AI Strategy Descriptions

Blake

The AI Guy
Joined
Feb 18, 2006
Messages
342
Location
Apolyton
For the curious, these are what may show up under cheat mode when you alt-mouseover an AI's territory.

Dagger:
The Dagger strategy involves building a lot more "City Attack" units, along with some "Counter" units, it does this at the expense of infrastructure and settling (but will do some settling and also get infrastructure which helps unit pumpage). Dagger also involves pursuing strong units via research.
AI's decide whether or not to dagger based on their personality and the availability of suitable UU's and combat units, an AI can decide to dagger from the start of the game (in which case it tends to chase ie Bronze Working), or it can find itself in a good position and decide to dagger from that point (ie Augustus might start off not daggering, but then get Iron and thus Praets and decide he wants to after all).
Also if an AI is attacked it may decide to Dagger as a means to counter-attack - this is mainly dependent on having real units.
Dagger wont activate if there's nothing to kill.
The name Dagger is basically a misnomer now.

Get Better Units:
The GBU strategy is active until the AI gets two units better than warrior/archer/longbow. Chariot+Horse Archer would do the trick, Axeman + Spearman would too, Elephant + Catapult. In the worst case it'd take Catapults+Trebs, or Catapults+Muskets to get out of GBU, what GBU does is reduce investment in units and prevents the dagger strategy from fully kicking in. GBU suppresses AI military buildups and war declarations, it figures that a military investment will damn it either way and it's best off just praying to get better units without being attacked. However being at war will largely override GBU and it'll just spam out whatever crap it has available.
Note: In the next version the special nature of UU's like Jaguar and Hwacha will be recognized, with those units counting as 2 "better units".

Crush:
The Crush strategy involves recruiting nearly every unit and sending them all in with the intention of causing some serious harm, it's used more often on Aggressive AI setting and AI's which are daggering are more likely to crush (GBU naturally suppresses Crush). The Crush strategy is triggered at the start of war planning and expires soon after, allowing the AI to then replenish garrisons as a guard against backstabs (As a newer strategy it needs some refinement).
Did an attack just really, really, really hurt? It was a crush.

Peace:
The Peace strategy currently does nothing. It involves a "personality profile" of the neighbors to decide whether or not the AI is safe from attack. While the Peace strategy shows as active, the AI thinks it's reasonably safe from being attacked by those it is at peace with - this is not to be confused with the AI's own feelings on attacking - Monty could show as Peace, despite planning to backstab his neighbor Gandhi... he just thinks Gandhi wont attack him. Once activated it will reduce unit spending levels, unless the AI is daggering.
Again, peace isn't working at the moment - it does nothing even if it shows up under chipotle.

Missionary:
The Missionary strategy is not active from the start of the game - instead it can only be activated by having a holy city and the religions of the neighbors also plays a role in deciding whether or not to missionary. Missionary overrides Dagger and causes the AI to try and secure it's position through diplomacy - using the logic that if it converts it's neighbors it'll be safe(r) from attack and on the right side of dogpiles.
Theocracy as a favorite civic also plays a role here, increasing the chance of Missionary.
Missionary AI's can be fairly scary since they usually have a huge shrine income and a bunch of buddies - but bad allies can also drag them down.

Production:
Involves a slightly increased weight on hammers and slightly decreased on commerce, is based on personality.

Culture 1:
The AI decides to pursue a cultural victory from the very start of the game (but only if the victory is enabled!). Stage 1 involves pursuing religions and wonders. Personality and traits play a large role, Theocracy as a favorite civic reduces the chance. Being a vassal increases the chance (can win in spite of the master) but only if it's sort of possible (it needs to have enough religions and enough cities).

Culture 2:
Once a bit further into the game the Cultural AI moves onto stage 2, this involves heavy investment in missionaries, religion spreading, temple building and so on and so on. Anything which gives culture is valued more.
Culture 2 deactivates Dagger.
Note that AI's in Culture 2 can tech very fast.

Culture 3:
Culture 3 involves putting A LOT more weight on culture and pursuing only techs which are immediately useful (Nationalism, Military Tradition, UU tech etc), it is activated once the AI feels confident that it can win a CV.

Culture 4:
Once the AI has no more useful techs on the radar it turns off research and cranks up the culture. This can be deactivated by a war declaration, if the AI is close to a not-so-useful-but-not-useless military tech (the AI might for example decide to pursue Artillery only during a war).
 
Great!

So we have strategy for military and cultural victories.

But i still don't see strategy for space race.
Basically a technology strategy.
Most of the cities should be improved for research and some improved to be production powerhouses (based on terrain) to get the fresh wonders and the spaceship eventually.

I'm not too good at food city management.
Perhaps they could be neglected by this strategy.
Although scientists with representation are not bad for research.
 
Thanks, Blake. I do have a question - in some games, I've noticed that no strategy appears for some AIs that are currently at peace. Is this supposed to happen? What does it mean?

And I second Kettyo's question about space race/technology -- are we eventually going to see a strategy for the AI to follow here -- or is this more or less the 'default' strategy?

Thanks again for all your work on this.
 
Very good to know Blake. I have a couple of suggestions as well. I didn't see any defense strategies (maybe these fall under war strategies (limited war, etc...) I suggest

Pillage Counter
- Create / recall units to a counter stack (w/ siege units) to kill a pillaging stack.
- A pillaging stack could be identified maybe by not including any siege units (after they are available).
- Increase willingness to sue for peace. IMO a pillaging war usually indicates the undesirability of capturing a city. Probably (if it is a player) we would just prefer to end the war.
- I recently had pretty much the run of an enemies territory while he threw units one or two at a time at my 12 unit stack.

Crush Counter
- relocate city garrison units from nearby cities to city under siege.
- hurry units (draft, whip, buy) in city under siege and nearby cities.
- consider switching to a civic that will allow hurrying units if the civ cannot do so now, especially if civ is spiritual.
- increase attempts to bribe other civ's against the attacker.

I've recently seen that the defending civs don't really seem to be fighting for their lives, but rather going through the motions. Of course we don't want to make them too good at defending right, or we end back up with the unit spamming problem...
 
So that's what these do. A couple of things:

Does the Get Better Units strategy know about barbarians? I know the mod shouldn't be optimised for raging barbarians, but without enough archers a civ has little chance of connecting resources under it.

Even at Culture 2, there isn't much of an excuse not to be in Free Speech. More emphasis on artists might be helpful too. I've seen many occasions when more mid-game culture would have gained an extra city by flipping.

Getting missionaries across the sea could help Cultural victory as well and the Missionary strategy manages it eventually so it should be possible.
 
Top Bottom