Spring RTS Engine

Yesterday i finally managed to make KAI 100% crashfree on any map (got bored of testing it on metal maps only)
It doesnt play anywhere as good on normal maps but they didnt do too bad, especially considering i was drunk while coding it yday
We had many very long games to test the stability so it looks like its real business!

There were two very funny games, one where Kelson wanted to play 2 humans vs 6 KAIs with 100% handicap, which was so short i wont post any screenies, and a big massive one in anglosaxon starfish (4 kais vs 3 humans no handicap):

Heres a human placing mexes:


Heres KAI doing the same:


The two starfish arms where Sabutai's and Diamond's bases used to exist:


They did attempt a hilarious rescue mission to save Diamond's last con, with a goliath chasing after it, making hlts to block the shots while masse sent in a transport around the island. It didnt work but it was close!

By the 30th minute all 3 humans were cornered but putting up a stiff resistance thanks to masses ultraporcing techniques, which were getting neutroned all the time (wait till it makes nukes):


In my drunken rage i removed radar usage completely as it was buggy and ugly, wasting resources, however this prevented AI units from finding the human base (yeah i need to code scouting properly), and due to the also crappy building code, the game was running much too slow, so we had to call it quits so i could go to bed...


EDIT:
so then today i readded it along with more "agressive" building placement


Anyways thats the update. By the way, The metal class is now 2x faster, a bit smaller and more practical, ill post it up soon!

Looking at the mini maps, it's not hard to see which ones are AI's and which are human.

ahhh... soo much fun with the KAI... i want millions of rematches

how much handicap was there ?

zero

very interesting then

any release date? I want to watch this beat up the other AI.

2 months at the very least... so much work still to be done

^_^, that is very godo news from over here, that gives me ample time to finish my coursework and exams then go fulldrive on NTAI, and I've got truckloads of ideas I've had to stall on.
I only hope that veylona nd submarine can keep up once our exams are finished.

my exams will begin this february and my prediploma in theoretical physics will likely take place this april, i dont know how much time will be left to work on aai

For the record, Peace and I managed to break the AI, not get beaten by them :) Not to claim we would have won each time, but the more we played them the more we dominatedddddd! (although, as he said, it wasn't made a 'good' ai yet, just a stable ai)

really? i thought you lost all you base and all you had left was a comm wild gun running in the first 4 minutes, while peace survived but was already taking a 10-to-1 beating in economic power...
oh well i must be getting old and remembering things wrong

Now that it looks like it's stable... I will be happy to test it, if you want it tested thoroughly. Based on the screens, I am quite sure that I will be able to find chinks in its armor

Moreover, I doubt it can play anything but XTA... stable or not. Which is a problem, but hey... baby steps and all that.

Lost my base? I didn't bother building one that first time - I dgun'd the enemy's then started building up in my corner. Then you had to go fix something with the AI. I had a bunch of stuff up, including a near-done lvl 2 kbot lab, and a sexy lvl 3/4 comm :)

On another note, have some more feedback on your metal class whenever I find you on spring.

How mod-portable will this AI be?

does it support aircraft !!!!!

MAKE IT SUPPORT AIRCRAFT!

I added aircraft support, took a whole nine minutes, plus twelve mins of testing. It build its base up, scouted and attacked with air, but the results are not satisfactory, even with air/land combined and im not surprised since ive been over-using the crappy beehive & friends test code for all that. I removed air support shortly after, happy enough with the proof of concept.

Im working on the last "instant gratification" features (eg, small bits of code that show visible improvement straight away), before moving on to the "extremely delayed gratification" things, where ill have to code many thousands of lines, it will probably not work and ill only find out if it does once im done! If it works, then well, planes and ships will be added quickly. If it doesnt, well Ill quit making KAI

As for mod-portability, ill be using a new system which should please players and mod-makers alike (especially mod makers), but i wont reveal any details about it yet since its a feature described above and might not work well or at all.

*evil cackle*
they would be surprised to see this built in the 23rd minute but both players died beforehand

it looks very impressive , allthough i need a 1on1 match with it on a not metal map to take my conclusions out

Please consider releasling a pre-alpha test version. You would definitely not be the first person to release an incomplete AI :)