forums.aiscripters.com

You should consider allowing it to play other civs.

It almost won a FFA where was Immortal in, if didn't got 2v1 in the end, with Binary and Immortal attacking at same time.

In theory some civs could also work similar, like Mayans with plumes, Cumans with cipchaks, Berbers with camel-archers, spanish with conquistadores.
But to be honest its not my goal, since I like to play mongols all in micro with ideally elite Mangudais. I'm sorry.
And it could use so much improvement, has many flaws, issues and limitations. I like rather to work on that, if even I have time for that.

I added a no-military-DUC version of MangudAI-all-in, which is intended to be used for Kimberly Azula's AI ladder for weaker AIs.
It is based on the 0.1.2 version, but I added esp some recent parts of the building-rules from version 0.3a, which uses up-build place-control.
There is a no-DUC-at-all version and a no-military-micro version, that uses FireBalls Sheep-claim, which I very much like.

For weak AI? Oh, it's for my ai thank u)

Version 0.3a, Military/Action.per
Line 47, rule too long.
Dang

Aynshtaynn wrote: ↑

Sun Nov 13, 2022 11:08 am

Version 0.3a, Military/Action.per
Line 47, rule too long.
Dang

Sorry, I test it sometimes on one, sometimes on the other version. Rule is split now.

wattle wrote: ↑

Sun Nov 13, 2022 4:00 pm

Aynshtaynn wrote: ↑

Sun Nov 13, 2022 11:08 am

Version 0.3a, Military/Action.per
Line 47, rule too long.
Dang

Sorry, I test it sometimes on one, sometimes on the other version. Rule is split now.

No worries, I was just excited, that's all. I still am though, let's see

Great unit micro, but just runs ahead with 1 Mangudai, causing it to be easily picked off before other units arrive. Very vulnerable against monks for the same reason, the AI just builds an army for the enemy because it doesn't bother to build big enough of a group to kill a monk before the monk converts the unit. Even 3 or 5 mangudais would be better than one.

Good early economy with good Castle age times, but I feel like aiming for 50 villagers only severely holds the AI back as the game goes on.

This is an AI that one could say "Man, good AI", and it can be even better.

Uploading version 0.3b beta for Arandis tournament. [edit: 0.3c is the successor to be used for PoA]
It's an old version from where I somewhat endet up I think in Nov last year. Not sure how much it differs from 0.3a, but the attack I think I changed to attack-now, since the TSA+patrol-attack did override Fireballs ranged-micro-code.
Checked vs HD AI (medium strong flush), if it gets back stone, it usually can get in the stone and drop the castle. If not it is doomed.

Unfortunately I can't see myself taking time for scripting in the foreseeable future.

PoA submit I guess