NTai XE10.1b - Page 2

NTai XE10.1b

Here is where ideas can be collected for the skirmish AI in development

Moderators: hoijui, Moderators

User avatar
AF
AI Developer
Posts: 20687
Joined: 14 Sep 2004, 11:32

Post by AF »

Download here

I doubt NTAI generates that many commands either, and if it's still a problem I'll push it down to 2/3
User avatar
AF
AI Developer
Posts: 20687
Joined: 14 Sep 2004, 11:32

Post by AF »

MUHAHAHHAHA

masse has just reported in a KAI vs NTAI game that thuds firing high trajectory now work.
mikedep333
Posts: 46
Joined: 25 Jul 2005, 20:07

Post by mikedep333 »

AF wrote:MUHAHAHHAHA

masse has just reported in a KAI vs NTAI game that thuds firing high trajectory now work.
Nice, which AI won the game?
User avatar
Masse
Damned Developer
Posts: 979
Joined: 15 Sep 2004, 18:56

Post by Masse »

well... KAI won
cause of the "odd" behaviour of NTAI's units
User avatar
Min3mat
Posts: 3455
Joined: 17 Nov 2004, 20:19

Post by Min3mat »

nice AI now alantai! ^-^ com dguns and it doesnt even stall that much!
User avatar
AF
AI Developer
Posts: 20687
Joined: 14 Sep 2004, 11:32

Post by AF »

NTAI doesnt go lvl 2 quickly enough to compete with KAI currently.

That and I suspect that something is amiss. Overrall I am not happy with 0.29 at all.
Chocapic
Posts: 556
Joined: 16 Oct 2005, 03:35

Post by Chocapic »

AF just a note for you considering your overall performance as a coder:
Make less releases and increase diference between them , cuz as they are ppl start to loose interest in dling thousands of zilions of ntai versions.
mikedep333
Posts: 46
Joined: 25 Jul 2005, 20:07

Post by mikedep333 »

Chocapic, are not his open source test releases a way of speeding up the development cycle as people report bugs and send patches?
Chocapic
Posts: 556
Joined: 16 Oct 2005, 03:35

Post by Chocapic »

mikedep333 wrote:Chocapic, are not his open source test releases a way of speeding up the development cycle as people report bugs and send patches?
That is true allthough the coder has to do some testing first, you dont see game companies releasing 1000 versions do you ? Its only supposed to be released when you see no bugs (after quite some testing).
User avatar
Argh
Posts: 10920
Joined: 21 Feb 2005, 03:38

Post by Argh »

I'd have to say that, while in an ideal world this software would get extensively playtested in a wide variety of configurations first... probably spamming out versions is about as good a way as any. It's not like we have to download every version, test it, and report back, after all- I usually catch every second/third version and give it a quick whirl.
submarine
AI Developer
Posts: 834
Joined: 31 Jan 2005, 20:04

Post by submarine »

Chocapic wrote: That is true allthough the coder has to do some testing first, you dont see game companies releasing 1000 versions do you ? Its only supposed to be released when you see no bugs (after quite some testing).
alantai clearly stated that these builds are test releases and no release candidates. therefore anybody who doesnt want to test buggy ai releases can wait for alantai's larger releases...
Chocapic
Posts: 556
Joined: 16 Oct 2005, 03:35

Post by Chocapic »

submarine wrote:
Chocapic wrote: That is true allthough the coder has to do some testing first, you dont see game companies releasing 1000 versions do you ? Its only supposed to be released when you see no bugs (after quite some testing).
alantai clearly stated that these builds are test releases and no release candidates. therefore anybody who doesnt want to test buggy ai releases can wait for alantai's larger releases...
I didnt say i had downloaded any though.
Anything that has too much releases is a pain in the ass for me.
User avatar
Argh
Posts: 10920
Joined: 21 Feb 2005, 03:38

Post by Argh »

If you haven't downloaded any, why are you complaining about a non-problem :roll:
User avatar
Argh
Posts: 10920
Joined: 21 Feb 2005, 03:38

Post by Argh »

NTAI 29a is not compatible with NanoBlob, and crashes immediately on the map Two Valleys. It also didn't handle my current hacked-up version of XTA :P

I was able to get my "hack mod" to work with SpringSP (btw, that's a neat interface... does it do anything else beside provide a better Battle Screen?) but I, like previous testers, saw sync errors fairly quickly in that made the game unplayable.

