PDA

View Full Version : BoB And Beyond --- will AI Bubble Range be option for customers?



LEXX_Luthor
09-13-2006, 01:24 PM
Can the AI Bubble range -- the distance from the player's aircraft where the Simplified flight models and air combat takes place -- be changed by the player, or server if AI Bubbles are applicable to Online play?

Explanation:: To get very large AI aircraft formations for BoB And Beyond, Oleg said he will use AI Bubbles similar to Rowan's AI Bubbles in their BoB and MiG Alley sims. The AI Bubble is a sphere in space, centered on the player's aircraft, beyond which the AI aircraft use simplified flight models and (possibly) simplified damage/weapon modelling and simplified AI tactics, or even maybe a purely "roll-the-dice" statistical numbers combat. The player can not see this from the cockpit as its designed to happen beyond visual range, and Oleg said he will not allow External View, at least of aircraft, to work outside the AI Bubble far from the player's aircraft.

Now, as computers gain more power over the life of BoB And Beyond (the "beyond" part after BoB), its possible that more and more of the air war can be simulated by using larger AI Bubbles and so less simplification. This is like how FB 1.0 clouds and Excellent terrain mode grafix challenged the best computers 4 years ago, but today's computers are far better than even today's FB/PF 4.0 New Clouds and Perfect mode grafix. The same could happen with AI Bubbles during the life of BoB+, as computers in a few years may be able to run more of the air war simulation without needing as much simplification in FMs and possibly DMs/WMs.

If so, can customers increase the AI Bubble range depending on their computer hardware, as he/she upgrades over the next few years. If not, we will have to depend on Oleg to increase the AI Bubble range....if this even makes sense to the AI Bubble concept, as it has never been explained to customers by Oleg and Rowan/Shockwave. I *think* its possible that AI Bubbles may give non-sensible results in dynamic campaigns, and Online Wars if applicable, as the combat beyond the AI Bubble is abstracted. There have been poasts about strange campaign results at the Shockwave's BoB2 board that I can only think are caused by Rowan's AI Bubbles. I assume its possible that the entire air battle is one big AI Bubble when the player does not fly and plays the role of air force commander alone. This is a very interesting subject and we would love to hear more about it. Its possible that the the AI Bubble concept is restricted to only bombers in formation, which could be a good thing. I dunno.

ThankS!

alert_1
09-14-2006, 04:47 AM
I hate bubbles, they can degrade even pretty realistic sim pretty badly...I hope that SoW wont be the case though

joeap
09-14-2006, 06:44 AM
Very thoughtful "poast" as usual Lexx...let me add something, doesn't the Falcon series use AI bubbles? Isn't it universally praised for its excellent dynamic campaign? I think abstraction can work in some situations of the data and probablities are carefully tweaked. I have played some old games, like Steel Panthers and Fighting Steel that used abstract methods (based on WWII records in part) that got realistic results. If implemented carefully it could work.

Capt.LoneRanger
09-14-2006, 07:40 AM
Yes, Falcon4 used this method and it was really successfull with it - as you allready stated, Joeap.

EAW used it, too, and, just to shock you, IL2 does, too. Just not for AI at the moment. But if you fly at 15.000m you'll see the nice vision bubble around you. Even when approaching a city, you'll see the effects of that bubble-system, as details are only visible, after you approached them to a certain degree.

At this state, you cannot even dare to think about how far the AI-Bubble (and this is all your post seems to be about?) will be and what effect it will have. From what we know now, this bubble enables BoB to have a whole war running, just like in F4. And how many units does F4 simulate? 30.000? What would it help you, to be able to view every singe one at one time?

joeap
09-14-2006, 08:25 AM
http://forums.ubi.com/images/smilies/agreepost.gif Exactly, we don't need to see every unit, just those in the combat area. As I said it would need tweaking, the bubble that exists in IL2 is not ideal for medium or dive bombers as it is too hard to see targets on the ground IMO. Although IRL that would vary.

