=◬= Illuminati AI =◬=

[scripter64]

First of all, that's not a triangle, it's a cyclops pyramid. Second, it's not , it's ◬, or maybe it's ))).

)))

[Aleph]

; changelog 0.5d -> 0.6
; - Illuminati now plays Celts and Goths!
; - new Celt strategies: Scorps + Halbs, Scorps + Woads
; - new Goth strategies: Huskarls, Champflood, Longswords
; - new universal strategy: pocket Krushboom
; - new 2-part semi-dynamic gatherer percentage system
; - will now build more lumber camps, especially in the late game
; - will now buy food/wood with excess gold in 1v1, too
; - will now gather troops under castle if it has one (instead of TC)
; - will now try to counter eagles with champs
; - will no longer retreat from a tower if it has enough infantry and superior numbers/age
; - will no longer smush vs. FCing aztecs/mayans
; - will now stop skirm production when saving food for castle-up
; - will now escrow imp-up
; - can now train skirms in castle age, too
; - Aztecs can now train jaguar men
; - can now also build a tower on its second mining camp (e.g. when castle dropping)
; - pocket can now detect its flank player, and (if the flank is also Illuminati) the enemy flank (otherwise it just detects closest enemy)
; - pocket krush will help its flank (still needs more work)
; - capped spearman production in feudal war
; - fixed mangonel micro
; - fixed some issues with town center placement (no space / too far)
; - fixed a bug that could cause pocket to get stuck in dark age
; - fixed a bug in tower placement vs human opponents
; - fixed a bug that could delay boar relure
; - fixed a drush detection bug
; - fixed a bug that could cause buildings to not be placed
; - fixed a bug that caused the tower to be built prematurely
; - fixed a bug that caused the university to be built prematurely
; - fixed a bug that caused the monastery to be built prematurely
; - fixed a bug that stopped archer production in feudal age
; - fixed a boar steal bug where the scout would only hit the boar once
; - many minor tweaks and fixes
; - added goals 11 (under attack) and 392 (attacking) for better AI cooperation

The TG stuff still isn't very good, but the new civs are great in 1v1.

[Felipe_EL_Guapo]

Hi Aleph, just a quick feedback. I haven't had the pleasure to see goths and celts yet, rather I run some games with Mayas. Very good job improving the feudal. I liked a lot that when luring, 4 hunters my attack deer, but then they redistribute to dead deer (Cheesontoast must see this). It is much more competitive now against flushers. Great tower destruction code involving even villagers ahahah.

I wanted to see how have the klew been evolving, so I run it on GA to increase the probabilities. I could see some issues here. It tend to go for exploring the enemy instead of luring deer (maybe they were too far). A big issue I think it's mining stone prematurely when Klewing (at the same time when is moment to get gold). I think an eco going up to castle with only 20 pop cannot afford two mining camps and two out of 19 villagers mining stone). Actually, onces getting castle it barely have wood for barracks. Secondly, this villagers are susceptible to be attacked, so is an extra problem for illu. Thirdly, a saw a game were It went first for stone, then it didn't have wood for gold mining. Fourthly, if it successfully reach castle, it still wont have wood for a tc, and if it very very lucky and got it, then it will be still running out of food to keep creating EW for two barracks and vils for two tc. In short, it was more effective without mining stone... till got an advantage in castle and stable economy. Finally, sometimes too many guys in berries, so now is slower than before. 100% when Klewing gets the 800f much faster than the wood for the buildings, that is way the Klew is no longer close to 13min (probably the stone camp is what has the greatest impact though).

[Aleph]

Thanks Felipe!

I wanted to see how have the klew been evolving,

I haven't changed the klew since 0.5. In fact I haven't tested the Mayans at all lately - I probably should have.

It tend to go for exploring the enemy instead of luring deer

If it finds all resources very quickly, with meso civs it will try for a boar steal if the deer are not found or far away. In any case, you can't klew without deer.

mining stone prematurely when Klewing

OOOOOh i completely forgot about the klew when writing the stone mining for tc rule. At least it's easy to fix.

it still wont have wood for a tc

It's a klew! It shouldn't be building TCs, it should delete its TC!

100% when Klewing gets the 800f much faster than the wood for the buildings, that is way the Klew is no longer close to 13min

Interesting, I'll need to test it some more. To be honest the klew was never really close to 13minutes except in that one rec of yours. Consistently doing a 14 minute klew is good enough. But of course if I can make it faster, I will try to.

