Immortal v0.10f

Upload your latest creations here.
Please note that by posting comments to posts in this forum you forfeit your personal copyright on whatever you post. AI Scripters reserves the right to re-use objective comments in script reviews without requiring the posting author's permission. Credit, where appropriate will be given in such cases.
Post Reply
TheMaximalBeing
Forumer
Posts: 97
Joined: Fri Jun 02, 2017 11:00 pm
Location: Australia

Re: Immortal v0.10e

Post by TheMaximalBeing »

CHANGELOG Version 0.10d/e
  • can play Romans, Armenians and Georgians
  • various things so that mule-carts are treated like regular camps
  • default strategy only chatted in AI games
  • can correctly handle unbalanced enemy/ally teams
  • can predict enemy locations properly with unbalanced teams
  • always allow spearmen training while being scrushed (not tested)
  • different resign threshold for humans and AI
  • 105 taunt detection
  • groups can attack gates and palisade walls if pathing is blocked
  • can train petards to blow up enemy walls (very rudimentary)
  • cancel MAA if archers are out; less likey to start maa if enemy has ranged units (not tested)
  • more likely to research heresy and faith when needed (not tested)
  • can return to dangerous dropsites slightly sooner
  • more builders for tc/castle when under attack
  • fewer monks when enemies researched faith/herasy (partially bugged on de)
  • chat strategy switches (only allies)
  • allow delayed defensive monks/mangonels
  • pocket archer players much more agressive when enemy still in feudal
  • fixed bug where builders break attack-ground + ballistics micro (not tested)
  • move-hide state remains closer to enemy tc for slower groups
  • enable attacking single monk/siege groups
  • always attack single groups if enemy has low pop
  • limit when straggler chopping is allowed
  • reassign straggler choppers to LC when too many
  • generalized vill hopping (to rapidly reassign shephards to any resource)
  • enabled arena wall deletiion
  • enabled wall deletion around camps
  • enabled wall deletion for boar lure
  • AOC version now supports all FE civs
Last edited by TheMaximalBeing on Tue Apr 30, 2024 9:12 am, edited 1 time in total.

TheMaximalBeing
Forumer
Posts: 97
Joined: Fri Jun 02, 2017 11:00 pm
Location: Australia

Re: Immortal v0.10e

Post by TheMaximalBeing »

WK version will not be released as I have gone over 10000 rules.

TheMaximalBeing
Forumer
Posts: 97
Joined: Fri Jun 02, 2017 11:00 pm
Location: Australia

Re: Immortal v0.10f

Post by TheMaximalBeing »

CHANGELOG Version 0.10f
  • archers will now dodge mangonel projectiles (very rudimentary)
  • will no longer attack towers that are protected by castles
  • will not attack towers with too many guard-towers + units nearby
  • enemy scouting smaller radius for arena (not tested)
  • more trade for arena
  • will add siege to defend forward castles/towers
  • limit to 3 TCs if forward enemy castles
  • research imp sooner if forward enemy castles
  • don't delay imp for more military if forward enemy castles
  • prefers monks to camels/spears if enemy is an age ahead
  • new function to check if we can path to enemy tc
  • build camps further from map centre on arena (rather than from enemy)
  • fixed missing blacksmith for certain BOs
  • fixed missing market for smush BO
  • vills will attack unprotected siege units
  • vills will no longer retreat from trebuchets
  • vills will no longer over-estimate the strength of siege units
  • no longer considers castles when finding enemy players
  • groups will attack walls that are attacked by other groups
  • prevented groups from using forward enemy buildings for enemyPos when TC not found
  • groups can attack closer to enemy TC on arena (not tested)
  • can dive under tc to attack monks/siege (not tested)
  • less likely for groups to walk under enemy forward buildings
  • archers will not delay attack vs siege, monks, scorpions (not tested)
  • archers prioritize siege, monks, and repairers more
  • prevent new units running into forward enemy castles
AOC/WK versions not released as I have run out of rules

TheMaximalBeing
Forumer
Posts: 97
Joined: Fri Jun 02, 2017 11:00 pm
Location: Australia

