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

[Kaljamaha]

Another thing that I think should be looked at is AI dominion design. Now, granted, I'm no pro, but sometimes the AI makes design decisions that are utterly suicidal. It has effectively lost before the game even began. One example, from Dom I would be Marignon with Turmoil 2, Cold 3, Death 3. Can't remember if it had sloth also. Granted, those are worth a lot of points, but I'm pretty sure that whatever the AI bought, it wasn't worth it. In the Dom II demo, turmoil seems to be a favorite of many AI opponents, and I can't figure why, as I'm always strapped for cash. As someone mentioned earlier, human designed pretenders for AIs would probably make a lot of sense. As I play mostly SP, it always pains me (when playing Dom I) to see the AI shoot itself in the foot with an idiotic domain.

Next, some anecdotal evidence of AI brilliance. I was playing Abysia, and was besieging the Tien Ch'i capital. I had Anathemant Something, plus three Warlocks as combat mages (probably not a great choice). Anyway, while sieging, there was a break-out attempt, as there always is. However, this time was different. It consisted of one commander, a mage with 3 Astral and something else, plus a few traditional troops. What happened was that the normal troops drew my army's attention long enough for the mage to slay all three of my Warlocks with Magic Duel. Naturally, the mage died in the end, but if this was a deliberate move instead of a random happening, it shows an amazing cost/benefit analysis by the AI. Anyway, just wanted to share the story,it gives me some faith in the AI.


K.

[Gandalf Parker]

OK Mort, Lets pin down castle tactics. Im just rambling out loud here...........

So you would have the AI maintain (a scout? priest? small army?) which moves seperate from his larger expansion army. When it finds a province that has (I forget what the formula players say whether its high resource or high gold). The province needs to have no bordering neighbors that are not owned by the AI (avoiding front line positions) and no bordering neighbor that already has a castle. That sounds good.

And by preference it should have 2 bordering provinces which are water or mountain? Im not sure how code would go about recognizing a chokepoint. Two water neighbors which dont also border on each other? It might be able to see that. The impenetrable mountain walls like on the Eye map Im not so sure about. That would be "a province which should be a neighbor but doesnt have a #neighbor statement" which is what really makes them impenetrable mountains. That extra-thick borderline is just for visiblity.

But even if it cant specifically recognize chokes (I dont know if it can or not) that might not be too bad . Mountain areas are supposed to represent a difficult terrain for movement in Dom2. And if the 2 water neighbors happen to be next to each other thats not so bad since I often discover a need for some coastline defense.

Hmmm but the formula so far would want me to own those 2 water provinces. And what about the water nations? Hmmmmm I think this formula is getting "smarter" than its worth.

[Particle]

Originally posted by Kaljamaha:
Another thing that I think should be looked at is AI dominion design. Now, granted, I'm no pro, but sometimes the AI makes design decisions that are utterly suicidal. It has effectively lost before the game even began. One example, from Dom I would be Marignon with Turmoil 2, Cold 3, Death 3. Can't remember if it had sloth also. Granted, those are worth a lot of points, but I'm pretty sure that whatever the AI bought, it wasn't worth it. In the Dom II demo, turmoil seems to be a favorite of many AI opponents, and I can't figure why, as I'm always strapped for cash. As someone mentioned earlier, human designed pretenders for AIs would probably make a lot of sense. As I play mostly SP, it always pains me (when playing Dom I) to see the AI shoot itself in the foot with an idiotic domain.

K.

Yup, I've noticed this also.
As for forts....hrm a complicated question. I would say the AI should build castles 3x or 4x more often than right now. Sometimes I see 1-2 AI castles before turn 40, so 4-8 castles before turn 40 would be ok.

[-Storm-]

Just a little addition to the army setup:
IMHO the AI should build 50% LI before turn 10, 40% until turn 15, 30% until turn 25, 25% until turn 30, 20% until turn 40, and >20% after turn 40.
As for the rest of the units, they should be HI - HC - Creatures - Heavy Bow/Xbow of course.

