AIs behaviour/improvement

[deanosbeano 0]

Quote[/b] ]You know, you could just ignore such threats and publish your opinions.

as you can see , i am not shy to publish my opinions and its not the beta nitpicking that i am talking about ,its the bigger debate,

also me and you know body can stab you over the internet, because we are adults, but it isnt always adults that play ofp and write here. and it isnt always the physical hurt that hurts the most,its the mental hurt.but anyway thats ottt for this problem, i merely quote them so there is not doubt to whom i am speaking.

i personally now use this option

Do you wish to view members signatures when reading threads? no

the end see you in arma.

[dreday 1]

4) I'm not sure that BIS can improve the ArmedAss AI much beyond what has already been done in the OFP community by scripting, without  seriously affecting the minimum system requirements.

This point has already been debated a million times over.  It has been mentioned by many members (myself included) that VME mood’s AI is much more impressive than what we saw in ArmA videos.  I am able to run platoon sized engagements with VME CQB soldiers while taking very little hit to the performance on my system, which is just slightly above average by today’s standards.

Now don’t get me wrong, VME AI is far from perfect and it is very badly documented (which further limits its utility), yet it is a clear improvement over the core OFP.  

No one is expecting the ArmA to have human-like AI, but I don’t see why something similar to VME AI can not be implemented at core engine, rather than scripting level.

Peace,

DreDay

[Friedchiken 0]

I agree that the VME AI is very good. I just downloaded the mod again and I'm very impressed with the quality of firefights around the area of Lipany I can form.

I would encourage BIS to place some AI improving scripts in their missions if it would make the OPFOR AI more aware of the tactical situation.

Maybe some scripts that lay out a danger plan for AI under attack by a player sniper? I mean someone needs to plug a large shell into the man-bush if the idiot is out in the open.

[mr.peanut 1]

4) I'm not sure that BIS can improve the ArmedAss AI much beyond what has already been done in the OFP community by scripting, without seriously affecting the minimum system requirements.

This point has already been debated a million times over. It has been mentioned by many members (myself included) that VME mood’s AI is much more impressive than what we saw in ArmA videos. I am able to run platoon sized engagements with VME CQB soldiers while taking very little hit to the performance on my system, which is just slightly above average by today’s standards.

Now don’t get me wrong, VME AI is far from perfect and it is very badly documented (which further limits its utility), yet it is a clear improvement over the core OFP.

No one is expecting the ArmA to have human-like AI, but I don’t see why something similar to VME AI can not be implemented at core engine, rather than scripting level.

Peace,

DreDay

VME would be what I meant by what has already been done in the OFP community by scripting to improve AI.

[dreday 1]

VME would be what I meant by what has already been done in the OFP community by scripting to improve AI.

In that case, wouldn't you agree that it would be fair to expect the core ArmA AI be about as capable as VME's? I hope that would be the case in the final product, but it sure did not look that way from the videos...

Peace,

DreDay

[benreeper 0]

No I wouldn't because VME is a mod. It is an addition to a game and you can't expect a game to have in it everything concievable thing that CAN be in it. It's the same as vehicles. OFP had about 30 vehicles in it and mods have created about 500 vehicles. Would you expect BIS to have put 500 vehicles in the game because mods have done it? Most games don't even have incompetent AI to begin with because it is not cost effective to develop. You are not going to have a developer devote money to a feature that most gamers (not PC OFPers) don't even notice and do not see as a feature. Just look at E3 this year. The big stories were about graphics and physics, there were none about AI. Heck BF2 removed AI to devote their time to other areas. I ask you this: if the AI isthe same as it is in OFP will you still buy the game? Believe me, BIS knows the state of their AI. They play the game themselves and are doing what can feasibly can be done. I'm not saying you can't make suggestions to them for upgrading the AI, I'm saying that they know already.

What I would like to see is more scripting control over the AI and more ways to override it at the core.

--Ben

[lwlooz 0]

I ask you this: if the AI isthe same as it is in OFP will you still buy the game?

