Target Selection

[Storm]

Storm your arguements are cries for help in a world where you dont need help

Please drown.

but your threats to leave only comfort me and make no effect on the arguement

Threats? I've just given up trying to debate retarded newbies. I'll let you go blaze the game as much as you please. I admit my defeat against idiots. Thank you for my patience.

[AF]

hmmm? You?!?!? Admit defeat?!?!?! I cant believe my ears (well eyes technically).

Eitherway you where all defeated before this debate started, the SY's chose GroupAI they arent getting rid.

Eitherway we are all in agreement that GroupAI is a powerful tool to create unfair situations and remove certain aspects fo the game.
Suggestions to fix this are:
Ability to see who has what AI installed
Ability to downlaod AI from other people ingame
Ability for host to turn GroupAI off
Get rid of GroupAI completely.
Central review system so only fair AI can be used and standardised.

[Warlord Zsinj]

Please people, noone has countered any of the points I have raised. I suppose I'll have to wait for SJ or 10053r.

---------------

To respond to the non-posts (although I am beginning to head Storm's way of utter frustration):

better or worse than human is up to the human, if he dosnt want an ai to be better well he just has to make himself better.

This is a simplistic answer to a complicated argument, and a complicated problem.
People aren't expected to calculate complex math equations. People aren't expected to build precision instruments or attempt high-speed construction. Computers do this. There are simply things that humans cannot compete with.
But unlike calculators, computers and other machines, the end user isn't some consumer at the end of an assembly line, the end user is the very person the machines are invalidating.

can do that already just setup some waypoints for patrol and set a massive build que of peepers, but then its a bad more because a peeper lasts about .5 of a second vs any AA by the time they have guardians.

Again, you are not actually dealing with any of the points I raised. The guardian point was to show how an AI can be used to abuse a certain system of the game. Imagine getting all my bombers to do make sure that they always off-screen bomb (or line bomb), etc, etc.

Also what is a skilled player? Is it a smart player or just some player who can click faster.

A smart player. I have never argued for more clicking - as may have noted with my attack on Warcraft and Starcraft for similar reasons. I think that the AI's reduce the level of intelligence required to play the game, rather than increase it.
SJ unintentionally alluded to this when he said:
"Although it might of course be better for a newbie to let the ai handle everything."
I don't want to take SJ's words out of context, so I'll let them speak for themselves.

Also how many people lost to skirmish ai? it was a cheating ai and even on hard it was a walk over.

This is hardly a comparison. The Skirmish AI of TA is 7 years old. And was rushed even back then. And with some tweaking it can still get a decent base up and put up an average fight. What happens if you throw 8 years into the mix, as well as complete 3rd party interaction. This is a non-argument, and you are really scraping the bottom of the barrel if it is all you can bring against my arguments.

-----------

Alantai: Storm raised points, they were generally ignored. Your aggressive repost does nothing to solve the situation.

Warlord, If soemone does do that and your in a situation where an AI & human combo beats you where the AI gives a grossly unfair advantage, note their name down and never play them again, simple as that.

This is a silly solution which doesn't approach the problem at all. It will simply divide the community into those who use the AI's and those who don't. And it doesn't solve the fact that the AI's are a bad idea in the first place at all. It is "ignore it and it'll go away" logic, which never worked in the first place.

And mongus; What the hell are you talking about? Are you reading the same thread as we are?

[AF]

Storms points and the points of you all worrying that you will not be able to have fair games because of n00b b*stards and coders who aer up to no good decide to code a GroupAI that gives them so much more than they should do that they cannot possibly be beaten by a human is questionable but it is still a valid point that does not have a clear solution.
I have attempted to list possible aides in preventing these fiascos and gut wrenching problems. Remember I am in a position where I need the GroupAI to do what I am doing with skirmish AI and it is the only interface I have till the GlobalAI interface is finished.
The GroupAI is already in use for such things as metal maker AI etc, which will be missed if erased unless built in as a sub-behaviour meaning another discussion and more argument.

As of now there is a very, very, very small chance of you getting GroupAI removed but that is a very, very unlikely outcome. Your aims should therefore be placed on limiting the power and abilities of the GroupAI interface to provide an AI with what it needs, but that makes my job a lot harder than it needs to be.