LEXX_Luthor
09-14-2006, 10:48 PM
No -- flight model and AI tactic bubbles, not just grafix draw distance.

joeap
09-15-2006, 01:57 AM
That's what I understood and I agree again.

Bearcat99
09-15-2006, 05:54 AM
What I am wondering is if the new AI SEEK (http://www.aiseek.com/AI_acceleration.html) technology will possibly be implememnted into BoB. It looks promising.. and from what I have seen from 1C... they like to build for the future.... look at this engine... which is long in the tooth but still robust, so perhaps one of the things that may extend the release time for BoB to our benifit is making sure that is is at the front of the wave... and not at the back of it. That AICPU could solve some of the AI issues.... hopefully anyway. We shall see..... Download those vieos and pay attention to the AI.... I know many have already seen this.. but the first time I went there I just read the info.. I didnt DL the videos. Interestingly promising stuff..... It also seems as if the Tank battle videos are using the PPU technology.

I see the day when PPU and AIPUs will be onboard as more and more software is written to exploit thier features.

Aviar
09-17-2006, 01:11 AM
Just wondering how this AI bubble would work online.

I understand how it works offline with one Player, but how about an online coop with 20 'Players'?

Aviar

blackTIE
09-17-2006, 10:45 AM
I hope it is build with AISeek in mind, but not that it's necessary. It would be nice to up the framerate in a later stage just by adding a AISeek Intia AIPU.

Oleg and 1C:Maddox are known for their forward looking engine and it would be nice to just put in a PhysX PPU and Intia AIPU, but my system doensn't have the PCI-slots for it. It woult be nice to have the GPU, PPU and AIPU all on one card at least.

[update] AI and physics seems to have a lot in commong according to the samplevideos. Maybe AISeek en AGEIA can join hands to make a killer product. The porting of PlayStation 3 titles woudl be easier in the future with that combination.


Originally posted by Bearcat99:
What I am wondering is if the new AI SEEK (http://www.aiseek.com/AI_acceleration.html) technology will possibly be implememnted into BoB. It looks promising.. and from what I have seen from 1C... they like to build for the future.... look at this engine... which is long in the tooth but still robust, so perhaps one of the things that may extend the release time for BoB to our benifit is making sure that is is at the front of the wave... and not at the back of it. That AICPU could solve some of the AI issues.... hopefully anyway. We shall see..... Download those vieos and pay attention to the AI.... I know many have already seen this.. but the first time I went there I just read the info.. I didnt DL the videos. Interestingly promising stuff..... It also seems as if the Tank battle videos are using the PPU technology.

I see the day when PPU and AIPUs will be onboard as more and more software is written to exploit thier features.

LEXX_Luthor
09-17-2006, 06:50 PM
Thanks everybody. AI Bubbles don't get the attention they need from the community, and the developers are not talking; Rowan/Shockwave included.


alert_1::
I hate bubbles, they can degrade even pretty realistic sim pretty badly...I hope that SoW wont be the case though
Tell more. I never had a sim that used AI Bubbles.


joeap::
...doesn't the Falcon series use AI bubbles? Isn't it universally praised for its excellent dynamic campaign? I think abstraction can work in some situations of the data and probablities are carefully tweaked. I have played some old games, like Steel Panthers and Fighting Steel that used abstract methods <span class="ev_code_yellow">(based on WWII records in part)</span> that got realistic results. If implemented carefully it could work.
We can't ever *know* because we can't see it in action beyond the AI Bubble's range. I have seen some reports of "weirdness" in Falcon's dynamic campaigns. I don't know. Recall, Oleg says he will disable External View for all aircraft outside the AI/flight model Bubble. Thus, we can't know, not even for experiments that are needed by 3rd Party developers of either static or dynamic campaigns.

