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 ;
TRiBaL_Warriors
- THE RULER
- Guru
- Posts: 655
- Joined: Sat Feb 20, 2010 10:55 am
- Location: FRG
- Contact:
Re: TRiBaL_Warriors (new AI)
It's poison and it's blood and big fire, big burn - into the ashes and no return
-
- Waheguru
- Posts: 5891
- Joined: Fri Jan 16, 2009 8:36 pm
Re: TRiBaL_Warriors (new AI)
RULER! Welcome back
- Esty
- Waheguru
- Posts: 1159
- Joined: Wed Dec 30, 2015 7:37 am
Re: TRiBaL_Warriors (new AI)
Hi,
Just reuploaded my AI with some changes :
This code is not really efficient...
Just reuploaded my AI with some changes :
- my flank will mine stone in feudal
My flanks will make defensive towers
Counter : now my AI will counter with the adequate units
TSA is now used
Code: Select all
(defrule
(building-type-count lumber-camp == 1)
(building-type-count-total watch-tower < 1)
(up-pending-objects c: watch-tower < 1)
(can-build watch-tower)
=>
(up-set-placement-data lumber-camp 0 c: 15)
(set-strategic-number sn-allow-adjacent-dropsites 1)
(set-strategic-number sn-placement-zone-size 2)
(up-build place-control 0 c: watch-tower))
Author of TRiBaL_Warrior, Yggdrasil, PharaoN, WoodyWood Pecker and currently dette
-
- Waheguru
- Posts: 5891
- Joined: Fri Jan 16, 2009 8:36 pm
Re: TRiBaL_Warriors (new AI)
Try c: 0 instead of 15 to center the placement zone at the camp and please be sure to set the sn to ignore tower elevation.
- Esty
- Waheguru
- Posts: 1159
- Joined: Wed Dec 30, 2015 7:37 am
Re: TRiBaL_Warriors (new AI)
New update :
significant modifications :
- Pocket ai can now decide to rush with KTS instead of Booming (it will tell what he gonna do game-time == 10
- Flank will make towers exactely where i want it to build (on own LC, own gold mine..) for a better defense
- retreat system is corrected... It was buggy because my script was a mess
After some tests i can say that my AI is improving. It's stile weak, compared to the great Ais i found here but its good to know that with a really basic knowledge it's possible to make something able to work.
significant modifications :
- Pocket ai can now decide to rush with KTS instead of Booming (it will tell what he gonna do game-time == 10
- Flank will make towers exactely where i want it to build (on own LC, own gold mine..) for a better defense
- retreat system is corrected... It was buggy because my script was a mess
After some tests i can say that my AI is improving. It's stile weak, compared to the great Ais i found here but its good to know that with a really basic knowledge it's possible to make something able to work.
Author of TRiBaL_Warrior, Yggdrasil, PharaoN, WoodyWood Pecker and currently dette
-
- Waheguru
- Posts: 5891
- Joined: Fri Jan 16, 2009 8:36 pm
Re: TRiBaL_Warriors (new AI)
Downloaded, thanks
- Esty
- Waheguru
- Posts: 1159
- Joined: Wed Dec 30, 2015 7:37 am
Re: TRiBaL_Warriors (new AI)
Hello,
Back for the new version of TRiBaL_Warrior AI !
First of all, i finally fixed the retreat system which makes a real difference in skill with the old version. My units are not wasted that easily anymore
Moreover, Daedric AI influenced me a lot in the many attacking rules i used here to complete the old ones (i copied some and adapted some others and didn't forget to specify him as the author when needed). I know its not great to copy the codes, and the more i'l feel comfortable with scrpting, the less i'l do it. But i'm stile rusty
NOW, Pocket and flank have each two strategies available :
- flank now will go for a huge feudal rush, massing archers/skirms with some defensive towers or will go for castle with one archery+towers.
Pocket will go for a fast castle Kts or will try to BOOM fast and prepare a strong imperial army.
Just tested some 4v4 Tribal vs Daedric and then Tribal vs Juggernaut. was not bad at all. It maybe shall start to build forward in case it got military superiority... and make bombard towers.
Plz test it and gimme feedbacks about this new version.
Now i want to make it work on other maps : Island, Nomad (copied already some codes like : (set-strategic-number sn-percent-attack-boats xxx) ready to be used !
Back for the new version of TRiBaL_Warrior AI !
