@krogothe:
That would meen we would have to bug betalord and zaphod about this. It would be much much easier to do it as Alantai suggests it. I wanted to use a harmless AI as a test run for his idea. There would be no harm done and we would be abel to see how things work out. When devs have the time and interest then we can have more advanced method of chosing and controlling groupAIs.
@Alantai:
I ment a formation AI that kept the units in a fixed formation even when moving with the artillery at the back.
Group AI and Global AI
Moderators: hoijui, Moderators
about what Uber ai are you talking about krogothe?
there is no such thing.
having the ability do disable the few ais that exits rigth now will lead the ignorants to disable them without even knowing what they do, thinking they are harmfull.
and about the proof of concept, well if you cant proove what you are telling about how harmfull they are, it seems to me they are not such.
there is no such thing.
having the ability do disable the few ais that exits rigth now will lead the ignorants to disable them without even knowing what they do, thinking they are harmfull.
and about the proof of concept, well if you cant proove what you are telling about how harmfull they are, it seems to me they are not such.
i give up, think what you want, make groupAIs if you want, believe me if you want...
If i see what i call uber-groupAIs out ill counter by simply making a better, all encompassing one (a modified KAI) and not releasing it, which should raise my winning % quite significantly (i believe enough to piss off everyone i play with), afterall it will combine the human thinking and strategy making ability with the AIs precise calculation and near infinite microing ability. Of course you guys dont believe this is possible and think AIs are retarded in all apects but hey why bother? Im wrong arent I?
i keep getting sucked into these pointless arguments where even if i win theres no reward in it, i gotta tell myself every day to stop doing it...
Release all groupAIs if you want, make TA realistic, think that one day there wont be poor people anymore i cannot be arsed arguiing anymore...
In fact, dont even dignify this post with a reply, i agree with whatever you guys say...
If i see what i call uber-groupAIs out ill counter by simply making a better, all encompassing one (a modified KAI) and not releasing it, which should raise my winning % quite significantly (i believe enough to piss off everyone i play with), afterall it will combine the human thinking and strategy making ability with the AIs precise calculation and near infinite microing ability. Of course you guys dont believe this is possible and think AIs are retarded in all apects but hey why bother? Im wrong arent I?
i keep getting sucked into these pointless arguments where even if i win theres no reward in it, i gotta tell myself every day to stop doing it...
Release all groupAIs if you want, make TA realistic, think that one day there wont be poor people anymore i cannot be arsed arguiing anymore...
In fact, dont even dignify this post with a reply, i agree with whatever you guys say...
well i think you are giving up too fast krogothe.
why dont you try to put up some "uber ai" toguether and try it against a good player.
Im deffinetly sure you will be surprised, and will try to "take over" the controll you gave to that ai over your units.
Is what happens when you play shared commander.
there are thousand of decisions an ai cant make at the moment.
why dont you try to put up some "uber ai" toguether and try it against a good player.
Im deffinetly sure you will be surprised, and will try to "take over" the controll you gave to that ai over your units.
Is what happens when you play shared commander.
there are thousand of decisions an ai cant make at the moment.
What we need, is to know what are this "monsters" really capable of before banning them ony bc we fear they could own us.
And dissabling group ais, will lead to even slower adoption of this really cool technology.
As far as i can recall, there is not even one replay boasting central build.
That is because, even using them requires some skill.
And dissabling group ais, will lead to even slower adoption of this really cool technology.
As far as i can recall, there is not even one replay boasting central build.
That is because, even using them requires some skill.
I don't think this is about banning group AI's, mearly giving players the ability to enforce fair play in a game that is not supposed to use them. I have a feeling that most games would have them enabled, and only disabled in seriously compeditive games between people who want to know who the better player is.
thank you! you understand! (feels great to have braincells huh?)ILMTitan wrote:I don't think this is about banning group AI's, mearly giving players the ability to enforce fair play in a game that is not supposed to use them. I have a feeling that most games would have them enabled, and only disabled in seriously compeditive games between people who want to know who the better player is.
So what? At least they are happy and they had a choice in the matter!Decimator wrote:I disagree, Titan. Implementing controls would likely have the opposite result. It would be disabled as a demon in all the nub games, and it would be enabled in the good players' games because they know better.
I KNOW Im right about the UberAIs, but i now know that the community will be safe because all the crazy people who want a choice-less world of uber-AIs actually cant seem to make those AIs themselves, because they are too lazy to learn c++ themselves, because if they really wanted them instead of just chit-chatting on the forum they would learn it and get some work done. This will give time for the devs to implement server-side control and once thats in place then i couldnt care less about this subject...
I have all the evidence i need to support my UberAI thesis, but you guys (mongus anyways) are always right remember? You never made an AI (and if you did it must have sucked major balls) and you think you know all about them right?, call me a liar if you want :)
As long as people are too damn lazy to make something happen instead of just talking theory and showing a complete ignorance to the power AIs can have then I can concentrate on making KAI. without worries. At all.
Thats it, im catching my plane in 20 mins, see you guys in 2006...
i wont aswer you krogothe till you come back and have the time to read the responses and dont run away posting just insults and degradating the persons that think different than you, or dont have your abilities.