Of everyone here, who knows how to write a GroupAI save SJ me and napalm? I bet you the possible users outnumber the people who could do it. And of those people, who would have the time to code something and research the necessary tools to code it without ti interfering with other things? For one you have to code the AI to understand your actions and figure out what it wants you to do *he told me to attack that missile tower, does he want me to kill all the missile towers or just those in that area, or does he mean AA weaponry and no just missile towers? *

AI dealing with playing the game for you?! Maybe those sorts of things would do in a skirmish AI but in a normal human v human game? For one the sort of strategy from an AI that aGorm pointed out would be terribly predictable once put into action and could easily be eliminated with an *every time AI does that you do this*. Perhaps you could implement a tactical agent system for attack like I have planned for TAI? But will you really be willing to start numerous tactical agents by hand while your enemy is fighting you and pummelling your defences?

RTS AI is pathetic; it is the most backwards area of AI research and does not compare with AI in FPS games etc.

I would say that if people where pushed to this the first AI of the sort you and storm fear would be a variant of TAI or another skirmish AI, simple but does the job, iffy about control since it's not designed to share the interface with the human and easily sought out by the other player.

Then there is who would do that?

This is a silly solution which doesn't approach the problem at all. It will simply divide the community into those who use the AI's and those who don't. And it doesn't solve the fact that the AI's are a bad idea in the first place at all. It is "ignore it and it'll go away" logic, which never worked in the first place.

Yes but those people will then find themselves as a minority as nobody would play against them, they'd end up playing against the AI itself and they'd loose all fun from the game and drift away.

And mongus; What the hell are you talking about? Are you reading the same thread as we are?

Mongus was saying that this debate is over and leads to no conclusion; though it may still be productive it is overall a negative and not a positive.

Please people, noone has countered any of the points I have raised. I suppose I'll have to wait for SJ or 10053r.

I think it would be wise that you post a brief summary in bullet points of what your points are as some people obviously have lost track of them.

Now lets see where this heads:

Current way spring is heading the AI will have to take in more and more information. Buggi├óÔé¼Ôäós satellites, SY's energy web plans, huge maps, they're all making the game wider in a way humans may be fine with, but the AI will not. Calculating how to deal with energy webs on an epic size map may just aswell cripple a GroupAI and the system its on when joined by everything already there and satellites and weather and terrain properties of the new map format etc.....

Count yourselves lucky that a sufficient GroupAI to cause what you predict is still hard to code and is getting harder as we go along. Humans are abstract, the AI deals with numbers and values and calculations, something that a human tends not to save resources and build times which aren├óÔé¼Ôäót precise values in an algorithm. Yes the AI may be able to build things at best but don├óÔé¼Ôäót we all follow a pattern when building in the first minute or two for certain maps? Is that unfair? After all the AI would likely do the same as we have done by testing different combinations till we find the best that suits us. Maybe the fact that humans can learn is enough? Machine AI that learns is costly system resource wise and takes a lot more time and effort and thus would only ever be worth it for a skirmish AI and even then it would need to be restrained because of system resource issues.

But unlike calculators, computers and other machines, the end user isn't some consumer at the end of an assembly line, the end user is the very person the machines are invalidating.

Your logic is flawed in terms of consumer, we are the consumer and the customer, if the AI however controlled then the AI would be the customer and the human the consumer. What you are referencing is a skirmish AI and skirmish AI can be beaten. A skirmish AI guided by a human player however is what I see your worries are actually about.

People aren't expected to build precision instruments or attempt high-speed construction.

Are you suggesting they use game trees to solve the equation that is the battlefield? I thought the battlefield was variable; constantly changing, and such a tree or construction would be very costly and could never be finished in an rts environment without the use of an infinite amount fo supercomputers.

People aren't expected to calculate complex math equations.

We have heuristics for that, something the AI cannot do, we can gage things and compare and take in every detail while an AI has to have those defined to use them, and then an algorithm.

There are simply things that humans cannot compete with.

There are simply things an AI cannot compete with, remember we are sentient they are not.

Again, you are not actually dealing with any of the points I raised. The guardian point was to show how an AI can be used to abuse a certain system of the game. Imagine getting all my bombers to do make sure that they always off-screen bomb (or line bomb), etc, etc.