First of all, i finally fixed the retreat system which makes a real difference in skill with the old version. My units are not wasted that easily anymore
Moreover, Daedric AI influenced me a lot in the many attacking rules i used here to complete the old ones (i copied some and adapted some others and didn't forget to specify him as the author when needed). I know its not great to copy the codes, and the more i'l feel comfortable with scrpting, the less i'l do it. But i'm stile rusty
NOW, Pocket and flank have each two strategies available :
- flank now will go for a huge feudal rush, massing archers/skirms with some defensive towers or will go for castle with one archery+towers.
Pocket will go for a fast castle Kts or will try to BOOM fast and prepare a strong imperial army.
Just tested some 4v4 Tribal vs Daedric and then Tribal vs Juggernaut. was not bad at all. It maybe shall start to build forward in case it got military superiority... and make bombard towers.
Plz test it and gimme feedbacks about this new version.
Now i want to make it work on other maps : Island, Nomad (copied already some codes like : (set-strategic-number sn-percent-attack-boats xxx) ready to be used !
Author of TRiBaL_Warrior, Yggdrasil, PharaoN, WoodyWood Pecker and currently dette
-
- Waheguru
- Posts: 5891
- Joined: Fri Jan 16, 2009 8:36 pm
Re: TRiBaL_Warriors (new AI)
Nice changes, it's very important to have specific pocket/flank strategies
-
- Waheguru
- Posts: 1905
- Joined: Sat Nov 08, 2003 9:22 am
Re: TRiBaL_Warriors (new AI)
I did a test game vs team 'evil kitties', 4v4 green arabia, all random civ.
Basically the game developed into two 2v2. One side lost very clearly, as the frankish Tribal Warrior chose to boom leaving his aztec partner alone in 1v2. Promi only managed to lure a few soldier and overall didn't do much damage, as Tribal Warrior towered up and managed to advance to castle far before Promi. However he didn't get bodkin arrow upgrades and this was fatal against Jugger's +2 armored Knights. Later the franks did a counter attack with cavaliers and and lots of spears, but lacked a lot of blacksmith techs, so Promi could actually stop them with castle age troops. This sort of strategy needs a bit more impact, as giving up a player is a serious deficit. Also I am not convinced if 4 TC into a castle drop is the best build order, I think the three TC boom that various AI do with meso civs is stronger.
The other side went very different as the spanish Tribal Warrior chose to KT rush. With both pocket players attacking the other flanks, which benefited the towered up korean tribal warrior way more than barbarian, team Tribal Warrior put up a way better fight here. Also the gothic Horde was focusing more on eco than the spanish and should have helped Barbarian more than they did. In the end both flanks ended up with heavily damage eco, Barbie a bit more than the Korans. However the korean tribal warrior was down to one lumber camp and did suffer from some bug as it stopped gathering wood (the camp was not so far away, so idk why they stopped). In the end before it come to a decision here, the other big deficit of the other side was coming into play when Jugger switched sides.
Even though Tribal Warrior lost in the end, it put up a good fight. That was better than I thought, a nice AI.
Basically the game developed into two 2v2. One side lost very clearly, as the frankish Tribal Warrior chose to boom leaving his aztec partner alone in 1v2. Promi only managed to lure a few soldier and overall didn't do much damage, as Tribal Warrior towered up and managed to advance to castle far before Promi. However he didn't get bodkin arrow upgrades and this was fatal against Jugger's +2 armored Knights. Later the franks did a counter attack with cavaliers and and lots of spears, but lacked a lot of blacksmith techs, so Promi could actually stop them with castle age troops. This sort of strategy needs a bit more impact, as giving up a player is a serious deficit. Also I am not convinced if 4 TC into a castle drop is the best build order, I think the three TC boom that various AI do with meso civs is stronger.
The other side went very different as the spanish Tribal Warrior chose to KT rush. With both pocket players attacking the other flanks, which benefited the towered up korean tribal warrior way more than barbarian, team Tribal Warrior put up a way better fight here. Also the gothic Horde was focusing more on eco than the spanish and should have helped Barbarian more than they did. In the end both flanks ended up with heavily damage eco, Barbie a bit more than the Korans. However the korean tribal warrior was down to one lumber camp and did suffer from some bug as it stopped gathering wood (the camp was not so far away, so idk why they stopped). In the end before it come to a decision here, the other big deficit of the other side was coming into play when Jugger switched sides.
Even though Tribal Warrior lost in the end, it put up a good fight. That was better than I thought, a nice AI.