Very sorry to have to report that this version has so many problems, AF- before it started slowing down, it was clearly building some very clean, nice bases, sub-bases and was using a halfway-decent unit mix...
User avatar
Masse
Damned Developer
Posts: 979
Joined: 15 Sep 2004, 18:56

Post by Masse »

from AF: (he couldnt log for some odd reason)

@Argh, the 0.3 core works 100% fine with nanoblobz currently, I'll be releasing a test build 5 soon.

@chocapic and so on...: I've released 0.29, and there is a lot of differences betwene it and 0.28, aswell as a miriad of tweaks and fixes. However somethign is amiss, and I fixed it onyl to find a very dissatisfied result overrall that was a symptom of bad structure. So I finished the core build fo 0.3, and soon I'm porting forward the agents and releasing a special version of 0.3 early with basic universal mod support and the 0.29 agents and build tree, aswell as general build tre eimprovements ( destroying NTAI's utter dependence on kbots, and a basic water/hover branch with code to use it when necessary)

AF said: NTAI blobs generated 300 units in 4 minutes under nanoblobz 0.1

AF said: ask Argh for a copy of the ailog when playing his modified XTA
User avatar
AF
AI Developer
Posts: 20687
Joined: 14 Sep 2004, 11:32

Post by AF »

^^, that and technically this wasnt meant to be a test release, but rather just a general update of 0.28, fixing several bugs and adding a few extras.

M0ngus posted an error at darkstars.co.uk, I fixed ti btu I got annoyed and fustrated with 0.29's build system whcih still stalls on soem maps. So I forward proted the agents to 0.3, see test build 5. Test build 5 will be either 100% stable, or totally messed up. But ti works with nanoblobz, which is new, and it shouldnt stall either. However I've only forward ported the core commanders task tree, any other setup will revert to universal mod algorithms (still crappy).

A side line of this is a new advantage in that TA:K is pretty much fully supported now, with the exclusion of decent build algorithms.
Chocapic
Posts: 556
Joined: 16 Oct 2005, 03:35

Post by Chocapic »

@chocapic and so on...: I've released 0.29, and there is a lot of differences betwene it and 0.28, aswell as a miriad of tweaks and fixes. However somethign is amiss, and I fixed it onyl to find a very dissatisfied result overrall that was a symptom of bad structure. So I finished the core build fo 0.3, and soon I'm porting forward the agents and releasing a special version of 0.3 early with basic universal mod support and the 0.29 agents and build tree, aswell as general build tre eimprovements ( destroying NTAI's utter dependence on kbots, and a basic water/hover branch with code to use it when necessary)
you do realize that there were 10 releases/updates/bug fixes between .2 and .3 dont you ?
(im not counting the .29a nor .28b1 and b2 and maybe more i cant remember)
User avatar
AF
AI Developer
Posts: 20687
Joined: 14 Sep 2004, 11:32

Post by AF »

And they where spread over a long time.

Also if you look at the difference between 0.2 b1 and 0.29a, the leap betwene them is astounding. 0.2 didnt even attack, and had terrible scouting, infact I'm nto sure it even had that either. If I was making releases in the same way I was back then the current version nubmer would nto be 0.29a, it would be 0.45, but I didnt consider them big releases because I had plans and I had earmarked them for 0.3, so what if I made a truckload of little changes in the mean tiem and fixed outstanding bugs?

Do you want the best or do you want to wait a month ebfore I release and have a substandard AI to play with? If so, 0.23 should be around the corner....
Chocapic
Posts: 556
Joined: 16 Oct 2005, 03:35

Post by Chocapic »

Do you want the best or do you want to wait a month ebfore I release and have a substandard AI to play with? If so, 0.23 should be around the corner....
well to be honest i really prefer that method yes, but its just a matter of opinions.
(ive been developing my ai for sometime now and it can play "ok" already (considering the current ai's available), still im not releasing it until it does somethings that i want).
But yet again its really just an opinion.
User avatar
hrmph
Posts: 1054
Joined: 12 May 2005, 20:08

Post by hrmph »

Please continue with the continuous updates, it is much appreciated. Kudos, your AI is looking great!
Post Reply

Return to “AI”