Somehow ILMTitan, you are totally rigth and i sound like a fanatic. ( and im, of this game
)
The thing is, and i have asked this without answer:
Is there currently any Group AI that plays the game instead of the player?
and, are current group ais harmfull?
if you can answer that, i think its totally related to the need of disabling them.
do you use them?
e: about the competitive games, .. im not trying to sound all knowing, but you know how to use the default group ais? you think they dont involve skill? you think the use of them makes you a worst player?
e2: the "uber ais" are only a myth at this point, and are being confused with "player ais".
Somehow ILMTitan, you are totally rigth and i sound like a fanatic. ( and im, of this game
)The thing is, and i have asked this without answer:
Is there currently any Group AI that plays the game instead of the player?
and, are current group ais harmfull?
if you can answer that, i think its totally related to the need of disabling them.
do you use them?
e: about the competitive games, .. im not trying to sound all knowing, but you know how to use the default group ais? you think they dont involve skill? you think the use of them makes you a worst player?
e2: the "uber ais" are only a myth at this point, and are being confused with "player ais".
There are some "encompassing-Ais" right now, like NTAI, JCAI and the like, that are supposed to put up a challange for the single player.
How come they dont win everytime, although much work has been put into them? How come many players regard AI-Opponents as no challenge?
Thats because our AI is not a real AI according to scientific standards but merley a very limited set of scripts.
I think Krog fears a econ/micro-managemant script that handles the dull and time-consuming tasks for the player, while he can concentrate on fighting.
But i can tell you for sure that even such a script will have weaknesses, to be exploited by a human player.
There is no über-ai right now, and there wont be in the near future, and i should know, scince i studied computer-science.
How come they dont win everytime, although much work has been put into them? How come many players regard AI-Opponents as no challenge?
Thats because our AI is not a real AI according to scientific standards but merley a very limited set of scripts.
I think Krog fears a econ/micro-managemant script that handles the dull and time-consuming tasks for the player, while he can concentrate on fighting.
But i can tell you for sure that even such a script will have weaknesses, to be exploited by a human player.
There is no über-ai right now, and there wont be in the near future, and i should know, scince i studied computer-science.
Our all encompassing AI's do not put up a challenge for one or all fo the following reasons:
There is too large a scope and we havent had the time to make our AI's as robust as they need to be yet
We've focused on getting a working AI that runs 100% stable rather than a challenging AI
We dont get user feedback unless it's because they've had a crash or they're just saying it's good or bad (no suggestions? AI should build this instead or more of thsoe or it should build this at the start or it should do that rather than those or it'd be better ifs......)
People dont tell us when problems occur (100+ people downloaded the original 0.28.10 from File Universe even though it was guaranteed to crash, yet only 2 people reported a crash)
We as AI develoepr may have other things todo (I could be working on NTAI right now but instead I'm revising for exams).
Lack fo engine information (I cant tell fi a wreckage can be resurrected or not, I ahve to give an area command and hope the unti resurrects something if anything and in doing so risk that unit stallign and becomign non-productive)
Engine problems. (Economy in AI's can be pushed back quickly if con untis stall which they do, and thus the untis become a waste of resources as the full potential of the AI is never met. And this is alll because the engien fails to make a single interface call when a unit is idle 30% of the time.)
Inexperience (NTAI was amde to work best on XTA Core, but only because that was all I knew, I knew nothign about how to play AA, and nobody told me or helepd em even though I asked continuosly for months, I didnt even get any help with Arm, I ahd to copy the core buidl tree and replace COR with ARM at the unit name prefixes then make a few tweaks os it didnt stall and hope for the best, and ontop fo that nobody pointed out glaring faults I'd amde in the buidl trees, or made suggestions. NTAI could ahve been amde 10x as challenging if an experienced player looked at the build tree for a mintue ro two and wrote out a new one, just as a list of unitnames, or expained where I went wrong. Instead everyone thinks I'm asking them to learn C++ and recode the entire construction system They where onyl lists of untinames not complex calculations!).
AI developer strategy faults (
AAI and submarine falling into base mentality
Me not knowing any strategies and having little or no experience playing online
Krogothes and KAI's reliance on all metal maps for a good game
JCAI's non expansionist game policy
OTAI's disregard for a units value or enemy threats
etc.....)
Most of the above would be eliminated if we had a little help, even if it's just a suggestion for a minor change or a comment that isnt just "I had a good game with this AI"
100+ people all downlaoded and getting a crash yet only 2 people actually reporting that a problem existed at all? If those 2 hadnt said ought I'd have left that there and it'd still be there now and nobody would have said anything.
But really feedback is much appreciated, not just the rampant flood of "good game"/"crash", or the "I want this feature", just suggestions for changes,(not additions changes/tweaks/editations/minor improvements etc).
There is too large a scope and we havent had the time to make our AI's as robust as they need to be yet
We've focused on getting a working AI that runs 100% stable rather than a challenging AI
We dont get user feedback unless it's because they've had a crash or they're just saying it's good or bad (no suggestions? AI should build this instead or more of thsoe or it should build this at the start or it should do that rather than those or it'd be better ifs......)
