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

[CaiusDrewart]

Right now, the AI gets a +4 combat bonus against the human on Deity. Despite this, Domination Victory is currently quite easy, even on the highest levels--at least in comparison to past Civ games. The AI is tactically hopeless, and weak city defenses means conquest has never been so easy or fast. The +4 combat bonus is just not enough to compensate for that.

So my question is: if the AI were to remain as inept as it currently is, what combat bonus would it take for Domination Victory to become extremely difficult?

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

How high would the Deity AI's combat bonus have to be for you to consider Domination Victory truly challenging? At what point would you fail the majority of your attempts? Would +10 strength to all AI units be enough? +15?

If you want, you can give me two answers, one for if you're playing as Scythia/Sumeria and one for if you're not.

 

[Bleser]

Agreed. I'm about to start a new game and am struggling on how to define it so that I don't end up quitting by the Industrial Era because I realize I'm already in the lead and none of the AI pose any kind of threat if I get aggressive beyond annoying me with their repetitive denouncements.

It seems the AI knows very early on in the game how to assault the human player - but after the first one or two attacks it just never happens again. If the AI could amass a non-obsolete-unit army as you move through the eras it would be a challenge at any time. This seems to be the biggest threat to this game becoming boring if the AI can't get it together. I should feel pressured to defend myself (last game was an immortal difficulty and I won with ease on my first try). I know the Civ V AI took time to tweak to learn how to effectively use ranged units and gather the right kind of assault force; hopefully Civ VI AI will learn the same tricks.

So, to answer your question, I wouldn't add strength but fix the AI's ability to upgrade its military, amass a large army and actually attack beyond the ancient era.

 

[LDiCesare]

You can't lose a city to the AI mid/late game, no matter how strong it is. You could check how much bonus/strength is required of a unit to attack a city. This could be a good value.
Then, it depends on nukes. I don't know the effects, but you can probably use only nukes in order to win, in which case a bonus wouldn't solve anything unless the ai can take cities.

 

[Takfloyd]

No combat bonus is needed. In fact, the current one should be removed. AIs never needed combat bonuses in the past to be dangerous.

They just need to hire a halfway decent AI programmer.

 

[LDiCesare]

Unit upgrades wouldn't make the ai any better. You can mod away the strategic resources requirements for units to begin with and see what happens.
But the ai, even with up to date units, never attacks cities and that's the killer;

 

[Unconquered Sun]

AI combat bonuses are an ugly admission 1UPT has failed.

Short answer: +10...maybe.

Longer answers:

- remove the bonus altogether and let Deity AI insta-upgrade anywhere on the map without cost or strategic resource need. Crude, but hey.

- realize 1UPT Firaxis-made AI will never work with the complexity of siege, support, how city defense penalizes attacks, etc and rebalance the entire army roster. Make cities stronger, but let Melee units attack them better. Remove melee units bonus against anti-cav, making anti-cav superior to melee on the field. In this way Deity AI can be programmed to swarm and take human cities with frontline troops (melee), which is much easier than coding siege AI logistics, placement, and defense.

While at it, probably rethink current pacing of unit technologies and strategic resources. The game might benefit from removing strategic resources as requirement, but have them as significant +combat strength boosters.

 

[FurballRocker]

People like to use combat to add to the spurious claim that 1UpT is a failure but generally the way the AI uses units and even forms attack forces in a 1UpT environment is ok.e.g. they usually put melee up front and ranged behind. They group them together when creating an attack force etc. It is the actual end combat that it falls down, much of which could be fixed with a few simple changes to priorities which would make them much more difficult in combat and thus domination victory..

