forums.aiscripters.com

It's just a bug. Lie! There are some bugs.

Some of the versions with my Cruzada have an horrible eco balancement. It's the main reason because I decided to develop Marathon's idea. And I suspect that my economical sns are not neither the best.

The result is that Cruzada has an all around better and more violent flush than Crusade (although with a smaller diversity of strategies) but a much worse castle and imperial ages. Better said, horrible castle and imperial ages.

But I think that I'll can test my new eco balancement next weekend and perhaps upload it.

Hey there,

I've just played a couple games with Cruzada to check it out and see if i could see any things that might help you so:

The second lumber camp seems to be dropped rather far away, what in my tests resulted them to be forwarded also.

First mining camp got dropped on middle gold on gold rush right away (this was rather close though im guessing about 20-25 tiles. still was unprotectable for your script though

Cruzada seems to go up to feudal-age as soon it got the 500 food? In the 3 tests i ran it once went with 24 pop, once 25 pop once 26 pop. And in all cases it didnt really seem to deal with everything proper. Only having wood for a single range at first not having enough gold to get archers going right away and such sort of things. In one of them (i believe it was the 26 pop up) it started making a archery-range with 4 villagers, im guessing because it thought it was quite late.

Anyhow i think it would be more effective to have a serten number of villagers to up at, at wich you know for sure it will do what it has to do.

In the one game you manged to reach castle-age your eagles + skirms (double counter on Daedric's Xbows/skirms) You manged to destroy Daedric though, i should really add counters to eagles haha.

Good luck! Hope this can help you some :3

Hi Armelon:

Camps. You are right, Cruzada sometimes builds them too far. I am using 1.c system to place them even in low resources maps, although I suspect taht with UP there are tools to place them in wise sites either in good or bad maps.

Feudal jump. There is a requisite, 24 villagers (this number will depend of the selected strategy) or 1000 food. I had it higher, but there are AIs which are too agressive to allow a slow feudal time and after the economy improvements that I've taken from The Horde and other AIs Cruzada is often able to build an instant archery and with fast gold incoming to don't delay at least the second archer. Keep in mind that if my enemy hasn't started an army Cruzada can attack even with just 3 soldiers (initial scout and extra two men). But I'll keep attention anyway in my next tests to see if Cruzada really needs to boost again the minimal number of dark villagers.

Eagles/skirms. I'ts natural. For the moment I have a single strategy for Mayans, which makes them to train only xbows in castle age as a default plan. But my big scripts always have a counter system because as player I like to face AIs which adapt their armies to what I am using. And Cruzada usually prepares eagles vs skirms and change from xbows to skirms if enemy uses a some archery (including skirms) unless enemy uses too counters to skirms, in such case it will forget xbows and skirms. Another example are huskarls, which convince Cruzada to uses a combination of eagles (or m@a), scorpions and monks.


P.D: Ty for the feedback.

New update, but bugged.

I've completed the new dynamic gatherer calculator, but it sometimes fails in the castle age. I suspect that the reason are the rams, they cost too much woodcutters, so the minimal number of villagers is often higher that I can have. Perhaps is another bug that I haven't found.

The system consists in calculate the number of villagers in every resource considering the selected units to train, their costs, time training, number of buildings, possible bonuses which affect these stats, etc. I've used values in the excel file for it.
It later reduce them depending of the completed economical upgrades, adds extra woodies for farms, reduce gold miners for relics/trade units, etc.

And in the Mayan Flush, Mayan Crossbow and Mayan Arbalest files I have rules to send the extra villagers (if there are)depending of the age phase.


If the problem is with my rams, a possible solution could be to leave them just for the case in which I have an "excedent" of villagers for the extra-percentage-xxx ones, and similar with onagers (they are just a complement for other antiarcher units) or perhaps castle units (all of them with a really fast training time, although always with a small number of buildings).

Any feedback is apreciated. I'm sure that my idea is good, although there are things to change.

A small suggestion is to use like a 5% escrow on wood to maybe save up for them? (I don't know if you're using escrow already or not).

(IIRC rams are 175 wood)
If the AI gathered 500 wood/minute, at ~%11-12 wood escrow, you could get the cost of a ram in three minutes. Unless my math is off.

My AI doesn't use escrow, although perhaps I'll do something similar to what you say: stop the production of a few wood and gold units to accelerate rams if really needed.

But my problem is not that Cruzada can't train rams. Cruzada does it even too easily. The problem is that when my AI trains rams it often sends every villagers into wood and gold, keeping 3-4 farmers, or even none of them. My probable solution is to convert rams and other siege units in an "extra", to don't use my basic villagers, just for when I have an excess. This solution could require your suggested escrow or to stop wood/gold units unless Cruzada have enough resources.

Well, I have now an idea to implement. The problem is that I have a very important exam next June-July, so my progresses will be much slower until then. But it doesn't matter, I've started a new AI this winter, I want to finish it in a future and everything goes right I'll try to make Cruzada in something as dangerous and more complete than Crusade a couple of years ago.

Cruzada updated.

I've resolved the bug which ruined my economy when castle or siege units requested.

UUs and siege are now considered by my eco balance system to calculate extra gatherers, not for minimal ones. This system is less accurate, but erases the risk to colapse my economy if there are not enough villagers to keep all my buildings spamming units.

Perhaps it could be improved adding escrows.


The other thing to improve is my obsolete attack system. Perhaps I learnt to do it in the future.

Thanks for the update, Campeador!
It's great to see a new Cruzada

Hello,
I tried this AI and displayed ERR5001: File open failed. File Extra Units stuff Caste and Sieget.per. Although still works.

Oh, it seems there's an extra load at the bottom of the main file. You can fix it by opening Notepad, drag "Cruzada Maya.per" onto Notepad to open it, go to the very bottom and add a ";" in front of (load "Extra units stuff Caste and Sieget"), so it's like ;(load "Extra units stuff Caste and Sieget"). After that, please save and see if it works.