need to rely on range vs melee imbalance.

I fail to see what we would be getting out of it.

Well, needing to use skill to equalize rather than dominate, would increase challenge.

You could go the other way to make the game easier for the AI, like nerfing ranged vs melee, but that ends up more like the religious combat everyone hates.

But given the things people want seem impossible commercially, like much better AI coding, imperfect but better options should be on the table.

 

[Gtdead]

Well, even if you give the AI a +100 combat bonus, as long as it keeps reshuffling, not claiming obvious kills and retreating without any apparent reason, it's not going to be any good. It will just make combat tedious.

I'm not sure exactly how it worked, but in Civ 5, AI was really good at killing my stuff. I had to exploit it's love of stealing workers to capture some cities.

 

[Acken]

You could go the other way to make the game easier for the AI, like nerfing ranged vs melee, but that ends up more like the religious combat everyone hates.

I sincerely disagree on that statement. Balancing melee and range and making sure combat is more leveled between humans and AI should be the first step.
Then you give bonuses.

Also people don't want a human like level AI.
I think asking for at least the level of Civ5 BNW is not too much to ask. Civ6 AI is terrible. And not on tactical combat alone. It's across the board stupid and diplomacy is even more a disaster than civ5 which is a remarkable feat.

 

[HF22]

I sincerely disagree on that statement. Balancing melee and range and making sure combat is more leveled between humans and AI should be the first step.
Then you give bonuses.

Oh, I agree with everything you say here. Also I like how the AI is vaguely ok at religious combat.

Just noting the paradox of the community demands.

 

[Browd]

The AI's ability with religious combat is an interesting illustration of the differences between many players' stated desires for an AI that will kick butt and what they actually appear to want. The AI's tactic of swarming a civ with as many Apostles as possible is a very effective strategy (although easily enough countered with inquisitors around your holy sites), but from the tenor of forum posts, it appears that some (many? most?) human players despise how effective the AI is at eliminating the human player's religion -- no fun, no fair, bad AI, turn off religious victory, it's broken -- the complaints go on. I suspect what most players want is a carefully tuned AI that poses a modest challenge throughout the game (while never holding a grudge when the human player stomps around the map backstabbing and conquering as he sees fit), and then gets out of the human player's way so they can win their hard-fought, well-earned victory. I've got to believe that designing an AI that's just good enough to always lose is a pretty challenging task -- building an AI that would always crush the human player might actually be easier.

 

[Heikki Mustonen]

For me it seems that AI has huge problems dealing with city walls. Specially if you build them after they have started the attack. Cities without walls they seem to surround and attack at some point but they have problems bringing support units to right places to attack city walls. If you have walls in the city AI just seems to run around the city and get slowly killed with ranged units and city attacks.

 

[SCBrain]

My solution:

Increase the upkeep per unit at higher levels dramatically. At the moment, it is possible to create an infinite-sized army consisting of slingers and warriors (and Sumerian war-carts if available) even at deity FOR FREE. If you add the "-1 Gold per unit" card, you can support an infinite army of warriors, slingers, spearmen, archers, galleys, chariots, and battering rams.

(A related observation here is that there's too much gold in the game, after the first trade routes come on line. I always have hundreds and then thousands of gold surplus per turn, and I run out of things to spend it on. Money should be SCARCE in game, as it has almost always been in history, so as to create drama and tension.)

My proposal: In the ancient era, at Prince or King, the first unit is free. Second unit costs 1 GPT. Third unit costs 2 GPT. And so on. A big army, if you can produce it, is going to be expensive, even for lower difficulty levels. If you have seven units moving around in the ancient era, it will cost you 21 gold per turn. You can look at your balance and know whether you can afford to enlarge your army. If you don't have +7 gold for the eighth unit, you'll be running a deficit, and when you hit zero, your units will be disbanded. In later eras, the costs can be 1/2/3/4/5... and 2/3/4/5... instead of 0/1/2/3/4...

Of course the numbers might need some adjustment to play well - maybe the first three units would be free and then the prices climb would more slowly than in my example above - but this shouldn't be too hard to optimize.

Then, you could address the problem of easy higher levels by increasing these costs accordingly for the top difficulties. Instead of 0/1/2/3/4 costs in the ancient age, it would be perhaps 1/2/3/4/5... for Emperor, perhaps 2/2/3/4/5 for Immortal, and 3/3/4/5/6.. for Deity. That would be a real challenge! If turns out to be too hard - bring the numbers down a bit; but no one would say it's too easy.

