Make The AI Great at Conquering Again

[Trav'ling Canuck]

Have Devs said anything about war AI?

Not directly, that I know about. But Ed did say that he thought the AI in Civ 6 was pretty good at most aspects of combat and their focus for Civ 7 was on things the AI was particularly bad at, like movement. Military Commanders were motivated about alleviating a nuisance for the players (moving units one at a time past terrain choke points), but also by a desire to make the AI more effective at getting its units into proper position. So to the extent that the AI can use Military Commanders to get its units into better positions to attack (and defend?), that should improve the AI's ability to wage war in Civ 7, at least in theory.

Obligatory reminder that the dev team's view of what constitutes the AI being good at war likely differs from the average Civfanatic. What we should mostly hope for is that the dev team makes the combat AI sufficiently accessible to modders that they can tune it to be a challenge for everyone's personal preferred difficulty level.

 

[Prester John 2]

Have Devs said anything about war AI?

Ed Beach said they were quite content with the tactical AI in Civ6 so it seems not much has been changed there. What they did fix is the difficult movement of armies.

However, IMHO, the movement wasn't the only problem: army composition, the lack of late game units and inefficient sieges hampered the AI conducting a war massively. Often the Barbarians were more of a threat than other civs.

 

[Krikkit1]

I think the Crises can Really help in this.

Right now if an AI can eat another AI it means a snowball... either
eating another AI helps and you should all be doing it
OR
eating another AI is more than you can handle (corruption happiness, etc.) so good AIs won't do it too much

However, if conquering another AIs cities can give you good benefits in this age... but you are very likely to lose some of them in the Crisis... then its not as much of a snowball.
Especially if taking all a civs settlements doesn't kill them....it just means they operate a low level rebellion until the Crisis.. when they operate a Major rebellion to get a chunk of their land back.

 

[Disgustipated]

Speaking of Civ 6 war AI, on a hunch I reduced combat strength of crossbowman and archers in the game files, and the AI was able to conquer cities after this. I only played one game, however, so this isn't an accurate test. Because I've noticed how hard it is to conquer a city in Civ 6 that has walls and a crossbowman in it. And the AI certainly can't conquer your city with those things, but even against other AI's it struggles. I just feel like the combat values of ranged units is unrealistically high.

I would say just getting rid of city bombards alone would help tremendously in this regard. But as I mentioned above, reduce the effectiveness of ranged units.

 

[RedCourtJester]

Speaking of Civ 6 war AI, on a hunch I reduced combat strength of crossbowman and archers in the game files, and the AI was able to conquer cities after this. I only played one game, however, so this isn't an accurate test. Because I've noticed how hard it is to conquer a city in Civ 6 that has walls and a crossbowman in it. And the AI certainly can't conquer your city with those things, but even against other AI's it struggles. I just feel like the combat values of ranged units is unrealistically high.

I would say just getting rid of city bombards alone would help tremendously in this regard. But as I mentioned above, reduce the effectiveness of ranged units.

I suspect that's just reflective of an intentional design intention to err combat on the side of the player being less likely to die to invasions. It's setting a high floor for failure, the Civ equivalent of an MMO like GW2 or FFXIV removing extraneous/overly "choosy" abilities and modular stat boosts so that it is much, much harder for casual players to fail, get frustrated and quit. It's...player plot armor.

 

[Krikkit1]

I suspect that's just reflective of an intentional design intention to err combat on the side of the player being less likely to die to invasions. It's setting a high floor for failure, the Civ equivalent of an MMO like GW2 or FFXIV removing extraneous/overly "choosy" abilities and modular stat boosts so that it is much, much harder for casual players to fail, get frustrated and quit. It's...player plot armor.

