Dominions 2. AI. Suggestions, that how to fix it.

[Zerger]

Originally posted by MStavros:
Okay I decided to make this topic, since the AI is very weak.

As I see, the biggest problems are, that the AI:

1.) isnt recruiting enough units.
2.) is trying to operate with small 'blitz' armies, I never ever seen a bigger AI army in the demo so far.
3.) isnt protecting its provinces very well
4.} isnt making smart strategical moves


Tell me what you think. I think these are the major problems with the Doms 2. AI.

1 - I agree. However some AIs had nice amount of units in some of my games, but all of those were weak units. They were toast against my heavyweight army.

2. - Very true

3. - Hm I don't agree with this. This part is working well. When I leave a province with a small defense force only, the AI always trying to conquer it with a small army, and he is getting it almost every time.

4. - I agree. One of the biggest problem is, that the AI has no counter for any sort of
efficient strategy.


2 main problems with the AI:
* The AI won't recruit an effective army. Mostly it is trying to attack with tiny - mediocre armies.

** The AI has no counter for any sort of
efficient strategy.


Setting indy str to 7++. It helps a bit, that will slow down the human player, but it also slow down the AI. Also I noticed that when I set it to 7, 8 or 9, and I reach the AI, its hellish easy to conquer the AI's provinces, since the AI had killed the indy armies, and propably the forces of the AI are almost totally dead when I reach the freshly conquered provinces.

===>> This won't help at all in anything.

[ October 29, 2003, 12:54: Message edited by: Zerger ]

[apoger]

The main issue is that humans use very focused strategies, compared to the computer that tends to dabble in a bit of everything.

Sadly also (just like Dom I) it looks like the computer doesn't use it's pretender for anything beyond sitting in the capital doing research, casting rituals, and making magic items. It doesn't use the predtender for combat advantage the way a human would. As far as I can tell, it also doesn't use the pretender for site searching (in person).

On the impossible level Illwinter should have hardcoded more focused behavior. Only allowing certain units, certain research, and certain spells to cast. It wouldn't be as good as a human, but could mimic some human behavior.


Hey IW- If you guys want I could come up with a few simple rules for each nation. Perhaps you could code that into a new level of AI.

[ October 29, 2003, 13:06: Message edited by: apoger ]

[ywl]

A scripting language which can handle the game situation is nice on paper. But it's a tremendous task - that's why I put a on my comment. So, unless the developers say that they're willing to work on it, or somebody could send them a language spec and preferrable a compiler, it'll be more or less remain just an idea.

It might be fun and easy enough to give the AI different preference to exploration, conquering, etc. But that will only change the weakpoint AI in strategy to something else . It wouldn't really solve the major issue here.

Overall, I don't think it's even realistic to expect an AI can play such a complicated game well. But some simple fixes should be able to make the AI more challenging - and we have proposed some.

[licker]

Originally posted by Gandalf Parker:
The area of unit selection is really hard to move from complaining to real suggestions. To keep a database of each race and each gaming situation would be extreme. They might as well move the AI to a seperate program so it doesnt slow down the MP play. Many of the people complaining about the AI are playing with independents set at 2 or 3. The strategies and units for such a game would be very different than if it were set in the middle, and that would be different than the high setting. And thats just ONE of the pre-game settings.

OK Ive heard that the AI should not buy cheap units but instead buy a "real army". So insteadof figuring out how to get the most units with the least upkeep (patrolling strtegy) they should buy the most expensive? That would lead to armys of hvy cavalry, which I guess isnt too bad.

Personally I think an AI best bet is to buy everything. Start at the cheapest and buy 1, move up thru the choices then start over. They would reach a point of not being able to buy another hvy cavalry or elephant or hydra and would use up the left over amount on cheap troops so they would end up with slightly more lesser troops than expensive ones. But that wouldnt be too bad.

What about leaders? How do you make sure the AI buys a mix of leaders? It has no memory so it cant say "I bought this Last time and now need to buy this". Hmmmm well actually it might be able to fake that. If turn number divides by 5 evenly then buy the most expensive leader THEN buy troops. *sigh* still not great but maybe it would be better?

Certainly the AI can look at its army composition at the start of its turn (units and leaders) and build appropriately no? Say it sees a deficiency in heavy troops, or mages, it can start building those troops until its reached its equilibrium level. This is sort of the approach MoO3 used actually for fleet compostion, it was moddable and eventually the modders were able to 'force' the AI to build more intelligent fleets by changing the ratios of scouts to beams to missiles...