 

[HF22]

The AI's ability with religious combat is an interesting illustration of the differences between many players' stated desires for an AI that will kick butt and what they actually appear to want. The AI's tactic of swarming a civ with as many Apostles as possible is a very effective strategy (although easily enough countered with inquisitors around your holy sites), but from the tenor of forum posts, it appears that some (many? most?) human players despise how effective the AI is at eliminating the human player's religion -- no fun, no fair, bad AI, turn off religious victory, it's broken -- the complaints go on. I suspect what most players want is a carefully tuned AI that poses a modest challenge throughout the game (while never holding a grudge when the human player stomps around the map backstabbing and conquering as he sees fit), and then gets out of the human player's way so they can win their hard-fought, well-earned victory. I've got to believe that designing an AI that's just good enough to always lose is a pretty challenging task -- building an AI that would always crush the human player might actually be easier.

Spot on. As a developer, why would you spend $1 on fixing the combat AI, when the religious combat example suggests the playerbase would hate you for it anyway? We all need to decide what we want, before the developers can have any chance of delivering it.

 

[Matthew.]

Ehhhhhhhhh.....

I don't know. I get the point, but I think AI bonuses can't be ignored in that point. Bombarding a human player with apostles is one thing, bombarding a player with units gained through bonus cheats is another.

Let's look at RTS games for a moment. AI generally does okay, since what they lack in tactical and strategic calculating, they make up for in speed. Let's say Starcraft 2, since it is somewhat recent people are familiar with it, though any RTS game would work. The general player base probably wouldn't mind an aggressive AI which pushed your base with roaches early on, as long as it still used game rules. After all, a roach rush is a roach rush. However if you give that same AI cheat bonuses and it rushes you with 30 roaches by the time you get 1 stalker out, then it becomes not okay. Sure, it is challenging, but it is breaking the game rules and it is no longer fun trying to micro 1 stalker around to abuse an AI rushing you with 30 roaches.

This is the impression I, personally, get when reading the same point on religious combat. Are players upset because their empire is being bombarded with apostles? Or are they upset because AI's are using their free settlers, production bonuses, and early faith districts to break the game rules turing it into 1 stalker vs. 30 roaches?

Now yes, Civ is too complex to expect an AI to perform without ANY sort of bonuses. I'm pretty sure we all agree on that point at least. However I'm willing to bet that if the AI was touched up and polished to play the game closer to the game rules, then apostle rushing a human player probably wouldn't be met with such criticism. Perhaps I'm wrong, but the way I see it, the complaining has more to do with AI breaking game rules than the tactic of sending a wave of apostles.

 

[WillowBrook]

The AI's ability with religious combat is an interesting illustration of the differences between many players' stated desires for an AI that will kick butt and what they actually appear to want. The AI's tactic of swarming a civ with as many Apostles as possible is a very effective strategy (although easily enough countered with inquisitors around your holy sites), but from the tenor of forum posts, it appears that some (many? most?) human players despise how effective the AI is at eliminating the human player's religion -- no fun, no fair, bad AI, turn off religious victory, it's broken -- the complaints go on. I suspect what most players want is a carefully tuned AI that poses a modest challenge throughout the game (while never holding a grudge when the human player stomps around the map backstabbing and conquering as he sees fit), and then gets out of the human player's way so they can win their hard-fought, well-earned victory. I've got to believe that designing an AI that's just good enough to always lose is a pretty challenging task -- building an AI that would always crush the human player might actually be easier.

Isn't this what difficulty levels are for? If they really can program an AI to stomp the average human player militarily, that should be deity. The current deity AI should be at most king. I haven't played much with the religious side, but swarms of apostles seem entirely appropriate for deity; I haven't seen such swarms on prince.

IMO, most people should think deity is too hard. Smart marketing could help make it more acceptable. Of course, they have to prove they can actually make it truly difficult first...

 

[Browd]

So you are asking for 8 different versions of an AI, each precisely calibrated to put up a decent fight, but rarely win, at their designed difficulty level. Sounds easy enough.....

 

[WillowBrook]

So you are asking for 8 different versions of an AI, each precisely calibrated to put up a decent fight, but rarely win, at their designed difficulty level. Sounds easy enough.....

No, I'm saying that, IF they can build an AI that can always stomp the human (which you suggested would be easier than whatever they're trying to do now), that should be deity despite whining from a subset of players. I have my doubts as to whether that's possible. But if it is possible, then you reduce the AI bonuses and/or give them handicaps of some sort for the other levels, as has always been done.