Hopefully the Crisis can help with that…a player that is behind (just lost a major chunk of territory)..particularly a himan player should have the Crisis adapted so that they can actually gain during it (reclaiming some old territory as the conqueror deals with the plague/rebellion/barb invaders

 

[RedCourtJester]

Hopefully the Crisis can help with that…a player that is behind (just lost a major chunk of territory)..particularly a himan player should have the Crisis adapted so that they can actually gain during it (reclaiming some old territory as the conqueror deals with the plague/rebellion/barb invaders

Oh yes, I am hoping that the overall better parity and "rubberbanding" afforded by three eras will allow things to feel much more dynamic and less "well I executed the science/culture/econ/religion hacks, now just gotta play cookie clicker."

 

[Nigel_Tufnel2]

Commanders will definitely help get many units to one spot in formation, but if they have done nothing to improve how battles play out the AI will be a joke again. How many times in VI did you see half-damaged Melee units dance around a near-dead city and not take it out? Or catapults or Archers diving into bushes rather than shooting? V's AI could be merciless, esp as Shaka or Attila. Revisit that battle code, not whatever was changed for VI.

 

[Prester John 2]

Commanders will definitely help get many units to one spot in formation, but if they have done nothing to improve how battles play out the AI will be a joke again. How many times in VI did you see half-damaged Melee units dance around a near-dead city and not take it out? Or catapults or Archers diving into bushes rather than shooting? V's AI could be merciless, esp as Shaka or Attila. Revisit that battle code, not whatever was changed for VI.

Yes, and Vox Populi brought that to another level. It is really hard waging a war against a well prepared enemy with Vox populi. Why don't the developers look into their code?

 

[Lonecat Nekophrodite]

AI should be able to do naval warfare properly now.

 

[Foufa131]

Have Devs said anything about war AI?

I think that they only said that they have way more people working on the AI, which is good.

 

[Foufa131]

AI should be able to do naval warfare properly now.

They did talk about the naval warfare, so I'm also optimistic. Even tho I never really cared that much about naval conquest.

 

[Disgustipated]

It's too soon to say they will do it "properly", whatever that is. But there are reports that the AI is building a navy now, so that's good. Always hilarious in Civ 6 When Norway of all civs only has like 1 boat.

Of course that can be a double edged sword. If they are building a Navy and neglecting ground defenses, that would be easy for the human to exploit.

 

[Trav'ling Canuck]

I think that they only said that they have way more people working on the AI, which is good.

Technically I think they said they had double the number of people working on AI. Which some folks are interpreting as they now have 2 people working on AI, since they had 1 in Civ 6, while others view it as 1 person now working on AI ( or 2 people working part-time on AI), since the 1 person who did AI for Civ 6 was generally understood on this forum as being part-time on AI as they also had other dev team responsibilities.

 

[Trav'ling Canuck]

It's too soon to say they will do it "properly", whatever that is. But there are reports that the AI is building a navy now, so that's good. Always hilarious in Civ 6 When Norway of all civs only has like 1 boat.

Of course that can be a double edged sword. If they are building a Navy and neglecting ground defenses, that would be easy for the human to exploit.

They are building a navy. In a lake. That is 4 hexes big. And bordered by that same AI's territory all around.

You will not succeed in invading that AI via that lake. The AI has seen to that.

 

[Prester John 2]

I guess the tricky part is really what to build (out of all the options), where to station them and where to deploy. I guess it would be a good start to force the AI to have some ranged garrison troops in their cities and forts during a war. Secondly, keeping one to two ships close to their most important seaside cities. If they are invading another civ they have to at least partially switch to a wartime economy and keep producing units/reinforcements, not just an initial wave of units. Of course, this has to be weighed with a long time strategy AI that is not indefinitely locked into building units by keeping them in a war.
Recently I discovered the Civ6 mod Late game AI which at least makes the late game AI build some units. Downside: Now with troops at hand the AI civs are constantly at war with each other.

 

[Lazy sweeper]

The other big issue since civ V, is no more complete kills.
Archers can bombard, and not take damage. Archers should have a bombard action, like the Persians in civ VI, but after, they should enter melee, even if behind a wall or up a cliff, if the other unit
has enough mobility and skills, it should be able to counterattack and kill the Archer, or get killed.