1) AI has lots of gold, especially on higher levels so should prioritise upgrading units with their gold or probably even better have a much higher priority than it does now with it being top priority for aggressive leaders such as Monty or Scythia to distinguish from aggressive and none aggressive leaders.
2) AI should attack melee units in range of their cities as priority instead of ranged.
3) AI should attack with melee units garrisoned in a city (where the garrison unit wouldn't die or the enemy unit wouldn't die and thus the garrison unit be moved out of the city).
4) AI should be "less afraid" to attack cities...Seen many massed attacks where the 10-15 units just dance around my city getting picked off by my small number of units and never attacking my city even though it is surrounded. On higher difficulties in particular getting the AI to human wave against your city walls would be an improvement as they can quite easily replace the lost forces with bonuses they have to production and gold.
5) AI should be more prone to pillaging.
A recent example for (4) and (5), AI surprise warred me and at one point had my city completely surrounded with every city tile having an enemy unit on it with a force of 2 catapults, 4 crossbows and the rest being horsemen with actually reasonable positioning so the catapults and some of the crossbows were in range multiple times yet the AI never attacked my city (it didn't even bombard my city with the ranged units) and only pillaged one district out of 3 districts, 5 mines and 4 farms.
That example isn't a failure of 1UpT as the units were well positioned and it was almost a perfect city assault but for some reason the units never actually attacked my city. The same problem would have occurred if they were stacked because in effect the command code to actually attack the city never occurred for whatever reason and if that bit of code was fixed it would have been a perfect attack.

 

[bbbt]

On diety level? Give the AI cities free walls on founding might make them harder to steam roll at the start.

I also wonder if removing fog of war for the AI would help it a good bit.

 

[FurballRocker]

On diety level? Give the AI cities free walls on founding might make them harder to steam roll at the start.

I also wonder if removing fog of war for the AI would help it a good bit.

A much better solution to the early steam roll would be to allow the AI to do the same rather than take away the human ability to do so.i.e. by increasing their combat abilities which leads back to my point (4) above particularly.
You see lots of early wars but often very few cities taken or other civs wiped out at what is the optimum time to do so.
The problem at the moment is the standard human start if to take over their nearest neighbour which expands your empire greatly for comparatively very little cost compared to what it would involve if you hadn't taken AI cities and also provides you with extra room to expand peacefully afterwards but the AI is poor at doing this even to other AI's so it is very easy for the player to expand very early on but harder for the AI to do so.

As i said earlier all it really needs is some small tweaks to the AI to make them a lot more effective at combat such as targeting the right targets rather than providing them with more bonuses or literal cheats.

 

[Cymsdale]

If the AI was halfway decent at moving around its units the production bonus would be enough. Either the AI moves it's armies around randomly, or the logic is trying to be "too smart". It really is frustrating how bad it seems.

 

[Kruos]

Remove 1UPT + Re-hire the AI guy of Civ4. Problem solved.

 

[CoconutTank]

What I'd like to see is a very very minor generic strength bonus for AI units (like only +1 or +2 strength) or none at all, and for damage modifiers involving typing to be increased or doubled, either all of them globally, or to only adjust positive modifiers for AI and adjust negative modifiers for human players. So then your artillery really suck against units, and your cavalry really suck against spearmen, whereas the enemy's warriors will clean up your spearmen that are out of position, and hopefully that gives the AI some more direction as well on how to maneuver.

 

[georgjorge]

I'm posting here because maybe one of you can tell me what is going on with the AI.

I'm playing Deity, Pangea, going for Domination. After conquering three AIs, I end up at the doorstep of Brazil. Now Brazil is advanced, more advanced than myself, and has seven cities. This is about 1200AD. I attack with my army, and within ten turns I encounter exactly three enemy units (nevermind that those are two Crossbows and a Chariot, the AI not upgrading can be solved through mods). So I take three cities without any resistance, and make peace for the rest of his cities except the capital. Then I attack his neighbour Spain, who has six cities, and encounter a single Field Cannon and nothing else in ten turns of combat.

How is it ever possible that a Deity AI, with all its bonuses, has actually less units than it has cities in 1200AD?

This is post-patch, with no mods running at all. I'm genuinely curious, because if this is going to happen more often - AIs being as easy to conquer as on Settler level - I don't know what I would do to make Domination victory ever a challenge.

 

[trojan13]

I'm posting here because maybe one of you can tell me what is going on with the AI.