<span class="ev_code_yellow">Abstraction = Realistic/Historic Results</span>

I fear the abstraction will be a mere formula calculation, based upon Historical combat ratios -- if so YES there will be "realistic results" -- 100% perfectly historical realistic results, every time the campaign is played. Perhaps there will be some statistical randomness thrown in to the formula. Who knows? We don't.

The community has always asked for "realistic" and Historical Combat Results(tm) from flight models alone.

They may finally get it, in a maths formula. Congratulations.

Honestly, I don't know how AI Bubbles are made so my comments are mostly worst case fears for dynamic campaign systems. We would just like to know, and to see what happens in External View beyond the AI Bubble -- if that's even possible as the AI action may actually not happen at all, except in an "abstract" AMD/Intel cpu math calculation.

heywooood
09-22-2006, 08:43 AM
I fear bubbles too...What will AI planes be up to outside of my bubble?...and why can't I see?

I like to watch, rumor has it, and if I cant start a mission, throw my plane on auto pilot, and go switching views to see what exactly is going on all over the map well....I just don't think I'll like it so much - yet another reason to suspend belief and lose immersion in my virtual experience.

How big is this bubble? Will AI planes drone into it - activate and participate in combat - then hit the invisible bubble exit wall and resume drone flight?....

Will it also affect ground equipment like armor, ships and trains?....How will their behavior change? How will it resume 'poast bubble'....please lettuce no.

Tater-SW-
09-22-2006, 10:38 AM
I actually LIKE the notion of realistic outcomes, even if by statistical methodology. From a player standpoint, YOU are the only one who should realy be able to alter the historical stream by some little amount.

What we have now, is (flawed) AI duking it out to the death while you do your thing. A classic example would be to build a historical airbase (with a proper number of AAA gun tubes), and have AI attack it with B-25s or something like that. See how many make it back. The 5th AF would have been a nonentity in a couple days with PF AI fighting the war. The AI just isn't up to the task unless it is better in a quantum way in BoB.

The real issue was raised by Aviar. Presumably in a coop AI bubbles would be 1 per player, and only stuff outside ANY AI bubble would be done in the simplified way.

tater

LEXX_Luthor
09-22-2006, 05:23 PM
Tater, about the B-25s bombing -- what you are asking for is more realistic AI behavior, which has to be programmed. AI Bubble math formula is not what you are asking for, and I fear that any "historical" statistical math formula is a quick escape from having to program AI behavior suitable for Air War simulation.


Be careful Tater, historical combat results work only for pure historical "re-enactment" Static campaigns and, if we dare say it, Static Online War with a fixed outcome every time the War is played on the server.

Oleg/UBI can pay ME a Monthly Fee and I still won't join a Static Online War. http://forums.ubi.com/images/smilies/51.gif

If the Dynamic Campaign or Online War begins to differ from historical correctness, combat results should not always follow historical combat results.

heywooood
09-22-2006, 08:15 PM
LeXX is trying to burst our bubble....first he can't see why a flight sim (Perfect Flotsam)needs perfect water...and now its all about BoBs bubbles... what next I ask you, what?

If histerically accurate timelines and concrete mission outcomes are in our future/past with BoB - and if these scripted campaigns are relegated to perfunctory mathematic calculations everywhere but where the players bubble of influence occurs...well then it may only end up a cinematic experience - very pretty to look at but not very interactive.

Somehow I just can't see it happening.

Tater-SW-
09-22-2006, 09:31 PM
Originally posted by LEXX_Luthor:
Tater, about the B-25s bombing -- what you are asking for is more realistic AI behavior, which has to be programmed. AI Bubble math formula is not what you are asking for, and I fear that any "historical" statistical math formula is a quick escape from having to program AI behavior suitable for Air War simulation.

I'd say that the AI is unlikely to be good enough to trust to produce historical results. Flight sims tend to be overly lethal, IMO.