Only fast units with better mobility should be able to retreat and survive to fight another day, or turn...

The direct consequence of the super strong units, which almost never die in a direct combat, are atrocious.
Every war or battle that could take one turn, and realistically, if a turn has a span of 250 years, it should; it now takes at least two or three turns, just to kill a unit.
Sieges that should last no more than a few turns, can take ages. This is IF there is an army on the ground.
Obviously if there is no Army and we are fighting against walled city with auto-bombard, the situation now reverse. Cities falls one after the other.
Good Archers will bombard a city to the ground, and not a single units will be lost in any conquering whatsoever.

Archer with bombard should get a rng ranging from 0% dmg to a maximum of 2% on other units. Navies on the other hand should keep a bombard advantage on land units.
Archers like the skirmishers should engage in melee fight every time, as every other unit, end fights should always end in a complete kill.


Units of all kinds should be able to be built a lot faster, in greater numbers, to give the impressions of a great army.
Few units that are super strong, almost never die, and are difficult to build, otherwise will suffice for AGES.

To resume combat we need Complete kills. Force AI to build hundreds of units. And of 100 units, sent to a battle, it may end in a win, but 99 units needs to DIE.
We need this Adrenaline that was taken away from us.

It is very difficult to understand, if you haven't played civ III and IV enough, you can't comprehend what it means to have to build
a 100 units Army, to just then lose all but three units, and with those three units left, you manage to finish off your worst nightmare of opponent civ,
and then SURVIVE and attack from your ALLY, that dont give a hole about friendship, diplo, or warmongering penalties ( Really?). To then kill
him off in one single Epic battle with your last standing Army of three Swordsman and a War Elephant.

Civ III and IV was like, ther's anything that money can't buy. Declare ware to your ally? 100 Gold and we'll betray our best friend.
Everything was fast. Wars were always very tactical, because Complete kills, just like Chess.... you can't afford to LOSE THE QUEEN or KING....

AI did understand the concept of GAME OVER, CHECKMATE given... and was alway Looking for the opportunity to launch a sneacky attack
DIRECTLY at your capital, to kill the King, burn the libraries, and checkmating you...

If the AI doenst understand the value of its pawn, it will never succeed. Units has to die quickly.

 

[Arcaian]

I'd really rather hitpoints aren't removed - so long as we're choosing not to produce buildings/wonders/districts/etc to build units, it feels very bad to lose a troop out of nowhere. If my unit is surrounded when I hit end turn? Yeah, that was my tactical mistake. If it's a 98% chance for me to defeat them and I randomly lose my unit entirely? No thanks, it's not fun - and I say this as someone who has been playing heavily since Civ 3, so your statement that going back to this would be "very difficult to understand" doesn't apply on that front. If we want instant death from a combat (especially when the argument is based off of the in-universe time it takes for a battle to complete), I think our troops should be mustered instantly on the same turn based off the population and infrastructure of a city.

 

[Voidwalkin]

If it's a 98% chance for me to defeat them and I randomly lose my unit entirely? No thanks, it's not fun

Agreed. I tried BG3 and hated missing on odds. I hated it in Xcom, too, to the extent that I actually rigged my party in such a way that it simply did not happen.

Adding chance into strategy games provides some maddening outcomes not to my taste.

I once had a 4 campaign in which I lost 3 heavily promoted swordsmen at 98, 96 and 92% odds, consecutively. I continued the game to conclusion, at one point simply throwing 4 units and won 3/4 battles at 8, 12 and 16, losing one at 64%. I just concluded that it has some sort of evening system, and rather than look it up, I just stopped playing. I intended to, never got around to it. Effectively ended my civ4 days.

 

[Arcaian]

Adding chance into strategy games provides some maddening outcomes not to my taste.

If the chance is impactful enough, I agree - I don't mind that my archer might do 15 to 23 damage on an attack, I don't feel like it's the end of the world if I get 15 on that roll. But if the downside is as impactful as losing something that I built instead of setting up my infrastructure, it's just too annoyingly random to me.