[Aleph]

; changelog 0.6 -> 0.7
; - Illuminati now plays Britons, Japanese and Vikings!
; - dynamic gatherer percentage system now makes use of cost data
; - completely redid escrow system
; - can now also escrow important buildings and units
; - added more researches to be escrowed
; - new strategy: knights + xbows
; - will now force flush build order in many situations
; - will now avoid enemy tower or attempt to destroy it, depending on the circumstances
; - will now change strat from pure xbows if countered too hard
; - will now also train a few spearmen if feudal enemy has a stable (but no scouts sighted yet)
; - will no longer make rams when at a disadvantage
; - will now also make mangonels in imp
; - will now retreat to castle (if it has one) instead of town-center when stopping attack
; - Goths will no longer use bombard cannons as main siege weapon (because they don't target buildings)
; - Goths and Japanese will now use trebuchets
; - improved defence behaviour
; - improved superiority evaluation to take more factors into consideration
; - improved boom
; - improved market placement for TG trade
; - improved trash war switch
; - improved some unit recruitment switches
; - improved some build orders and eco balances
; - improved garrisoning conditions
; - added a few more SNs
; - added "defensive" (TSA targeting) priorities
; - added a timer to boar kill to avoid lurer repeatedly turning around
; - tower placement now also takes the 2nd lumber camp into consideration (if there is one)
; - removed men-at-arms from mayan flush
; - increased camp distance on gold rush, especially on larger map sizes
; - fixed mangonels continuing to attack ground after no more enemies in range
; - fixed a critical bug that caused stone mining when klewing
; - fixed a bug with mayan imp strat selection
; - fixed a bug that caused sn-percent-building-cancellation to not be lowered in celtic castle drop
; - fixed a strat switch loop bug with goths
; - fixed sn-enemy-sighted-response-distance sometimes being too low
; - many minor tweaks

[scripter64]

Epic fix list
Thanks, Aleph!

[Aleph]

; changelog 0.7 -> 0.7a
; - Illuminati now plays Huns!
; - new strat: Cav Archers
; - improved spear and skirm training rules
; - improved duc attack when circumnavigating tower
; - improved retreat
; - will now also train scouts in trash war
; - will now also train champs or infantry UU in trash war
; - will now sell stone when needing gold for trash war upgrades
; - will now rescout target enemy as soon as they reach feudal age
; - will now resume search for enemy after successful boar steal if they aren't found yet
; - added research rules for units that Illuminati never used before (such as cavalry)
; - included scorpions and mangonels in forced retreat
; - fixed a build order bug in pocket krush
; - fixed miners mining far away resources (in early game)
; - fixed multiple escrow bugs
; - minor tweaks

[Felipe_EL_Guapo]

I tested version 7 with Mayas against Tron (1 Mayan by accident and 2 Mongols). Happy to see that the Klew is back!!! It is very interesting the use of the flare to avoid towers, attacking in the weak points. Unfortunately, Tron played even better in those games in GA (didn't know that Tron could perform so well with Mayas doing FC PA). The avoiding tower system worked very good in feudal, but then in castle I couldn't see it anymore and Illu kept attacking by the front. I will test this new version later.

[el+noob]

Aztec being with pocket takes 19 or 20 minutes castles, before the 15 or 16 to

I also think that way too hurries to attract deer, sometimes fails to find the two boars and all the sheep.

[Aleph]

I tested version 7 with Mayas against Tron (1 Mayan by accident and 2 Mongols). Happy to see that the Klew is back!!! It is very interesting the use of the flare to avoid towers, attacking in the weak points. Unfortunately, Tron played even better in those games in GA (didn't know that Tron could perform so well with Mayas doing FC PA). The avoiding tower system worked very good in feudal, but then in castle I couldn't see it anymore and Illu kept attacking by the front. I will test this new version later.

Yeah, TRON is pretty good with the Mayans. The flare was just there for debug purposes, I took it out in 0.7a - but the feature itself is still there, and improved.

Aztec being with pocket takes 19 or 20 minutes castles, before the 15 or 16 to

Do you have the rec for that? It's probably one of those TG bugs that I hardly ever find because I rarely test TGs.

I also think that way too hurries to attract deer, sometimes fails to find the two boars and all the sheep.

It will only start luring deer after having found both boars and at least 6 sheep, or after 6/7 minutes.