I would seriously considering buying the game. What people around here don't seem to get is that if the AI is malfunctional all other additions or improvements are rendered useless if you want to do combat with AI. Every single one of them. And the idea that the AI can be fixed with some sort of scripts is highly unbelievable for me.It is not the Inter-Squad-Relations on Level-2(Platoons) or more AI that needs improving.It is that very individual core AI processes and squad-thinking that needs urgent improving.This isn't a minor thing,anyone who plays COOP or wanted to set up a realistic scenario in OFP knows that the current AI ruins it.

OFP provides the basics for combat AI,but they need to re-define/tweak Engagement-Procedures(What point does their IFV-model with an ATGM make if the AI fails to properly use it?),Survivability-Decisions,Coordination.The basic thing is the AI needs to have some staying power.

It wasn't me that claimed you wouldn't notice the differences between humans and AI in Mulitplayer,but maybe Placebo was refering to TDM or CTF,well you won't notice a difference there because humans are headless chickens as well.

And yes,I am mentioning it because I believe people who create the basis for what hopefully in the future will become the first true Tactical Warfare Sim(Although i might have to save for VBS5 there :/) can come up with AI that puts up a proper fight so the players(thats us) actually have to use tactics.

I also keep posting about the AI issue on this forum,because there is so much noise around here,that a message like that can get easily lost and I strongly believe its not only me that thinks AI is priority #1 (Well,surely everyone who doesn't just make screenshots of new shiny addons all day).

Oh and I KNOW ArmA isn't finished yet,but I kinda have the experience that if you give hints at what would be nice before a development process,afterwards you get better results.I agree with all the fanbois that pointing out every single bug is silly,but as I said,AI is not a minor or small thing.

[Bloodnick 0]

Grenade use! That's one thing I really miss in ofp. Every 10th AI on a battlefield uses his/her grenades max 2 times. I haven't seen much of it even when I played with veteran difficulty. But I hope they implement a grenade warning system , like in COD2. There are more grenades used there than ammo (It's only a figure of speech, but the AI there use a lot of grenades).

I liked the first post where MGS3 was mentioned, I want the AI to aim in the area they thought they heard something and not lie down and open fire and get killed by atleast 5 headshots. And please fix the AI teamplay, they heal and supprt many of their squadmembers, but when there's only one soldier a squad of 12 is fighting against; they send soliders after him one at a time. I experienced this in Ia Drang, the vietcong sent one at a time after me in a bunch of tents, and when I killed one, another one came after me (plus they had a terrible aim; yes I played with veteran difficulty ). But back to MGS3 AI, they should be a lot more cautious when they hear footsteps or seen a mysterious leg disappear behind a building (Call for backup would be a nice feature). The Aa specops could maybe have Swat or SAS "AI", they move around silently and always stick together and use flashbangs to enter a room. But please, don't add real spetsnaz AI, then you have to fight against acrobats in a forest that can maim or kill you with their feets or take down two enemy soldiers equipped with knives

The vbs1 convoy scripts is a must for Aa, a mission with terrorists that's taken over a 747 or antonov, and you play as a civilian cop and walks around in the jet and takes down the terrorists and saves the day of 200 passengers . Or fighting in iraq and you get killed by a rebel drive-by by a Toyota pickup , or sitting on or inside tanks (large artillery tanks like Paladin that is), then you're more safe from AT-soldiers

P.S: Let the AI behaviour change from what kind of soldier classes they are or what kind of vehicle it is. I really hate it when the blackhawk or the hind acts as an attack chopper and not unloading the cargo first before slaughtering soldiers on the battlefield.

[funnyguy1 0]

Well, It`s not about the variety, it`s about quality...

By adding some extra things to AI It becomes better, while adding XX variants of M113 doesn`t make the core game more realistic. BIS can improove the AI by looking at what`s been done by the community, but of course they are the developers and do what they want. It`s not that I disagree with your post benreeper.