Point taken but this is a sub behaviour that should already have been implemented but hasn├óÔé¼Ôäót. *if miss target will it hit friendly unit?*. And considering this is a simple behaviour that cancels out something that is of great nuisance and is detrimental to the enjoyment of the game I suggest this be implemented, however a poll should be created and consensus reached first.

[AF]

I see now that maybe it could be made so that the GroupAI was only available to the GlobalAI or those who ont abuse it. Those who do abuse it could have their privilledge taken away by the enemy, any hack to remove this failsafe as described above would remove the ability to play against others.

And as for GroupAI being removed, we could say the same about cheating, the community polarised into the cheaters and non-cheaters.

[HellToupee]

If someone can create an ai that can consistantly outplay humans without cheating well then they would be a genious indeed. The only advantage vs non ai assited human player is, u could create all these ais to handle all the menial repetivitive tasks, so u can get the other guy while hes busy with his menial repetivitive tasks.


Also this freedom why would a human need ai help in calculating getting one quickly and why during game, to do this u can just memorise how.

[AF]

If someone can create an ai that can consistantly outplay humans without cheating well then they would be a genious indeed.

That is because:

The only advantage vs non ai assited human player is, u could create all these ais to handle all the menial repetivitive tasks, so u can get the other guy while hes busy with his menial repetivitive tasks.

AI's do not have heuristics and abstraction unless programmed specific instances, which are derived from humans. To make an AI make abstractions about the game is to enable it to do many things that an AI hasn├óÔé¼Ôäót done before in this field, that is what I am trying to do with the skirmish AI and that is what enables an AI to play strategically without the human telling it what strategy to follow and how.

And right now tti is a little abysmal, darkstars is a 4 person community, me napalm, triaxx and min3mat, maybe the odd drop in by aGorm and Liger after a long absence, and we know there are people in the AI forum who could contribute to theory ro code or just pseudocode but dont, maybe they're busy with their own skirmish AI plans though I doubt they'll be able to carry them through.

There simply isn├óÔé¼Ôäót enough involvement in AI to develop more than 1 project, that being the skirmish AI unless more people join in, never mind these GroupAI that you speak of.

For one these GroupAI where known about by some before the engine was even publicly released and this thread has only materialised in the last week. Napalm even posted a simple example of a peeper going to the middle of the map and exploring moving out in concentric circles, and until code was written at all, nothing was made with regards to AI, and a very long time passed by.

edit: sorry kixxe I changed the layout a bit so it was more understandable.

[Kixxe]

Im.. not qwite following you Alantai...

And I have offered my suggested solution several times. It is my opinion that radar dots should be kept as they are, but units should not auto-target them unless a target facility has been built. This suggestion keeps the game-screen dots, which makes it far easier to autotarget then when using the radar (reducing micro), but it balances the radar, and makes it so that radar coverage is no longer do or die. It makes attacking easier, and it makes aircraft less useless.

Your point raisen (btw this thread is going HEAVLY offtopic)

No. That would destroy the need for popup cannons. Offcourse, i could use em, but wheres the fun in clicking 16 dots evry attack?

Whats wrong whit radar do or die? You need intel, and you need that your enemy dosent have intel. Therefore radar/jammers. Radar is avilbole to almost all unit types, except lvl 1 units. And at the lvl 1 stage of the game, lvl 1 units dont relly have any benefit from radar.( and BTW most things in Spring are do or die. Build AA, do or die ect)

To have to click attack for every unit is micromangement yes, but an annoying type. So each time a unit attacks, i have to click on them to keep them from getting head to head whit my defences?

I can se it's usefullness in that air will be more powerfull, and that i have to do something when someones is outside whit a fidos who are shooting right outside my HLT base.

But it wont add anything to the game except that when you are building building, radaing, exploring, and building large armys you have to at the same time click dots to make your defence worthwile.
aint that a litte silly?

I see you've all ignored my requests to proofread your posts. What can I say? Maybe you all really are lacking in neurones, rather than just lazy, as I originally thought.

Aculy, we are both lazy. We wont take the time to spell check, and you wont(or will you?) take the time of just reading what we have to say, besides the errors and the overall annoyance( is annoyance a word?)

