Daedric is made with the intention to be used on my own lan-party's It has some ability's to team play. also quite a good oponent in 1v1's
Game-type: any (RM, PIDM)
Map: Any map expect migration
resources:any(standard)
victory: any(conquest)
civ: random
Currently:
Flush:
skirm-rush
archer-rush
mix-rush (skirms and archers)
Early-Eagles
Grush
castle:
Xbows
Knights
DoubleAgressive
Defensive castle boom
Castle-water
imperial:
Arbalest
Paladins
Imp-water
DM:
dm-arbalest
dm-paladin
dm-champion
dm-unique
Any feed back is welcome and apreciated
Daedric AI v1.10aa
-
- Swami
- Posts: 234
- Joined: Mon May 06, 2013 12:37 am
Daedric AI v1.10aa
- Attachments
-
- Daedric v1.10aa.zip
- (75.69 KiB) Downloaded 522 times
Last edited by Armelon on Thu Jul 06, 2023 3:54 am, edited 39 times in total.
- KiiNgZoNe
- Forumer
- Posts: 62
- Joined: Fri Mar 02, 2012 8:15 am
Re: Daedric AI
Hello there! Welcome to the world of !!! (as many call it )
I'm very glad there are new scripters coming! I finally get a rookie with me as well so I don't feel left out as a newbie
I'll be looking into this script and maybe suggest a few things (I'm not an experienced AI scripter myself, but I got a knack for some things )
Hopefully this doesn't destroy my poor AI in a punishing manner, 'cause I'm doing my best as well!
I'm very glad there are new scripters coming! I finally get a rookie with me as well so I don't feel left out as a newbie
I'll be looking into this script and maybe suggest a few things (I'm not an experienced AI scripter myself, but I got a knack for some things )
Hopefully this doesn't destroy my poor AI in a punishing manner, 'cause I'm doing my best as well!
- LightTree
- Waheguru
- Posts: 1608
- Joined: Wed Mar 10, 2010 12:21 am
Re: Daedric AI
Welcome .
It's really nice to see new scripters and more activity around, all the best luck on both Daedric and Vetrix .
It's really nice to see new scripters and more activity around, all the best luck on both Daedric and Vetrix .
-
- Waheguru
- Posts: 5891
- Joined: Fri Jan 16, 2009 8:36 pm
Re: Daedric AI
Armelon! Welcome to the forum
-
- Forumer
- Posts: 6
- Joined: Fri Jan 01, 2010 3:46 pm
Re: Daedric AI
Going to try it out right now!
- Promiskuitiv
- Waheguru
- Posts: 1364
- Joined: Thu Nov 10, 2011 1:07 pm
Re: Daedric AI
Welcome to the forums!
- marathon
- Waheguru
- Posts: 1984
- Joined: Fri May 14, 2010 2:30 pm
Re: Daedric AI
Hello Armelon, welcome to the world of !
Your AI seems pretty solid already. It has good strategic goals, good unit output, good age times. I tested it vs Eagles Revenge (Aztec wars) and it won with archer flush, not bad at all
You could drop first house near a villager.
You could check (dropsite-min-distance deer) before dropping hunting mill. For most civs, it is not worthy to drop mill when this is under 20 (Japanese or Mongols can perhaps waste the wood if it is 15 tiles).
You should wait a few seconds before dropping first mill. Sometimes the area there is not revealed well, and if you wait a forager may go there and reveal map.
Surprisingly strong for a first script.
Your AI seems pretty solid already. It has good strategic goals, good unit output, good age times. I tested it vs Eagles Revenge (Aztec wars) and it won with archer flush, not bad at all
You could drop first house near a villager.
You could check (dropsite-min-distance deer) before dropping hunting mill. For most civs, it is not worthy to drop mill when this is under 20 (Japanese or Mongols can perhaps waste the wood if it is 15 tiles).
You should wait a few seconds before dropping first mill. Sometimes the area there is not revealed well, and if you wait a forager may go there and reveal map.
Surprisingly strong for a first script.
-
- Swami
- Posts: 234
- Joined: Mon May 06, 2013 12:37 am
Re: Daedric AI
Thank you everyone for welcoming me so kindly! :3
And thank you Marathon for right away reviewing my AI.
I have made some small AI's before but without the userpatch thats probaly why it looks pretty decent.
Still trying to get the hang of all the things possible with the userpatch and hoping to be able to get more and more UP things in there.
And thank you Marathon for right away reviewing my AI.
I have made some small AI's before but without the userpatch thats probaly why it looks pretty decent.
Still trying to get the hang of all the things possible with the userpatch and hoping to be able to get more and more UP things in there.
-
- Guru
- Posts: 639
- Joined: Thu Jul 14, 2011 8:44 am
Re: Daedric AI
Daedric is very strong for a new scripter, and it even 3-0'd Crusade. Its flush seems well-balanced.
Here are some things I noted:
-Man-at-arms rush could start training militias already at dark-feudal transition (if not Goth) to have more units out faster.