I'm playing Deity, Pangea, going for Domination. After conquering three AIs, I end up at the doorstep of Brazil. Now Brazil is advanced, more advanced than myself, and has seven cities. This is about 1200AD. I attack with my army, and within ten turns I encounter exactly three enemy units (nevermind that those are two Crossbows and a Chariot, the AI not upgrading can be solved through mods). So I take three cities without any resistance, and make peace for the rest of his cities except the capital. Then I attack his neighbour Spain, who has six cities, and encounter a single Field Cannon and nothing else in ten turns of combat.

How is it ever possible that a Deity AI, with all its bonuses, has actually less units than it has cities in 1200AD?

This is post-patch, with no mods running at all. I'm genuinely curious, because if this is going to happen more often - AIs being as easy to conquer as on Settler level - I don't know what I would do to make Domination victory ever a challenge.

I suspected that AI also got profit from overflow exploit (they popped a unit per turn) but AI lost exploit jelly post patch.

 

[isau]

They need to drop the "conquer all capitals" aspect and change it to "own a certain percent of tiles" like it used to be. Since we can outright buy tiles now I'd put this around 70%.

Even Civ 4's AI would likely have failed miserably at the "conquer the capitals" criteria. The AI just can't handle it. Civ 4 AIs did win Domination Victories, and it was because if you just let them run wild they could out-race you to grabbing tiles. That is much easier to implement than Civ 5/6's rules.

 

[georgjorge]

@isau: I agree in general, but when an AI has way too few units in the late game and doesn't even upgrade those the exact victory condition doesn't matter because it's not going to get there anyway.

 

[elitetroops]

I'm posting here because maybe one of you can tell me what is going on with the AI.

I'm playing Deity, Pangea, going for Domination. After conquering three AIs, I end up at the doorstep of Brazil. Now Brazil is advanced, more advanced than myself, and has seven cities. This is about 1200AD. I attack with my army, and within ten turns I encounter exactly three enemy units (nevermind that those are two Crossbows and a Chariot, the AI not upgrading can be solved through mods). So I take three cities without any resistance, and make peace for the rest of his cities except the capital. Then I attack his neighbour Spain, who has six cities, and encounter a single Field Cannon and nothing else in ten turns of combat.

How is it ever possible that a Deity AI, with all its bonuses, has actually less units than it has cities in 1200AD?

This is post-patch, with no mods running at all. I'm genuinely curious, because if this is going to happen more often - AIs being as easy to conquer as on Settler level - I don't know what I would do to make Domination victory ever a challenge.

Pre-patch everyone in the game got the God of the Forge bonus, which for some reason was way higher than +25%. This caused the massive carpets of ancient/classical units. I've noticed also that with the patch they don't have nearly as many units. If they fight a war somewhere, they might send their entire army that way, leaving no defenders at all. Early on if you get a notification that someone has declared war on a city state, it's pretty much an invitation for you to walk in and take all their cities without facing a single unit.

I guess the devs adjusted military production with the God of the Forge bug before release and didn't readjust it when the bug was fixed.

 

[Rathelm]

If the AI was halfway decent at moving around its units the production bonus would be enough. Either the AI moves it's armies around randomly, or the logic is trying to be "too smart". It really is frustrating how bad it seems.

I think that's the issue. The AI is trying to be too smart and my guess is that it's cost benefit analysis is going a bit haywire. Especially when you consider the speed at which you can crank out units with the proper policies.

 

[MyOtherName]

"Fixing" the challenge by giving huge combat bonuses to an incompetent AI will push gameplay in a very unpleasant direction.

 

[CoconutTank]

"Fixing" the challenge by giving huge combat bonuses to an incompetent AI will push gameplay in a very unpleasant direction.

I agree, I don't want a blanket strength bonus for all of the AI's units. But I do think making the AI's units more punishing when they correctly apply their bonus damage modifiers would help. If, say, the enemy AI's melee vs. anti-cav bonus went up from +4 to +8 (as an example, not sure what the exact number are), then human players would need to maneuver their units better or be willing to take on bigger sacrifices.