[Warlord Zsinj]

As of now there is a very, very, very small chance of you getting GroupAI removed but that is a very, very unlikely outcome. Your aims should therefore be placed on limiting the power and abilities of the GroupAI interface to provide an AI with what it needs, but that makes my job a lot harder than it needs to be.

Of everyone here, who knows how to write a GroupAI save SJ me and napalm?

Oh please Alantai. You should hop off your high horse, and realise that your AI is not the be all and end all of Spring, and that Spring does not revolve around your AI creation. Honestly, if you read over your posts - it is all you talk about, but it never really adds anything to the discussion. "My AI will solve this" "Given my AI, this isn't such an issue" "Wait till my AI comes out, and I will make it that the problems will go away" "I will make the AI not do this, and this, and that". What AI? Until you have released your AI, I highly suggest you keep quiet about it, save where working on it is necessary. You cannot talk of the effects something will have on a game if it doesn't exist yet. Perhaps it will never exist, and allowances we have made for it will leave gameplay holes in our game? There are issues that need to be attended to now, that are facing the game now. Talking up your AI solves nothing, and generally diverts attention from the discussion at hand.

Apologies if that was crisp, but it is late, and I am beginning to get rather disenchanted with this discussion.

-------------

EDIT: responding to your post, kixxe, which appeared during my post.

You could click on defending targets - yes. There is nothing wrong with that, it just means that you have to tell your units to act on your intelligence. Or you could just spot for your defences to get LOS on them. I think having to LOS for your defences is a realistic solution (most military forces will get visuals on radar signatures before firing), and is a much better option then all the gameplay issues that auto-radar firing poses.

Whats wrong whit radar do or die? You need intel, and you need that your enemy dosent have intel. Therefore radar/jammers. Radar is avilbole to almost all unit types, except lvl 1 units. And at the lvl 1 stage of the game, lvl 1 units dont relly have any benefit from radar.( and BTW most things in Spring are do or die. Build AA, do or die ect)

I have explained numerous times that radar do or die makes defensive positions far stronger, and attacking almost useless. Basically, radar becomes less about keeping an eye on your enemy, and more about extending your unit's LOS. That's what radar is right now - it's extending unit LOS. It means that anyone with radar coverage has effectively extended the LOS of all units within that range.
And your level 1 example just shows a poor understanding of TA. Level 1 attacks would literally be practically useless in Spring, because I wouldn't have any means of blocking that enemy radar (jammers) or getting my own radar vision (mobile radar units).
And AA is not do or die. AA is just a single option in a massive permutation of possible strategies. I only build AA if my enemy is using aircraft. And then I can just use missile units. And then I could build aircraft. And then I could rather hit the enemy aircraft labs. So no, building AA is not do or die. If I cannot attack without radar coverage, or preventing radar coverage from the enemy, then yes, radar is do or die. This reduces my choice in spring, and therefore reduces my strategic options.

I can se it's usefullness in that air will be more powerfull, and that i have to do something when someones is outside whit a fidos who are shooting right outside my HLT base.

Pfft. If you let that HLT get destroyed, it is your poor play, not a weak game.

Aculy, we are both lazy. We wont take the time to spell check, and you wont(or will you?) take the time of just reading what we have to say, besides the errors and the overall annoyance( is annoyance a word?)

Annoyance is a word, Aculy is not. The reason I don't take your suggestions as seriously is because you are lazy and don't put the effort in to communicating your ideas. If you can't be bothered explaining your ideas in a way that is easy for me to understand, what sort of message am I getting about the nature of your ideas? Perhaps you put the same amount of thought into your posts as you put into your ideas?

[munch]

As I see it there is two main sorts of fun in a RTS game, one more action based and not needing much deep thought such as clicking radar dots and another more strategic.

Nicely put. This is what I was trying to say - only I took pages =)

Somebody mentioned a while back about Civ, where the gameplay is very strategic. I must admit to being somewhat of a slow thinker, so turn based games like that suit me better, but even so, I love TA because it has strategic level concepts (like using aircraft to spot for artillery) along with minute tactics (like rushing an enemy with a single fast unit that the comm can't keep up with and microing its movements so that the single weak attacker is always on the other side of the VP/lab to the comm). I love both of these aspects.

What it comes down to I've realised is decisions. You want each mouse click to be a significant decision game-wise.... and that might be a strategic or a tactical decision.....