(I guess I wasn't clear - The current AI and this hypothetical human-stomping AI wouldn't be used in the same game. I just meant that the current deity AI shouldn't be labeled anything like deity.)

As for the original question, I can't see how additional combat bonuses to the current AI would improve the game. 10-15 points would certainly make it more difficult, but you'd have to exploit the stupidity of the AI even more, so i don't think it would be more fun.

(FWIW, I'm still really enjoying the game after about 100 hrs and expect to enjoy a few hundred more before I require mods or a good improvement patch to keep enjoying it.)

 

[LDiCesare]

I don't think the critics regarding religious combat apply to regular combat.
First of all, the AI is not good at it. At all. It throws tons of apostles, suicides them with no support and ample time for you to heal from. It's bad at it.
Second, it's a pain because religious units prevent non-religious units from moving around. That's the sole reason why those stacks (carpets) are so annoying. Otherwise, it's pretty easy to deal with. DOW, kill a few of them, and you have reduced their religion influence to bits in a moment. Keep 3 aposltes rotating around a holy site and defeat all the incoming hordes. The religious AI is nowhere near competent.

 

[EpicWestern]

As a thought experiment, let's say you play a game under the following conditions:

--Deity difficulty
--Pangaea map
--Standard size
--You have a decent production start and access to horses and iron

See, I feel like with the bolded you're basically cheating already so its better to just not do that than to blame the developers. Just have stricter self-imposed rules for yourself and domination is significantly harder (but yeah, probably still too easy due to poor AI). My rules:

-- random civ
-- random map type
-- keep all starts
-- no stealing units
-- no super early attack
-- no initiating trades

 

[HF22]

I don't think the critics regarding religious combat apply. First of all, the AI is not good at it. The religious AI is nowhere near competent.

Well, then you get back to the 1upt issue. Because the AI religious combat IS far better than the AI military combat, as it gets all help of a simpler system.

So while the religious combat AI is only vaguely competent, its likely the strongest 1upt combat AI the developers can provide in a Civ context.

If its not good enough, then I doubt any Civ 1upt system will be good enough to provide challenge.

 

[Seven05]

See, I feel like with the bolded you're basically cheating already so its better to just not do that than to blame the developers. Just have stricter self-imposed rules for yourself and domination is significantly harder (but yeah, probably still too easy due to poor AI).

I'll add something similar: Just play on bigger maps.

Yeah, the AI sucks and will until they give us the DLL source code, then it will only suck for the people that don't use mods. But you can't setup the game to be ideal for one type of victory condition and then expect a challenge. It's kind of like playing a duel sized map against Kongo and complaining that a religious victory is easy.

 

[HF22]

I'll add something similar: Just play on bigger maps.

Do bigger maps, or continents, make it any harder though? Makes it take a bit longer, I'll grant, but I don't think its any harder.

Unless maybe when it means there is more space for barb spawns, because they do provide a challenge sufficient to at least not break immersion.

 

[Cymsdale]

The AI's ability with religious combat is an interesting illustration of the differences between many players' stated desires for an AI that will kick butt and what they actually appear to want. The AI's tactic of swarming a civ with as many Apostles as possible is a very effective strategy (although easily enough countered with inquisitors around your holy sites), but from the tenor of forum posts, it appears that some (many? most?) human players despise how effective the AI is at eliminating the human player's religion -- no fun, no fair, bad AI, turn off religious victory, it's broken -- the complaints go on. I suspect what most players want is a carefully tuned AI that poses a modest challenge throughout the game (while never holding a grudge when the human player stomps around the map backstabbing and conquering as he sees fit), and then gets out of the human player's way so they can win their hard-fought, well-earned victory. I've got to believe that designing an AI that's just good enough to always lose is a pretty challenging task -- building an AI that would always crush the human player might actually be easier.

If players are complaining about apostle spam because it is successfully converting their cities, they are being silly. If they are complaining about it because of the baffling decision to have religion and military units work on the same layer, that seems more legitimate to me.

Other than the units getting in the way, I have no problem with the AI spamming religious units. After all, I use the same strategy against them.

As far as military goes, I think the AI would be better off if it tried to act less "smart" and acted a bit more recklessly with its units. Especially if their military is much stronger. If the higher level AI used its production advantage to do overwhelming assault tactics, I suspect that would lead to quite a few player tears.