I just want to say that AI is something that can, and should evolve, no matter if this evolution is caused by community-made additions, or developers, or even developers inspired by the community-made addons. Quite the contrary(?) to the vehicles...

edit: To prevent this thread being locked I suggest continuing the generall discusion about the AI here, everything else can go to the next gen\suggestions\AI topick. I simply don`t want it to be locked...

[kernriver 4]

I also keep posting about the AI issue on this forum,because there is so much noise around here,that a message like that can get easily lost and I strongly believe its not only me that thinks AI is priority #1

Of course it is. If BIS only improved AI and did nothing else, ArmA would still be a must have. Because, for me it's all about gameplay, good graphics and sounds are fine, but gameplay is what gets me goin'.

[simba 0]

something that should'nt be that hard to implement is to have the possibility in the mission editor and the command menu to determine wish way the ai is going to attack, defend, protect, retreat ,....

for example you could order a mi 24

- to strafe quickly

- hover + search and destroy

- drop troops then protect them

-...

or you could order a soldier to cover you , walk and hide behind an armor , protect the rest of the squad while they are desembarking a vehicle ,...

but this is more work for the player , some of these things should be automatic.

[mr reality 0]

I would like to see the AI fallback when they run out of ammunition or if there is none available on the ground. Or maybe have the option as a unit commander to spread the magazines around the units at a lull in the mission without having to order individual drops and pick-ups. A soldier should be able to collect ammo from a dead buddies body without having to be ordered to do so (not sure if this is already implemented).

If the gunner of a tank or APC is badly wounded he's unable to operate it-thus the vehicle would have to retreat. The retreating aspect of OFP has to be done better as the current system tells the squad to retreat but they return to attack the same position from the same point with fewer men. No logic whatsoever. If a squad loses 60% (editors choice) of it's men they should suffer "morale" breakdown where the leader may order the remaining men out of the area.

Medics should not be in a squad as they are non-combatants. Squads have ALL soldiers trained in basic first-aid. This way the nearest man to you would heal you rather than the medic 100 metres away who invariably gets shot while leaving his cover to help you.

These ideas are probably more for game2 but i would think there doable.

[crashdome 3]

What I would like to see is more scripting control over the AI and more ways to override it at the core.

I hate to say this but I agree.

There isn't a soul in this thread that hasn't advocated change or threw in a suggestion. I've reread this forum umpteen times. This fanboy accusation crap is nothing more than a cover up for the fact that your voice isn't loud enough to drown out everyone else.

Personally, I give no care about anyone else's feelings in this thread in regards to how they value the AI in terms of priority, but having the experience under my belt of complex procedures in programming... this isn't something you can just make a firm decision on (i.e. BIS and what type of AI to use - let alone how far to take it) and expect a change this late in the game.

I only agree with ben because there is a valid point in that if you expose everything (sensors, states, etc.. even with a crappy AI base) you can mimic any type you want as efficiently as BIS can. The additional benefit is that you can tweak it if there is a mistake. If BIS makes a mistake it's broken until patched which can take alot more time and money than a few eager hobbyists.

I'm not being fanboy... just realistic.

[FatNinjaKid 0]

I once suggested in the game2 section to introduce a feature to the debug engine that would allow you to record behavior macros.

In the debug engine you push a record button and then everything you do gets recorded for, say, 10 seconds. When you are done AmrA will save a log file with the information about all the buttons you pressed. Something like:

<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE">

Time Direction Button

0.00 233 W

2.22 234 -

2.30 232 Q

4.55 180 W+D

....

... you get the idea.

Then in the mission editor you can define conditions or triggers that will make a certain AI unit suspend normal behavior (like disableAi Move) and execute the behavior macro defined in the log file. During all that time the AI should still be able to target and fire (with the weapon selected by the mission designer).

Now, of course that is not real 'artificial intelligence' and I'm in no way suggesting that you should script the whole mission like that. But there may be certain situations --like when entering a building or spotting a sniper-- when such a recording feature might really ad something to the game.

Basically that means that the AI could become as smart as the mission maker. The main advantage would have been that making intelligent missions would become much, much easier -- even for the BIS

(Of course, you could also include the log file in a .PBO addon file.)