I totally disagree with the statement "tactics has nothing to do with strategy", that's like saying a tree has nothing to do with leaves, or a game of chess has nothing to do with the chess pieces. Many a good strategy has been defeated by a worse strategy with a better tactical implementation. Both strategy and tactics clearly controlable in the same game are key to TA's success I think. Whatever changes are made I think it's important not to lose sight of that. If we can retain the ability to make decsions at all levels of strategy/tactics then I think we'll retain the fun and diversity of play that we have now: grand strategists can clash with brilliant tacticians. Moving towards supporting better strategy is good, but we mustn't squeeze out the tactical element in the process.

Perhaps, SJ, but while you move towards making greater steps in allowing strategic gameplay, you seem to be going at it somewhat blind (no intended offense), without a clear vision. For example, while you make all units target radar dots which reduces the micromanagement required, you also bring in a massive change in strategy itself, which I have argued consistently is not for the better (briefly: massive power to defensive players, radar required just to attack, aircraft become useless, etc).

Hopefully the above paragraph addresses the first half of this. However I can see what you mean about the blips... but only because of missile units. The inaccuracy in normal radar is so great that even with area effect weapons like MRPCs you're lucky if you hit a stationary target. If there is a swarm coming in then you have a better chance of doing some damage. LOS weapons such as lasers have no chance though. However, missile units can hit the targets because they correct their aiming in flight. For my money, the right solution to this is the one taken in the UberHack mod, which is that missle units become exclusively anti-air units. I can see that this might be a problem for incoming aircraft, but that's what AA defenses are for. If it's not possible to build fully effective AA then there's something wrong with the game anyway. The answer to effective enemy AA has got to be something other than aircraft (i.e. soften up with artillery/naval assault, or go for a ground assault), otherwise the AA is not doing its job.

I've gone on and on again haven't I?

Just my tuppence

Munch

[AF]

pfft, see my note on only allowing the GroupAI when sued byt he Global AI.

See thread on darkstars in AI forum
See Numerous threads ind arkstars, namely Pseudo.Cpp.

I suggest you reread, I mention ym AI when soemhtign has come across me that I had already encoutnered through thought over AI, though usually the manifestation was at darkstars behidn clsoed doors btu it's all opened up now so go read. And if you think that my skirmish AI plans arent the end all fo AI ins pring then maybe your best motivating the rest into their rival projects as I seem to be alone in my project save a handful and nobody else seems to care, that is what I was pointing out and that si why the groupAI you fear so much will never coem to light, nobodys bothered putting in the effort anymore.

[Warlord Zsinj]

I wasn't talking about those threads. I said that you shouldn't introduce it into this thread as evidence, as it doesn't exist yet.

And instead of waiting for everyone to do things for you, why don't you put your head down and get working? If you put the time you put into posting on this board working on your AI, it would be half way done by now.

---------

Nice post munch, nice to get someone else capable of expressing themselves logically.
I disagree with the all-missile-units-as-AA thing. It is a bandaid fix, and doesn't really solve the problem. I don't think missile units are the only ones that get a massive advantage from being auto-fire on radar targets. Anything with a bit of splash, or built in large numbers (such as defence installations *hint* *hint*) get an advantage.

To further explore while I feel it is a bandaid suggestion - if units are so inaccurate while trying to fire at radar targets, explain to me the purpose of having the function [of units autofiring on radar targets] is, other than making the terrain leading up to your base unpassable? And why is there such opposition to my suggestion of making units not autofire on radar targets, if firing on autotargets is so useless anyway?

[AF]

Because radar targetting does the targetting for you and it's annoying when you have to click all those dots.

I wasn't talking about those threads. I said that you shouldn't introduce it into this thread as evidence, as it doesn't exist yet.

I havent referred to the gameplay of TAI, especially since the only gameplay noticeable is theoretical or imaginative as it isnt in a workable state as of yet. However this thread wa mainly about AI and the prospects for gross misuse through the GroupAI interface which TAI uses, I was meerly pointing out the state of affairs and how it applies.
Now if you dont mind, TAI has no material on sourecforge and I need to start moving things and updating things. I'd also like to see who else has actually started a skirmish AI, because I'm tired of leading the only project that seems to have gotten anywhere while possible other projects could be hiding waiting for me to complete first and be harassed and flamed, remember there are high expectations of me and I haven't got much help.