Be careful Tater, historical combat results work only for pure historical "re-enactment" Static campaigns and, if we dare say it, Static Online War with a fixed outcome every time the War is played on the server.

Oleg/UBI can pay ME a Monthly Fee and I still won't join a Static Online War. http://forums.ubi.com/images/smilies/51.gif


I don't think campaigns should be entirely dynamic. To do so as you suggest requires modelling every individual ground unit---including infantry---at the same level as aircraft (including far better AI than PF currently has). As much as I really like air combat, it could shape events, and even be decisive now and again, but the bulk of the work of moving the war forward was done on the ground.


If the Dynamic Campaign or Online War begins to differ from historical correctness, combat results should not always follow historical combat results.

No, individual battles might differ, so what. The overall course of the war, however, would not change. You can have a dynamic war in the PTO, for example, but if there is ever an end result where Japan wins the war, it is broken.

I have no problem with a toggle for range of the bubble, but i'd want to see tests. If the results outside the bubble are statistically on par with reality, and the AI vs AI battles are not, then the latter are broken and better done without.

tater

heywooood
09-23-2006, 10:15 AM
Should the player be able to alter the outcome of the Battle in campaign mode?

Is that the real question here?

Lets say that, in the manner of a strategy game, I could command a squadron of jabos and attack only airfields and radar stations thus altering the outcome of the conflict..of course this would have to be done contrary to the scripted orders and mission objectives given prior to each mission. Would that be a good thing?

On the other hand - what if, regardless of the players prowess in achieving or exceeding mission objectives (or as in my case failing miserably at every turn), the war should go on regardless...as if the player had not participated at all?

Is this really about AI bubbles of awareness or is it rather about being able, through ones ability in the game, to alter or delay the outcome...so that the inevitable allied victory becomes somewhat less 'evitable' ?

Seems like this is more about mission editing than anything else. If the BoB simulation is to be accurate, the Battle must go according to history, regardless of the players achievements or failures.
On the other hand, if it is to be a combat flight sim based on the BoB but only using that as a backdrop, then the player should be able to alter the outcome by his involvement.

Tater-SW-
09-23-2006, 10:22 AM
Any attempt at global changes in a conflict based on the actions of a single player (or group of players), even if you control a whole squadron or even a wing, are pretty silly, IMO. It would also require a high-fidelity economic model, etc. It just gets too crazy. Political decisions are outside the realm as well. So the crazy leadership decides to change the targets to the cities in retribution, you do as ordered or face a firing squad.

tater

heywooood
09-23-2006, 11:05 AM
yeah tater I agree....however, this seems to be one more seperation between our genre and the general computer game/simulation marketplace.

Its like the Harlem Globetrotters vs the Washington Generals every night.

In other sims you can have varying degrees of impact in the outcome but in this more rigid environment you have to have a different view of what success is.

In other words - I can fly up in my Hurri and see a formation of 150 Heinkels or Dorniers and and say to myself "well lets see if I can get one before I get hit" and then I get hit and as I spiral down past those excellent renderings of WWII bombers and then see those beatiful white Cliffs and then splash down into that perfect water I can say to myself "wow, that was spectacular - lets do it again"...

On the other hand - most gamers will be dissapointed when they have a hard time successfully downing one plane - let alone getting back to the field and landing.

I think that narrows the market even more.

All of this would seem to put more importance into the aspects of the FMB and a dynamic campaign generator than AI bubble theory wouldn't it - I mean who cares what the AI equipment is doing if you can't see it from inside the bubble.

LEXX_Luthor
09-23-2006, 04:59 PM
Tater::
No, individual battles might differ, so what. The overall course of the war, however, would not change. You can have a dynamic war in the PTO, for example, but if there is ever an end result where Japan wins the war, it is broken.
The astronomical difference between USA and Japan as military powers means there can be no real alternative to the "end of the war" even through dynamic combat simulation, so this is a poor example to use. Only a Dynamic Campaign can simulate a possible (non-historical) 1944 or 1946 ending of the war, and in both cases combat results should generally be different than historical.