[lwlooz 0]

So just so I get the general logic of this.You people say that BIS is unable to do AI improvements/tweaking because it is way too complex even for a proffessional coder team that does that for a living and then everyone around here claims that the VME Script AI is the way to go or if BIS just provides us with Script commands,we will be able to do the AI ourselves?

So basically,it was impossible for BIS to do it in the timeframe of the Arma development ,but scripters who do this in their freetime will come up with great AI in no time?

I am all for having better ways to interact with the AI,I may not have experience in complex programming or a degree,but I deal with trying to improve OFP AI about 80% of my freetime (via Configs.. weapon and ammo values do my head in or Scripting).That is the only reason I so insist on it,because I have to deal with AI in OFP on a constant basis,so you can call me a uneducated moron fine,but I have some experience in getting frustrated with OFP AI

And to make this perfectly clear,I am not talking about wanting the AI to do things like rope down houses,sneak up to you and throat your knife,have a smoke,chat voice lines when in safe mode(That stuff is for AI scripts). I just want the AI in Arma to do this tidbit more and faster situation analyzing that they had in OFP but which just wasn't good enough and get rid of silly procedures like sending one single guy at a time onto the enemy or rush in close combat style with your bmp or tank.The very core AI that is.Anything else,I completly agree can be done if we have better script-interaction

[dreday 1]

No I wouldn't because VME is a mod.  It is an addition to a game and you can't expect a game to have in it everything concievable thing that CAN be in it.  It's the same as vehicles.  OFP had about 30 vehicles in it and mods have created about 500 vehicles.  Would you expect BIS to have put 500 vehicles in the game because mods have done it?  Most games don't even have incompetent AI to begin with because it is not cost effective to develop.  You are not going to have a developer devote money to a feature that most gamers (not PC OFPers) don't even notice and do not see as a feature.  Just look at E3 this year.  The big stories were about graphics and physics, there were none about AI.  Heck BF2 removed AI to devote their time to other areas.  I ask you this: if the AI isthe same as it is in OFP will you still buy the game?  Believe me, BIS knows the state of their AI.  They play the game themselves and are doing what can feasibly can be done.  I'm not saying you can't make suggestions to them for upgrading the AI, I'm saying that they know already.

What I would like to see is more scripting control over the AI and more ways to override it at the core.

--Ben

Please don't take this personally, but I have a real problem with this kind of attitude. This is like saying: "We are all going to die in the end, so why bother living a good life?"

Yes, most games have crappy AI. Yes, most games suck on other accounts. Does that mean that we should expect ArmA do this as well? I certainly don't think so!

I would like to believe that BIS developers are trying to make the best game that they can within the constraints of their deadlines and budget. It appers that they did not consider AI improvements to be a high priority. Numerous gamers have taken their time to post here (and to other places) in order to tell BIS that they would like to see more emphasis on AI improvements. Hopefully, it is not too late for the development team to evaluate this input and to make some feasible improvements (i.e. VME-like AI at the core level). Now what's your problem with it?

As to your other point about vehicle addons having the same impact on gameplay as AI addons... I don't even want to waste forum space replying to such a misguided notion. AFAIK, lwlooz has already replied to it quite eloquently.

I realize that this post might sound a little harsh. Please don't take it personally. I just find the attitude expressed in your post to be very unconstructive and defeatist.

Peace,

DreDay

[crashdome 3]

So just so I get the general logic of this.You people say that BIS is unable to do AI improvements/tweaking because it is way too complex even for a proffessional coder team that does that for a living and then everyone around here claims that the VME Script AI is the way to go or if BIS just provides us with Script commands,we will be able to do the AI ourselves?

So basically,it was impossible for BIS to do it in the timeframe of the Arma development ,but scripters who do this in their freetime will come up with great AI in no time?

No.  thats not what I am saying atleast. I am also not saying that "oh it's better than anything else so it's good enough"....

No...

I said that it is a complex decision to be made and requires alot of thought and work which is (not that I want it to be) somewhat difficult for them to address in such a short time-frame from now until release without them pressing it back another year.... at that point they might as well just drop ArmA altogether and focus on Game 2 only.