Now enough of TAI, I shall not mention it again in this thread and I a sure everyone else will follow suit.

And your tuppence is good munch:

The answer to effective enemy AA has got to be something other than aircraft (i.e. soften up with artillery/naval assault, or go for a ground assault), otherwise the AA is not doing its job.

I think that though you explain well you should think more of the bigger picture, aircraft are under powered in spring and many ahve complained of how inneffective they are now that auto targetting has been implemented. I shall have to ressurect the unlimited altitude && cylindrical/oval weapon range thread, at least that provided a solution to the problem, although controversial.

[Min3mat]

autotargetting is usefull when a horde of enemies are approaching / using LRPCs

[Storm]

hmmm? You?!?!? Admit defeat?!?!?! I cant believe my ears (well eyes technically).

If I had the time, I would go on and battle the wall of infinite stupidity posed by the countless brainless newbs on these forums, but alas, I have better ways to spend my time, especially since everyone of you are notoriously ignoring every single of my fucking points and reading your posts gives me a fucking headache. God I hate retards.

Why even bother, I'll go help Gnome rebuild the game instead.

[Yeha]

In Spring we want units to be intelligent, they should execute your strategic and tatical orders to the best of their abillities.

You want your units to fire on radar dots on the map, your units should know that and fire on them by themself without you having to constanly clicking on them.

The unit AI should help you execute your orders to your wishes, not stand and do nothing because you don't have enough time to click on every singel one of them. You should be able to lay out a good strategy and have you unit execute it and knowing your intention without having to individualy babysit them for maximum performance.

The group AI is there to realize this and giving the possibility to create intelligent unit behavior.
Not everyone thinks of intelligent unit in the same way, do you want your units to stupidly charge into a battle and get blasted to pieces, or do you want them to cowardly escape home when damaged a bit? Having the player able to choose the AI makes the unit behave intelligent acording to them.


Does this change the balance on uints? Yes it does, but in my opinion it makes more sense to balance the units acording to the engine, not the engine acording to the units. We haven't started to really balance the units yet, the game is still changing to much for that.

Would it be possible to create an AI that becomes to good? Possible i guess, but i don't think so, and if needed in the feature it wouldn't be hard to add a 'tournament mode' to the battleroom to only allow some set AIs or need everyone to have the same ones, but we will see about that when the time comes.

In short, your units should help you, not hinder you.

[Felix the Cat]

hmmm? You?!?!? Admit defeat?!?!?! I cant believe my ears (well eyes technically).

If I had the time, I would go on and battle the wall of infinite stupidity posed by the countless brainless newbs on these forums,

Yes, because we all know that elitism is the answer to all problems.

[aGorm]

Having just spent 20 mins attempting to get past the bad grammer and the even worse moods...

Can no-one understand what is wrong??

Firstly.... how the hell did shooting radar blips get into this? They have been argued about enough and arn't realy anything to do with these arguments, because they arn't realy that much of an alteration. Just becasue you units now shoot blips does not mean that there's any intelegence behind the fireing. It's infact no different to you yourself telling your units to attack the blood blip thats nearst, which as all you might as well do, because unless you happen to magicly know what that blip is, your telling your units to fire at a certain blip will make no difference.
SO leave the dam auto radar targeting out of this. All it does is save you the hassel of manuly clicking.

On to the problems with groupe AI's.
The problem lies in the fact that it basicly unbalances the game.
The situation would be as follows.
Player 1: Has his own AI's taht help do the little things that dont make realy a grand difference, but do save him alot of time, meaning he can concentrate on his game.

Player 2. Did not know how to make his own AI's. He does everything manuly. In the end, unless he's very good, he will lose.

Now your about to go on about the fact player 2 cant write his own not mattering. But it will. To counter the arguments that have already been mentioned (these are the ones Alantai gave.. if theres more tell me, i cant remember all of them.)

Ability to see who has what AI installed

How will this make it any easyer for player 2? Unless he knows in detail what this does, it will make no difference.