I don't see why this wouldn't work for Dom as well. You'd have a list of commander types (maybe also broken down into magic types for different mages) and a list of unit types (classes rather than specific units so indies fit in). Then a weight is assigned to each type, a crude example for just infantry could be:

LI: 25%
MI: 50%
HI: 25%

So the AI would look at its current army and see that it had only 10% LI, then it would build more LI to get the ratios set. Something like that should be made external so that modders can tweek away to their hearts content. In fact you can use lists like that to control many aspects of AI choices, you can do it for magic, artificing, recruiting, summoning, ... and have master matricies to control the embedded ones...

That's what I'd like to see anyway

[licker]

Originally posted by ywl:
A scripting language which can handle the game situation is nice on paper. But it's a tremendous task - that's why I put a on my comment. So, unless the developers say that they're willing to work on it, or somebody could send them a language spec and preferrable a compiler, it'll be more or less remain just an idea.

It might be fun and easy enough to give the AI different preference to exploration, conquering, etc. But that will only change the weakpoint AI in strategy to something else . It wouldn't really solve the major issue here.

Overall, I don't think it's even realistic to expect an AI can play such a complicated game well. But some simple fixes should be able to make the AI more challenging - and we have proposed some.

It would change the AI weakpoint to what exactly? Certainly if you think the AI is weak in many areas, improving some of them is worthwhile, it doesn't make the unimproved areas any weaker, other than on a reletive scale.

The easist thing to fix in the strategic AI is probably the unit balance within a nation. I would guess that if that were fixed (or made moddible to let the modders play around with it) other defiencies would be less obvious since the main hanicap of not fielding competative armies would be gone.

I'm sure it's a deeper issue than we touch on though, only the Devs can tell us what is realistically possible. We are just presenting different ways to look at the problem. (I hope )

[Mortifer]

Well if you ask me, separating various AI scripts sounds very good, however it would work? I have no idea, I am not an AI scripter.
The devs will know that this is a good idea or not.

Once again, the first and biggest problems are that the AI tends to use weak or mediocre units, also bigger AI armies are very rare. (in the demo)


I have a question: the AI is casting rituals in a proper way? IE. Is it summoning creatures and enlist them in his armies? Perhaps Beta testers could answer this?
(I barely saw summoned creatures in the AI armies in the demo)

[-Storm-]

Hey, maybe JK should comment these AI suggestions. He coded the AI...

[licker]

Originally posted by -Storm-:
Hey, maybe JK should comment these AI suggestions. He coded the AI...

Yep it would be nice for a Dev to step into this thread and give some reaction. Though depending on how much of the AI is hardcoded it may be difficult to really effect the types of changes suggested here...

[MythicalMino]

ok, here is a report on my current game...

Playing the Aran map, Indeps set to 4 strength. 5 other AI's, 3 Easy, 2 Normal...I am playing as C'tis (Miasma theme)

Caelum just attacked me with about 30 troops...the make up of the troops:

Wingless (had several of these)
Blizzard Warriors (Several of these)
Spine Horn Warriors (had several of these also)
1 Wyvern
Seraphine
Caelian Infantry
1 Temple Guard
also, they were attacking with their Prophet

In another border province, they have about 20 troops, mostly made up of Spire Warriors, and they have a Wyvern there also.


On my other border, Michtlan is making their move i think, against me:

110 Troop army:

made up of Warriors, Slaves, Sun Warriors and several Fiends of Darkness

They are placed as 4th place in Army Size. 3 other AI are ahead of them. Michtlin is a Normal AI, Caelum is easy AI.

Anyways, there is my report for this game....what i have seen so far....I have come across several times 20+ armies on my borders, but they pull their troops elsewhere...possibly to fight on another border with another nation? IT would seem....cause the AI's are losing/gaining provinces quite a bit...

[Gandalf Parker]

originally posted by licker
I don't see why this wouldn't work for Dom as well. You'd have a list of commander types (maybe also broken down into magic types for different mages) and a list of unit types (classes rather than specific units so indies fit in). Then a weight is assigned to each type, a crude example for just infantry could be:
LI: 25%
MI: 50%
HI: 25%
So the AI would look at its current army and see that it had only 10% LI, then it would build more LI to get the ratios set. Something like that should be made external so that modders can tweek away to their hearts content. In fact you can use lists like that to control many aspects of AI choices, you can do it for magic, artificing, recruiting, summoning, ... and have master matricies to control the embedded ones...

Seems simple and workable. On the other hand, lets pick one race and try to write it all out. Ulm should be simplest. It has a cheap and expensive Version of each of its units which is nice. But I suspect that anything we come up with to "fix" this will amount to fixing it for the indep 2 games people are playing (fast player conflict) and will need fixing again for 4, 6, 8. Then for levels of magic research, and levels of resources, etc. Im thinking that this level of "thinking" might end up being equal in size to the game now (Ive run into this problem before).