Perhaps I am wrong. Maybe they can fit a complete overhaul of the core AI in 3 months, but from my practical business programming experience I would say that is a laugh beyond laughs. It's not that I wish it to be, I just believe that exposing the methods is far more realistic than an overhaul of the core AI, which is IMO more difficult than you might think.

When you modify a 'core' anything you need to fully test the effects it will have on everything else. It's a BIG project!!! They could have redone it from OFP --> ArmA (and who knows... maybe they are?) but if I have understood every word they have said in official interviews - they have a few months and stated will not do any significant changes.

And I do believe the community can accomplish more. It's been proven... otherwise BIS would have released far more material than the community has - it would have released more patches and more improvements, but in comparison the community has always done more, faster. Maybe it's just my coding nature, but I do like exposure to more internal functions. It allows me to be more creative rather than expecting BIS to live up my dreams for me and crucify them when they don't.

I make my own dreams.

BTW, I have no idea what the VME script is... so none of my opinions are based on that.

[benreeper 0]

DreDay you do not understand my point about the vehicles and it is not important enough to reiterate.

As to BIS not thinking that AI is important enough; just because there are people in this forum (some 20 posters) that demand it to be changed does not mean that the majority of the people think it's a top priority. Oblviously most game developers, very successful ones, also do not think it's not a priority because AI in most games sucks! Did you go E3, well BIS did and they saw what gamers want to buy and play. You may not like it but BIS needs to find a publisher and they would like to make some money also. If they do not, they will be gone. I'm being realistic. There are the things that I would like to have and then there are the things that I can actually get. I have been working on a mod for several years and have stopped it several times because of the AI thwarting me. I kept going because I found work arounds and in the process I've also found out that the AI is STILL the best and it has to work in a huge playing environment that has no funneling.

BIS is NOT a bottomless pit of money that can throw 10 programmers on the AI: they have to devote their time and money to the entire game. Correct me if I'm wrong but most mod teams (for this game) do not have to create and debug the entire game codebase. You make it seem like they are ignoring us. GIVE THEM A BREAK! They owe us nothing and have already given us more than we have ever given them. I brought this game in June 2001 and basically play nothing else.

For the last several years there were many on this board, including me, that said if they would release an update to OFP, that had JIP, they would buy it. Well they are doing JUST that and now it's not enough for us.

If you can name me one game that has AI this good, in MP, with as may AI in the game at once, then I will concede this point and jump on the "Better AI Bandwagon". If you cannot, then you are saying that what we have now is the best and that should be worth something. BIS reads these boards and they have feelings. A little praise and grace goes a long way.

--Ben

[dreday 1]

Benreeper, I am sorry if I have misunderstood your point about the vehicles. I also appreciate you taking your time to express your view in a very well presented and logical manner. However, I still can't agree with most of your points.

I will address your closing challenge first. We both know that it is hard to compare the AI in OFP to any other game because no other game can match the scale of OFP. That is why no one is asking for OFP AI to be as good as the SWAT series or even GRAW. Yet, the scripted AI in the VME mod was a marked improvement over the original OFP AI. So to answer your question, OFP + VME AI are far superior to OFP on its own. Unfortunately, VME AI is somewhat heavy on the system resources, somewhat buggy, and poorly documented. There is only so much you can expect from a mod…

Yet, if the functionality of VME was embraced and streamlined in ArmA, we might have had the perfect solution. I don't know much about the VME designers, but I find it very hard to believe that they have had more time and resources than dedicated professionals at BIS. This leads me to believe that the OFP AI can be improved with only a moderate amount of effort. I seriously doubt that the VME solution is a bottomless pit of development resources that you seem to think it is. I say that as person who has managed software (albeit not gaming) projects for quite a few years.

You seem to believe that consumers don't care about good AI, and that most gaming companies don't care to code good AI. I disagree with both of these points. The AI in games is continuously progressing. Consider this, when OFP first came out it was praised for its AI; while now the same AI is considered to be weak. It is not that the AI has gotten any worse, it's just the other games have raised the bar (regardless of the scale of those games).