What we are talking about is, for example, during a dynamic campaign or Online War, Japanese *may* have an opportunity of attacking, taking, and holding (for a while) Fiji, as Fiji was a part of the historical plans of the Japenese -- if you are interested in researching, studying and having fun and enjoy simulating history beyond simple Static re-enactment of history. This is why essentially "AI cheating" with fixed historical Math formulae won't work.

A better "end of war" example to use here would be a dynamic campaign of RAF vs Luftwaffe, which could have gone a bit different, and would need no ground units fighting each other (AI ground anti-aircraft, searchlights, radar, intercept controllers, etc... are still needed.

No, for any Air War simulation, ground units don't have to be simulated "at the same level" as aircraft. These can be simple for an Air War simulation -- don't confuse this with Ground War simulation of the ground combat sims where the ground units are detailed, and air units simplified...and oh are they simplified in the ground sims http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif


Tater::
Any attempt at global changes in a conflict based on the actions of a single player (or group of players), even if you control a whole squadron or even a wing, are pretty silly, IMO.
A realistic behaving dynamic campaign or Online War does not "cheat" this way, and you know this.

LEXX_Luthor
09-23-2006, 07:15 PM
Perhaps the best demonstration of the need for options to create and play fully dynamic campaigns and fully dynamic Online Wars is simulation of wars that never happened. Some may find extending WW1 into 1919 to offer a most intense World War 1 simming option, particularly regarding advanced designs for ground attack aircraft and plans to use them massively, that were, fortunately, cut off early by the war's end in 1918.


Others....an air war at the time of the Berlin Airlift, or Cold War era Strategic Air Command vs Soviet PVO air defense which I find the most fascinating historical military aviation subject. The new Wings Over Europe by ThirdWire offers Soviet invasion of West European NATO from early 1960s to late 1970s, and is one PC combat flight sim that can't depend on Perfect Historical Combat Results.

These simulated wars would have no "historical combat results" to use, so they would have to be simulated dynamically.

Tater-SW-
09-23-2006, 08:18 PM
NO they wouldn't. A statistical method would still be far more accurate than the AI unless it is literally orders of magnitude more sophisticated than what we have in any version of il-2.

The AI just isn't good enough, sorry.

Historical results doesn't require a lookup table vs some actual battle, it just require ballpark estimates based on similar combats/combatants. The board and miniatures gamers have been doing this for decades, and they can produce pretty reliable results (Harpoon is a decent example).

The ground unit example was for the vast majority of combats, not the special case of the BoB. Had the Germans won the BoB, then what would the online war do? What would the abstraction of Sealion be?

The issue at hand is this AI bubble. The reality is that either the AI is up to snuff, or it isn't. If the AI does a decently good (meaning realistic historically) job, then fine, no bubble needed, use AI. The tests are simple enough to do. This requires that the AI use realistic doctrine, staying where ordered to be, for example, and not chasing contatcs all over down to the deck when they should follow orders and stay where they are ordered to be.

I'll believe the AI is good enough when I see it. When I see planes make a single pass on a ground target, they get the $#%@%$# home. I'll believe it when stacked air defense stay stacked, when close escorts escort closely, etc, ad nauseum. Hell, how about not just flying whole flights into mountains? Hard to have the AI decide the fate of the BoB if whole flights will crash because the AI is poorly modelled.

Sorry to say, I don't believe the AI will be good enough.

tater

Tater-SW-
09-23-2006, 08:19 PM
BTW, in addition to having doctrine coded for AI, we need things like morale.

tater

LEXX_Luthor
09-23-2006, 09:05 PM
Tater, those are great ideas on needed advances in cfs AI programming. Thanks.

LEXX_Luthor
10-20-2006, 09:46 PM
*bump*

We need to know more about this issue.

