Page 1 of 72

TRiBaL_Warriors

Posted: Sun Feb 21, 2016 7:59 am
by Esty
TRiBaL Awards.png
TRiBaL Awards.png (339.58 KiB) Viewed 8000 times
  • AI for RM standart settings - Dark Age start
  • LAST UPDATE : july 2023 ---
  • It plays every civ from UP and Wololo kingdoms
  • It supports Wololokingdom
  • It supports DE
  • It can perform solo and team games
  • It can play every map except migration

Re: TRiBaL_Warriors (new AI)

Posted: Sun Feb 21, 2016 10:27 am
by CheeseOnToast
Hi!

For a first script its looking like its starting to take shape, however I've been looking at some of the code and am quite confused to a bit researches I saw, for example:

Code: Select all

(defrule
	(unit-type-count archer-line >= 10)
	(can-research ri-iron-casting)
=>
	(research ri-iron-casting))

(defrule
	(unit-type-count archer-line >= 10)
	(can-research ri-chain-mail)
=>
	(research ri-chain-mail))
	
The researches here are for infantry and not archer units. Is this a mistake? I haven't tested it in-game yet but I will give it a look when I can!

Re: TRiBaL_Warriors (new AI)

Posted: Sun Feb 21, 2016 1:06 pm
by Esty
Ooh thank you sausagebread, i am a bite confused with some upgrades ; i modified it already. ty

Re: TRiBaL_Warriors (new AI)

Posted: Sun Feb 21, 2016 2:46 pm
by scripter64
Welcome to the forums, estyty! Downloaded :)

Re: TRiBaL_Warriors (new AI)

Posted: Sun Feb 21, 2016 4:13 pm
by Promiskuitiv
Also downloaded and welcome! :D

Re: TRiBaL_Warriors (new AI)

Posted: Mon Feb 22, 2016 5:06 am
by offwo
I tested this script 4v4 against The_Unknown and was generally quite impressed. It survived one and a half hours, although in part because one of my pockets had an attacking bug. I watched from the point of view of a mayan pocket.

Comments: (most of this was for the mayan pocket strat, I didn't really look at the other players)
- Although making a lumbercamp first is generally a better idea than making a mill first, I think you should go mill first if you detect foragers and haven't found sheep. Your mayan player had idle town center time (one and a half vills behind) and its villagers spent a long time dropping forage back to the town centre.
- It tried to build a deer mill, but its dropsite distance wasn't quite far enough to actually build it by the deer.
- I think your fast castle build order should go to the feudal age about a minute later. Ideally, for a fast castle build order, you should go up in such a way you only need to create two or three villagers in feudal (for each vill you make in dark rather than feudal, you gain 40 extra resources during the dark/feudal transition). Most scripts take the approach of not advancing until X villagers (usually 27-31), but I prefer simply delaying the research until you have amassed about 600-650 food.
- I think you can afford to take bow saw at the start of the castle age. It makes a big difference, and you never had less than 150 food in castle, so it shouldn't cause idle town center time. If you look at recorded games of pro human players on sites like aoczone, you will see that in 9/10 games, they will research bow saw right away in castle age almost instinctively. If you want, you could try researching wheelbarrow a few minutes earlier too.
- Your help system is good, but one flank went to help the other at one point. This isn't generally a good idea, but it worked this game.
- I think the mayan player mines too much stone in castle. Either it should go full stone mining during transition to castle and make some plumed archers to help the flank, or it should ignore stone (buy 100 for 4th tc) and focus on clicking imp ASAP. I don't think the castle it got 5 minutes into castle served much.
- Your AI built 3 blacksmiths while imping, 1 or 2 are enough.
- I think you should try building your barracks during transition or even after you have built your monastery and siege workshop. It would help it flood quicker, and leave the pocket defenseless for less long.
- Lategame, you should consider training some elite plumed archers against gunpowder units. It kept trying to fight massed elite conquistadors with EEW and halbs.

This sounds negative but this is one of the best first scripts I've tested, I will probably start adding it to my MP matchups.

Re: TRiBaL_Warriors (new AI)

Posted: Mon Feb 22, 2016 6:30 am
by Esty
Thank you offwo. Your comments will help me a lot. I start working on it.

- It tried to build a deer mill, but its dropsite distance wasn't quite far enough to actually build it by the deer.
Actually my eagles are not defined as "scout units" ; so they dont lure deers to tc. As soon as its fixed, i wont make it buid another mill.
I'll have a look about your idea to make a mill first in case i dont have early sheeps ; and i will try to delay my feudal research :) ! Let's see...
Lategame, you should consider training some elite plumed archers against gunpowder units. It kept trying to fight massed elite conquistadors with EEW and halbs.
YES ! I have not already made any counter rule. That's why it dont care about what is the opposite army. It's my next step :dance

Re: TRiBaL_Warriors (new AI)

Posted: Mon Feb 22, 2016 11:06 am
by Esty
Reupload !

Changes :
- Eagle-scout now lure deers to tc
- Azteks and mayas pockets make barracks during the imperial-age research
- No more 2nd mill built in dark-age
- timer for forcedrops to prevent idl tc in dark-age
- Flank rushers will make archers while researching castle-age for bigger castle army
- pocket boomers will make more vills before feudal-age research

I dont know how to force making a mill if sheeps are not imediately found. I'l look at it asap.

Thanx every testers ! I keep waiting any more comments.

Re: TRiBaL_Warriors (new AI)

Posted: Mon Feb 22, 2016 4:02 pm
by el+noob
Hello, I think the poket should go rush of knights in castles

Re: TRiBaL_Warriors (new AI)

Posted: Tue Feb 23, 2016 6:15 am
by Esty
Hello el+noob
Hello, I think the poket should go rush of knights in castles
I know that going Kts in castle age for the pocket are always more effective ; that's why most of the AI use this strategy. For my first experience in scripting i wanted to work on a strategy based on a boom. My idea is to optimize this strategy ; Actually, the interesting part of my work is to make my flanks able to survive till my pockets start their imperial attack. :D
When my AI sounds better in that way, i'l prolly work on other strategies.