Re: Immortal v0.10c

Post by TheMaximalBeing »

Bandit1990 wrote:
Thu Feb 22, 2024 12:24 pm
TheMaximalBeing wrote:
Sun Feb 11, 2024 1:52 am
Here's an experimental version of Immortal with the latest civs.
There were many changes that were necessary to make the mule carts work.
I have not had time to test it properly.
Woop! Great to see it playing new civs :)

One thing I've noticed on Arena with this version is that Immortal deletes its walls really early - like mid castle-age. I've done tests on v0.10c on Arena before and I don't recall them going that soon. In these scenarios Immortal is actually losing so deleting the walls ends up throwing the game. Is that something different or has it always done that?
Walls should be deleted much later in the most recent versions. At some point I would like to get rid of it but they are causing too many pathing issues atm :head

TheMaximalBeing
Forumer
Posts: 97
Joined: Fri Jun 02, 2017 11:00 pm
Location: Australia

Re: Immortal v0.10c

Post by TheMaximalBeing »

CnelKurtz wrote:
Tue Feb 13, 2024 12:48 pm
Hi!

I'm new here. After Viper's showcase of the last AI tournament I downloaded it and I've been playing it. Since then I have played about 50 games against it. First of all let me say this is fantastic work TheMaximalBeign! Congratulations and thank you, I'm having a great time facing Immortal.


I'll share some of my notes on it in case they're helpful for you.

Thanks for the feedback :D

1. I'm still not sure why it's ignoring wolves. I've tried to fix it multiple times. It seems to work fine in the scenario editor.

2. We (AI scripters) don't have the necessary commands to check for walled in units. We would need more reliable pathing commands that can check if a path to a unit is available.

3. Should be more likely to attack walls around a walled-in opponent in the recent versions. It may be a bit buggy sometimes (due to the limited + buggy pathing commands we have available).

4. boar laming was disabled for now. It's super easy to counter just by luring your boars early. But I might re-activate it vs. human opponents.

5. there are actually two militia strats: maa or drush. The timing is different for each. It can also go straight archers or skirms. Try byzantines, mayans or britons to make this more likely. If it detects a close opponent then it will likely switch to a drush strat.

6. It should be better at avoiding excess militia + MAA upgrade when not needed in the latest versions.

7. It doesn't really know how to counter a siege push at this stage. Could take a while before I get around to this.

8. Once it starts building towers it will usually fall behind in army numbers so it will likely end up needing more to defend itself.
Last edited by TheMaximalBeing on Fri May 10, 2024 10:04 am, edited 2 times in total.

TheMaximalBeing
Forumer
Posts: 97
Joined: Fri Jun 02, 2017 11:00 pm
Location: Australia

Re: Immortal v0.10c

Post by TheMaximalBeing »

clc8901 wrote:
Sun Feb 11, 2024 10:55 am
Tested Arms, Romans, and Georgs on Arabia and Arena. They worked well without any major bugs (so far). I'm guessing the monk gameplay is still being worked on, noticed Arms didn't build a church on Arena. Also, this AI highlights how silly the Monaspa are.
That's interesting. I never tested the Georgians properly but went for Monaspa just based on the stats.

venomus
Forumer
Posts: 61
Joined: Sun Sep 05, 2021 10:26 am

Re: Immortal v0.10f

Post by venomus »

Thankful that there is a new Immortal update so soon. Every day I play my games with this AI, literally hundreds of hours with it. :dance

paper
Forumer
Posts: 18
Joined: Thu Dec 19, 2019 8:17 am

Re: Immortal v0.10f

Post by paper »

In post-imperial deathmatch, Immortal v0.10f always builds trebuchets with escrow, even with little to no military. These trebuchets get in the way of the other units in the queue

Immortal v0.10e doesn't have this problem

User avatar
Aleph
Yogi
Posts: 488
Joined: Sat Jan 17, 2015 10:07 am
Location: Germany

Re: Immortal v0.10f

Post by Aleph »

thx for new :)

Post Reply