new hero skills build

hi!

i've edited the way AMAI chooses skill builds a little bit, now there are three possible skill builds for each 1st, 2nd and 3rd hero chosen by AMAI. skills are chosen by 60:30:10% possibility, which leaves possibility of identical choices, in case there should be, for example 90:10% probability of choosing one build over anther. i've also tweaked a bit probability to choses certain heroes.

this probably doesn't improve AMAI, but makes it slightly less predictable. i meant to test it more extensively, but had to hurry, since 2.51 is on the way...

anyway, here it is, only tft version this time.

the change is (in common.eai):

"function LoadSkills takes integer sa, integer hn, integer ta returns nothing
local integer perc = GetRandomInt(0,100)
local integer sk = 0

if perc > 60 then
set sk = sk + 1
endif

if perc > 90 then
set sk = sk + 1
endif

call LoadSkillsExt((sa-1)*HERO_LEVEL_NUMBER + sk*10 + (hn-1)*9*HERO_LEVEL_NUMBER)
call StoreSkills(ta)

endfunction"

i've also changed racial heroes.txt, heroskill.txt and neutralheroskill.txt (included in data.zip). 3rd file is xls.zip, skill tables are given there, along with some comments (regarding what is changed).

feel free to experiment and comment, have fun

No love for Alchemist? :(

And the skillset for Alchemist is bad. You should never give the A.I any opportunity to learn Healing Spray because the A.I will use them on the enemies .

Actually, Alchemist is one of the very few heroes who can go toe to toe vs a Demon Hunter of Blade Master earlier in the game and end up coming on top. Alchemist has the highest attack damage among heroes when not counting skills (no critical strikes) and is a great combo when paired with Blademaster. Focus fire with those two + armor debuff from Alchemist' Acid Bomb and any low level Int heroes dies very quickly.

thank you for time taken!

it's still under testing, i change values all the time, so it's far from finished. i think i've set to high overall values for the 1st tavern hero, (for human and orc increased from 5-6% to 21-22%) and that i should lower them to be closer to strategy masters.

regarding alchemist's healing spray, since you mentioned this before, i've payed special attention to him when he was present in game, and never actually saw him using spray except outside the battle. if you say it happens, i believe you, maybe he shouldn't have access to that skill

Well i've finally had chance to analyse your numbers. Some ideas i have implemented into the first hero bonuses etc and other missed out information.

The base prorities i am not changing. They are actually taken from the battle.net stats of hero share on all realms, so in actual fact very accurate to the same as you would expect to show up at battle.net. I need to update them though as that was a while ago and looks like realm stats have changed a little. Biggest change seems to be the orc hero stats with undead having the least change to the base stats.

thank you for the time taken!

my test system is being modified heavily, and i also used battle.net statistic in later testing stages (if you, by chance, still need them categorized in xls table, i have it)