People dont tell us when problems occur (100+ people downloaded the original 0.28.10 from File Universe even though it was guaranteed to crash, yet only 2 people reported a crash)
We as AI develoepr may have other things todo (I could be working on NTAI right now but instead I'm revising for exams).
Lack fo engine information (I cant tell fi a wreckage can be resurrected or not, I ahve to give an area command and hope the unti resurrects something if anything and in doing so risk that unit stallign and becomign non-productive)
Engine problems. (Economy in AI's can be pushed back quickly if con untis stall which they do, and thus the untis become a waste of resources as the full potential of the AI is never met. And this is alll because the engien fails to make a single interface call when a unit is idle 30% of the time.)
Inexperience (NTAI was amde to work best on XTA Core, but only because that was all I knew, I knew nothign about how to play AA, and nobody told me or helepd em even though I asked continuosly for months, I didnt even get any help with Arm, I ahd to copy the core buidl tree and replace COR with ARM at the unit name prefixes then make a few tweaks os it didnt stall and hope for the best, and ontop fo that nobody pointed out glaring faults I'd amde in the buidl trees, or made suggestions. NTAI could ahve been amde 10x as challenging if an experienced player looked at the build tree for a mintue ro two and wrote out a new one, just as a list of unitnames, or expained where I went wrong. Instead everyone thinks I'm asking them to learn C++ and recode the entire construction system They where onyl lists of untinames not complex calculations!).
AI developer strategy faults (
AAI and submarine falling into base mentality
Me not knowing any strategies and having little or no experience playing online
Krogothes and KAI's reliance on all metal maps for a good game
JCAI's non expansionist game policy
OTAI's disregard for a units value or enemy threats
etc.....)
Most of the above would be eliminated if we had a little help, even if it's just a suggestion for a minor change or a comment that isnt just "I had a good game with this AI"
100+ people all downlaoded and getting a crash yet only 2 people actually reporting that a problem existed at all? If those 2 hadnt said ought I'd have left that there and it'd still be there now and nobody would have said anything.
We said it was a skirmish AI not an attempt at sentient AI. ^_^.Thats because our AI is not a real AI according to scientific standards but merley a very limited set of scripts.
But really feedback is much appreciated, not just the rampant flood of "good game"/"crash", or the "I want this feature", just suggestions for changes,(not additions changes/tweaks/editations/minor improvements etc).
Shame that you wont be reading this Krogothe :/. First thing first: This has been discussed in the past, the end all discussion was here (but you wont want to be reading it 
): http://taspring.clan-sy.com/phpbb/viewt ... ro&start=0
Second: What almost everybody agreed on then that was that the most sensible thing to do is make Group AIs on/off able by the host.
The earlier quote of SJ was pretty correct: "Personally I would enjoy a game where you can sit back more and take in info while only occacionally having to direct the action. " (page 3 in the above thread for the whole post)
There is no reason whatsoever to not make the Group AIs on offable. I mean theres probably loads of games on maps you dont like or in mods you dont want to play at the moment , but nobodys forcing you right ? So let people swich of the Group AIs, who cares? You dont have to play with them, do you?
You can just host your own game with AIs on
Oh and to adress some questions and statements:
Mongus, if you´ve tried KAI you will know that AIs can give you a tough time.
Alantai, it does need to be server side (hardcoded option) you even say it yourself at the end of a post: "And if you're worried that someone will implement the same GroupAI without this measure incorporated, then it'd be very easy to incorporate it directly into the spring engine so GroupAI's are forced to use it else they simply wont work." So that works out just the same as making it a server option basicly
): http://taspring.clan-sy.com/phpbb/viewt ... ro&start=0Second: What almost everybody agreed on then that was that the most sensible thing to do is make Group AIs on/off able by the host.
The earlier quote of SJ was pretty correct: "Personally I would enjoy a game where you can sit back more and take in info while only occacionally having to direct the action. " (page 3 in the above thread for the whole post)
There is no reason whatsoever to not make the Group AIs on offable. I mean theres probably loads of games on maps you dont like or in mods you dont want to play at the moment , but nobodys forcing you right ? So let people swich of the Group AIs, who cares? You dont have to play with them, do you?
You can just host your own game with AIs on
Oh and to adress some questions and statements:
Mongus, if you´ve tried KAI you will know that AIs can give you a tough time.
Alantai, it does need to be server side (hardcoded option) you even say it yourself at the end of a post: "And if you're worried that someone will implement the same GroupAI without this measure incorporated, then it'd be very easy to incorporate it directly into the spring engine so GroupAI's are forced to use it else they simply wont work." So that works out just the same as making it a server option basicly
well that might be true or might be a complete mad statement.There is no reason whatsoever to not make the Group AIs on offable. I mean theres probably loads of games on maps you dont like or in mods you dont want to play at the moment , but nobodys forcing you right ? So let people swich of the Group AIs, who cares? You dont have to play with them, do you?
You can just host your own game with AIs on Smile
If it really gets the ppl into it, you might be the only one playing with ai's off, therefore hosting a game for yourself only