Finally I don't see what leads you to believe that I am out to hurt the feelings of BIS developers. I have nothing but respect for them, and you should see that if you re-read my earlier post. However, as a fan of a game, I have a right to express the constructive criticism over some of their decisions. How else would BIS know how we feel about the subject matter?

Peace,

DreDay

[benreeper 0]

You are not, and most users on this board, are not trying to hurt BIS's feelings, what I'm saying is there seems to be more of a climate of "wants" than "thanks".

I disagree about the state of AI in other games. GRAW's was not good and most other games are HEAVILY scripted. Their play environments are small and the player is funnelled along a path. If the game is not exaclty like this, there will definitely not be alot of AI in the action at once because it would slow the game down. Like you said, VME's slows the game dowm. How do you know that this is not the case with what BIS has tried already. AI is heavy on the CPU and most gamers will not appreciate it. They would probably complain about the game being to slow. An ugly game with good AI will get bad reviews and before you say "don't make an ugly game", if it was that easy to make the perfect game why doesn't everybody do it?

I am sorry, consumers do not care about AI when they purchase a game. They want graphics and then maybe physics. This is a fact. AI routines are very expensive in terms of CPU cycles and developers are going to save those cycles for something that will make the game more desireable to the buyer. Console games cannot afford AI at all (there are RARE exceptions) and although this is not a console game, it could be because that's where the money is. It is so mush easier to sell to the console crowd and that is why established developers are abandoning the pc for them.

We can only hope that BIS is not fated to this.

--Ben

[Speeeedy 0]

I am sorry, consumers do not care about AI when they purchase a game.  They want graphics and then maybe physics.  This is a fact.  AI routines are very expensive in terms of CPU cycles and developers are going to save those cycles for something that will make the game more desireable to the buyer.  Console games cannot afford AI at all (there are RARE exceptions) and although this is not a console game, it could be because that's where the money is.  It is so mush easier to sell to the console crowd and that is why established developers are abandoning the pc for them.

Im afraid i can no agree on that. There are customers indeed who will want the high quality graphics and thats the most important thing for them. I mean who doesn't love high quality graphics . However a majority of the gamers luckly still things gameplay is one of the most important factors. Because this is what will keep gamers playing for a longer period of time. If the gameplay sucks the gamers will put the game away soon after buying it (how else would you explain the succesful selling of the GTA series and the dissapointing selling of for instead Quake4).

AI is a important part of this gameplay. The AI can either make or break a game. Depending on the quality of this AI you play a game differently. Take for instead GRAW, while the AI is far from perfect they were able to create good tactical movements and use of cover in this game. Thanks to this it is impossible to go and play RAMBO in this game. You are forced to use tactical movement and plan out your team movements otherwise you won't make it till the end of the mission. The AI makes the gameplay in Ghost Recon.

While AI can take a heavy load (depending on the way its coded) on your CPU. You must not forget that Flashpoint is already 5 years old. In that time the CPU's have tripled there capacitiy and the graphics card even increased more. So thats no excuse on not to include a high quality AI.

While i don't except BIS to create the best AI at the moment. I do except them to increase the AI so we don't just get a newly polished OFP. Because while the people on this forum most likely will buy any BIS game. It would mean the majority of the people (and any game developer wants to sell his game as much as possible) will probably leave it alone which such an terrible AI.

And even if what you say is true regarding the graphics. AA won't sell well if it has to sell from the graphics alone. (While there not bad there neither state of the art)

Regarding your Console commit. I don't play a lot of console games so i can't really answer on your question regarding bad AI's. However i do know many of the successful games are ported to the console and most of the time this only means quality lose regarding the graphics (in the old generation consoles). Also the reasons many of the developers are developing games for the console has nothing to do with cost. In fact the cost of the new generation consoles is more expensive then developing for the PC. The reason is that market for these consoles is most of the time larger then for the PC.