[ November 14, 2003, 20:55: Message edited by: -Storm- ]

[Zerger]

Originally posted by -Storm-:
Just a little addition to the army setup:
IMHO the AI should build 50% LI before turn 10, 40% until turn 15, 30% until turn 25, 25% until turn 30, 20% until turn 40, and >20% after turn 40.
As for the rest of the units, they should be HI - HC - Creatures - Heavy Bow/Xbow of course.

Hrmmmmm I dunno. LI is very weak in Dominions 2. I think the AI should avoid to use them, MAYBE if a patch will make LI more useful......until than I don't wanna see LI from the AI at all. They are weak, routing easily etc. etc.

1 suggestion: the AI should use more army Groups on the battlefield, and use some tactics.
[Ok this is nothing new, I guess we all know about this already, but Aristoteles didn't posted any details about the tactics part.]

[ November 15, 2003, 13:33: Message edited by: Zerger ]

[Gandalf Parker]

Originally posted by Zerger:
Hrmmmmm I dunno. LI is very weak in Dominions 2. I think the AI should avoid to use them, MAYBE if a patch will make LI more useful......until than I don't wanna see LI from the AI at all. They are weak, routing easily etc. etc.

Ummm have you tried playing that way? Yes the LI units have some drawbacks, they also have some advantages (such as cost, and patrolling). But I dont think there are very many nations that can afford to use no LI at all. That would make an army stronger, but make the AI easier to beat.

I disagree with Zerger on this point. I haven't found a more cost effective way of getting more provinces quicker other than in the first few turns than with LI.

Maybe it would help if you posted what you would build in the first 4 turns as opposed to LI and that would help.

[NTJedi]

Storm has half of the idea correctly... which is as turns pass the AI decisions should also change. However all previous suggestions from anyone regarding a fixed percentage is wrong. This will cause human players to see Army_A, Army_B and Army_C as almost identical. The percentages used for LI, HI, ranged_units, etc, etc, ... these percentages should be a random range thus keeping human players guessing. Human players could easily master one fixed percentage.

Here's an example:
Day_1 thru Day_15: LI =20% thru 40%, HI = 20% thru 30%, LC = 2% thru 10%, HC = 2% thru 5%, Creatures = 5%.
Weak ranged units = 20% thru 30%, Moderate ranged units = 5% thru 15%, Strong ranged units = 1% thru 5%.
Day_16 thru Day_35: The percentages should change based on its kingdoms total strength.

The above is just an example... but I guarantee it would make the AI more interesting then any fixed percentage.

[ November 15, 2003, 15:18: Message edited by: NTJedi ]

There is a limitation that we are dealing with; as far as the AI as a whole. As you have repeatedly said "Should be" different themes regarding AI, they have said they will not be able to do so at least as far as this game is concerned.

I agree it would be better to leave in variables; but I think in order to give the AI the best chance for not being abused or exploited you have to find it's weaknesses and compensate for them. As in:

One of the AI's weaknesses in the demo is the low production of mages. Even mage heavy nations like Arco do not produce their main line of fighting commanders are standard commanders, and are not using Arco priestesses as they should, healing their stacks after a battle if they have time.

Research should be left to apprentices because they can be produced every turn and not upper end casters unless they have a high drain scale and need to do so.

Things of that nature. Which may change in the full game, I do not know.

Edit: For numbers I haven't found a better way to beat neighbor provinces with anything other than LI with anything but High Bless Jotun Woodsmen (Nature or Earth for the most part), High Bless Black Hunters (Nature), and Ulm Heavy Inf, and that is only if you attack with your pretender to keep up efficency if you have a SC pretender.

[ November 15, 2003, 15:29: Message edited by: Zen ]

[PDF]

Originally posted by Zen:
I disagree with Zerger on this point. I haven't found a more cost effective way of getting more provinces quicker other than in the first few turns than with LI.

Maybe it would help if you posted what you would build in the first 4 turns as opposed to LI and that would help.

Maybe LI can help you if you play a SC pretender , Indep <=4, and happen to stumble on Barbarians/militia provinces.
With Indep at 6 or more, a Rainbow Mage or more heavily defended provinces you won't get far with LI. They're Ok for patrolling, give fodder targets to enemy archers, but rout in doves if ever confronted to a pair of knights ..