-Xbow strategy may train spearmen even if enemy doesn't use any cavalry.
-Skirmisher rush mined a lot of gold without mining camp at late feudal age.
-Mining camps should probably be built with sn-allow-adjacent-dropsites 0 to leave villagers more room, since the closest gold piles will last for a long time and not enough villagers can fit.
-With pure archery flush blacksmith shouldn't be dropped before there are two ranges
-Daedric delays wheelbarrow too much with flush, which hurts economy. Wheelbarrow is a key upgrade for AIs, though it may be too expensive at early feudal with fast up time.
-After flush Daedric gathered too much food and gold and not enough wood. It hit castle age with >1000f and >700g in a game, and with villagers being only food-unit it trained it couldn't send the food. In another game Daedric hit 1000w before it could click up to feudal after skirmisher flush, since it lacked gold for a pretty long time.
-Late game camp rules probably worked with a timer instead of building a large number of camps once certain game time is reached. It's also good to build a new camp if dropsite-min-distance gets too large.
-With mining camps you should check whether there is gold/stone on map left instead of building more camps. Dropsite-min-distance for them is -1 if there's no resource visible or 255 if only resources visible are on another island.
-In a game right at start a villager went to hunt deer ~15 tiles away.
Welcome to AI Scripters
Here are some things I noted:
-Man-at-arms rush could start training militias already at dark-feudal transition (if not Goth) to have more units out faster.
-Xbow strategy may train spearmen even if enemy doesn't use any cavalry.
-Skirmisher rush mined a lot of gold without mining camp at late feudal age.
-Mining camps should probably be built with sn-allow-adjacent-dropsites 0 to leave villagers more room, since the closest gold piles will last for a long time and not enough villagers can fit.
-With pure archery flush blacksmith shouldn't be dropped before there are two ranges
-Daedric delays wheelbarrow too much with flush, which hurts economy. Wheelbarrow is a key upgrade for AIs, though it may be too expensive at early feudal with fast up time.
-After flush Daedric gathered too much food and gold and not enough wood. It hit castle age with >1000f and >700g in a game, and with villagers being only food-unit it trained it couldn't send the food. In another game Daedric hit 1000w before it could click up to feudal after skirmisher flush, since it lacked gold for a pretty long time.
-Late game camp rules probably worked with a timer instead of building a large number of camps once certain game time is reached. It's also good to build a new camp if dropsite-min-distance gets too large.
-With mining camps you should check whether there is gold/stone on map left instead of building more camps. Dropsite-min-distance for them is -1 if there's no resource visible or 255 if only resources visible are on another island.
-In a game right at start a villager went to hunt deer ~15 tiles away.
Welcome to AI Scripters
- Promiskuitiv
- Waheguru
- Posts: 1364
- Joined: Thu Nov 10, 2011 1:07 pm
Re: Daedric AI
Personally i don't like that you copied some code of The Horde, but if you asked him before it should be okay.
Marathon also has a good view on this: Something along the lines of "Why research something that someone else researched already" (correct me if i'm wrong) and while i generally agree with that i think AI strategies itself shouldn't be copied. But i don't know how much you copied anyway, just looked into your AI for one minute and already noticed that several rules came straight out of The Horde. I also really hope i don't sound hostile saying this, it's just something i personally dislike - but that doesn't change my view of you as a person (although i don't know you yet anyway).
I'll test your AI later and leave you some feedback, might take me a couple of days though.
Also
Would always select the x-bow strategy if the random-number is higher than 0 (even if it's above 100).
Maybe add some condition to check if a strategy has been selected already (so that the second rule can't fire if the first already did).
If these goals aren't strategy goals (like =>krush/=>x-bow rush) but only determine which units are allowed to be trained or something like that then forget about what i said (same if i overlooked something, i'm sleepy as fork right now).
Marathon also has a good view on this: Something along the lines of "Why research something that someone else researched already" (correct me if i'm wrong) and while i generally agree with that i think AI strategies itself shouldn't be copied. But i don't know how much you copied anyway, just looked into your AI for one minute and already noticed that several rules came straight out of The Horde. I also really hope i don't sound hostile saying this, it's just something i personally dislike - but that doesn't change my view of you as a person (although i don't know you yet anyway).
I'll test your AI later and leave you some feedback, might take me a couple of days though.
Also
Code: Select all
(defrule
(random-number > 100)
(goal my-random-number 3)
=>
(set-goal Knights YES)
(chat-to-allies "Going for double stable")
(chat-local-to-self "Knight-style picked")
(disable-self)
)
(defrule
(random-number > 0)
(goal my-random-number 3)
=>
(set-goal Xbows YES)
(chat-to-allies "Going for Xbows")
(chat-local-to-self "Xbows-style picked")
(disable-self)
)
Maybe add some condition to check if a strategy has been selected already (so that the second rule can't fire if the first already did).
If these goals aren't strategy goals (like =>krush/=>x-bow rush) but only determine which units are allowed to be trained or something like that then forget about what i said (same if i overlooked something, i'm sleepy as fork right now).