[CameronMcDonald 146]

Nice post, I agree with quite a few points in there.

I really don't mind having to improve the OFP AI via scripting, since I can do that half well, but like many people have said before, this dramatically increases the load on a typical system. And, like Speeedy says, surely the improvement in technology for compies should mean a corresponding increase in AI intelligence?

[lecholas 2]

When speaking about the gameplay. Many of people on this forum say that other games may have a good AI because they have small levels and can be scripted. That's true. But for many end-users who aren't going to make missions for the games it doesn't matter how the good AI is achived, if it is by predefined paths or also predefined but at the other lever - hardcoded. Did you play Brother in Arms Earned in Blood? It's AI was quite good when it comes to tactics. And let's get back to OFP. Did you notice that many mission in it take place at terrains that aren't much bigger than those from GhostRecon or BiA? So it is natural that the AI in OFP is compared to the other games. And for the end-user OFP loses in that contest because for many players it doesn't matter that could operate at bigger levels when it can't operate in squad based levels. Another thing. ArmA as OFP won't be (AFAIK) really simulator of big numbers of troops because of CPU useage and lack of Chain of command between the different groups. So we won't see company scaled battles. We'll se at best platoon scaled ones (I hope I'm wrong but I try not to expect too much from ArmA). And if we are to see platoon scaled ones we are justified to expect that AI on that level could be compared to other games which take place on small levels filled with up to 40 enemies.

Just some of my thoughts.

[firedude_edd 0]

Anyone played the Close Combat series?(NOT FIRST TO FIGHT, but the top-down strategy series). Every single soldier had his own morale, and reacted to everything around them.

-I'd like to see how moving troops move according to the situation. During combat, they rush from cover to cover, and when just patrolling through streets, they stay to the sides of buildings.

-AI uses everything as cover, including vehicles.

-Morale-troops under heavy fire may begin to panic depending on rank, leadership and other things.

-AI will effectively do things by themselves. If a maching gun position is seen, a couple of men will flank it on their own accord, depending on situation.

-Vehicles! Their AI must be improved. Helos under fire may try to regain altitude or even land if things go badly. Tanks need to move more realisticaly. Rather than surge forwards on their own, they allow infantry to scout ahead.

-Possibility to take prisoners. And be one. If the enemy are completely surrounded by a large allied force, and are running out of ammo, they might decide to surrender. Obviously some might decide to die fighting, and form a suicide attack,

-On the defensive. Troops will effectively form a defensive line so flanking enemy will have a hard time.

I have a lot more sugggestions, but that is it for now.

Edit: If there is any heavy combat scripting in the game, like Call of Duty I won't buy the game, as it is no longer a simulation, just an action shooter game.

Also, after reading earlier points about graphics, it is used psychologically to attact consumers to the products. This does'nt affect me for I play really old games too, where gameplay is the most important factor.

[lwlooz 0]

Hi,no offense to anyone,but I believe anybody on this forum who doesn't agree that AI is priority #1 hasn't played OFP a lot (too busy with modding?).

That the "mainstream"-audience doesn't care much about AI is very true,but I never understand BIS move into that market anyways as it is A) saturated and B) the competition is pretty good. Now the "Advanced FPS" or SIM-market,that is us (How many play sims or wargames next to OFP eh? Or do you all play BF and CS?) does infact care about AI,so if BIS considers their "secure" customers,that is us,AI does matter somewhat.

In any case,I said it already,but there is no FPS game out there BIS should look to for AI improvements.

I also get the point that this stuff takes a lot of time,especially a complete AI redesign with lots of added functionalties,but personally I don't know if it is impossible to improve the already existing core AI.

OFP AI already knows to look for cover in some degree,they already know when to lie down or take a slower approach,they even must know if they are up against a strong enemy,because they try to flank you. It is just that those procedures are somewhat slow and faulty at times (Resulting in AI running straight in or past you). So,improve the already existing squad and individual AI and vehicle AI(which should be possible with the extra CPU - afterall OFP's AI is already progressive,their is just a upper limit of their smartness).