Ability to downlaod AI from other people ingame

Who say's the AI's player 1 is useing are going to benift the playing style of player 2? Unless they both play teh exact same why, player 2 is still at a dissadvantage, and possibly more so of one, if the AI infact messes up his stratagie.

Ability for host to turn GroupAI off

If the host has a groupe AI he realy likes why the hell would he turn it off?

Get rid of GroupAI completely.

thats wrong to... means no formations (currently) no smart metal makers... etc... It does have some use's!

Central review system so only fair AI can be used and standardised.

Alot of AI's will be fair... but as said earlyer up, there may be none that complament player 2s style of play.

Well, I've gone on lots so ill let some people get back.

The bottom line is... Yes Groupe AI is OK for a few things... like the formations and the metal managment and stuff, but letting people make there own to help them play is unfair and above all, unfun (if there is such a word)

aGorm Out

[SJ]

Some random quotes and answears.

If people can program their own AI's, and pit them at their foes, it is simply a matter of time before they are exploiting minor engine bugs (you'll never get all of them).

Well in that case we should fix the problem with the engine and not shot the messenger. This is especially true since most things that can abused by an AI could also be to some degree abused by a player that just learns to click fast enough. We might not be able to get all bugs in a single release but spring is supposed to be an ongoing process and over several versions the bugs will be found and eliminated.

Perhaps, SJ, but while you move towards making greater steps in allowing strategic gameplay, you seem to be going at it somewhat blind (no intended offense), without a clear vision.

Sure we dont have a final vision for how Spring might finally look. If we had much of this board discussions would be meaningless since the development would be set in stone anyway.

As I said earlier, when playing a game, the only variable should be the player. Thus, in any game, if you are defeated, the invariable conclusion is "you are better than me." What happens, then, in Spring, if you make the game itself the variable, so that it is no longer "you are better than me", but "You had a better AI than me."

A few thoughts on this and similar arguments

1: You massivly underestimate how hard it would be to create a good AI which means there will probably be only a few major ones created by groups of people which will more or less automatically mean they will be available for download.
2: Even if one assumes that a single person can create a superior AI what will give him the biggest ego boost, winning a few games or seeing everyone else use his AI?
3: Assuming that someone keep his superior AI secret, why is it worse to lose to someone because he can build a superior AI than because he can click faster than me ?
4: To make everything fair should I also outlaw computers above 2.6 GHz and mice with more than 1600 dpi since those might give the opponent an advantage over me ?

This is hardly a comparison. The Skirmish AI of TA is 7 years old. And was rushed even back then. And with some tweaking it can still get a decent base up and put up an average fight. What happens if you throw 8 years into the mix, as well as complete 3rd party interaction.

Uhm as far as I know you would still not get an non cheating AI that can beat a skilled player. Especially not a player that can use AI subroutines to handle his own routine tasks.

I have explained numerous times that radar do or die makes defensive positions far stronger, and attacking almost useless.

Then the answear would be to weaken defenses, not to force the player to click like mad on the minimap. Although as a counterpoint i have also seen several ppl saying that defenses are now too weak and must be strengtened ....

Player 1: Has his own AI's taht help do the little things that dont make realy a grand difference, but do save him alot of time, meaning he can concentrate on his game.

Player 2. Did not know how to make his own AI's. He does everything manuly. In the end, unless he's very good, he will lose.

First see my comment above, but as a somewhat extreme example. Assume that player 1 know about metal extractors but player 2 dont. Should we then outlaw metal extractors since these will clearly give player 1 an advantage against an otherwise equally skilled player 2?

[aGorm]

No... U missed the point (just)

In spring... metal extractors are avalible to both sides... and you can hardly say one complaments a certains style of play, because there fundemental to the game. They are usefull no matter how you play.

Someone having AI's to take care of functions for him, that are either not availible to the other play or are no use to the other player if availible, means that player 1 does suddenly have an advantage.
Having just one would prob make no difference, but having say 20 AI's would give you alot of extra time. (or 1 AI that does 20 things... just incase someone goes well maybe we should say each player is alowed 3 AI's or somthing.)

Following that last point... has anyone thought of that?
Maybe there should be say 20 AI's to chose from, and you can pick 3 befor you start, and assign them in game to wahtever. Then noone could 1)write there own.
2) have a realy big advantage over a non AI useing person.

Is taht maybe a compramise?

aGorm