Tater::
I'll believe the AI is good enough when I see it. When I see planes make a single pass on a ground target, they get the $#%@%$# home. I'll believe it when stacked air defense stay stacked, when close escorts escort closely, etc, ad nauseum. Hell, how about not just flying whole flights into mountains? Hard to have the AI decide the fate of the BoB if whole flights will crash because the AI is poorly modelled.

Sorry to say, I don't believe the AI will be good enough.
Agreed. Those are some of the ideas that sims need to be successful in the long run. AI programming is always in need of improvement. Interestingly tater, you assume the AI can never be "good enough" for the distant combat, but you accept that its "good enough" for the close combat the player is involved in within the AI Bubble.


tater::
Historical results doesn't require a lookup table vs some actual battle, it just require ballpark estimates based on similar combats/combatants. The board and miniatures gamers have been doing this for decades, and they can produce pretty reliable results (Harpoon is a decent example).
Look-up table is the same thing no matter if the specific numbers are taken from historical battle or "ballpark" guestimates. You know this, I know this, the community here knows this. Board and Miniature gamers don't have "AI" as their entire combat simulation is dice rolling from lookup tables. The agreed upon movement rules can be considered to be somewhat like AI, but the humoid player is generally his/her own AI in these old style board/miniature games.

We can agree! AI is always in need of improvement, and we also know that AI programming is the most difficult thing in any PC game -- far more difficult than even the famous "3D modeling" that gets the most attention, at least until the customer tires of the Perfect Polygon grafix and leaves the game on the shelf -- for games requiring steep learning curves such as combat flight sims.<div class="ev_tpc_signature">

__________________
http://forums.ubi.com/images/smilies/10.gif Flyable Swedish "Gladiator" listed as J8A...in FB Gold...and...Aces Expansion Pack

"You will still have FB, you will lose nothing" ~WUAF_Badsight
"I had actually pre ordered CFS3 and I couldnt wait..." ~Bearcat99
"At the altitudes this community flies at, diving is not an option." ~Stiglr
"Gladiator and Falco, elegant weapons of a more civilized age" ~ElAurens
"109Z flew briefly, after being hit by a bomb. Go-229 also saw combat, when the factory was overrun." ~pingu666
:
"Where you did read about Spitfire made from a wood?
Close this book forever and don't open anymore!" ~Oleg_Maddox http://forums.ubi.com/images/smilies/16x16_smiley-surprised.gif

Vidarts
10-21-2006, 07:08 AM
Well, I am not so much in favour of a too simplistic AI FM even outside my visual range. There are a couple of people who like to do little movies (doing nice advertisement for the game, too http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif ). To do this, they need the possibility to switch to outside view of AI aircraft from time to time. Following Oleg this won't be possible in the future. And even if it were it would be not nice do watch unrealistic ufo bejaviour.

And I agree to some others: if I consider what effort is put into modelling ground units I think the effort should be put rather in aircraft modelling. It's a flight sim and not Battle Field or something like that ...<div class="ev_tpc_signature">

Vidarts
Life is just another bugged script.

tagTaken2
10-21-2006, 08:43 PM
I'm playing MiG Alley at the moment. The AI bubble you are referring to is switchable, nicht wahr? In the preferences page, there's a complex AI off/on tickbox. Assuming that we are talking about the same thing, I'd suggest that it could simply be a toggle, or even a line of code changed in a couple of years, when we're all up to speed.

LEXX_Luthor
10-21-2006, 09:48 PM
http://forums.ubi.com/images/smilies/25.gif

Thanks tagT, that does sound interesting. Are there any experiments that can be performed to see if the toggle is for AI Bubble on/off or for selecting AI complexity within the player's AI Bubble?

Do you have full external view of all aircraft flying everywhere during a MiG Alley game run?<div class="ev_tpc_signature">

__________________
http://forums.ubi.com/images/smilies/10.gif Flyable Swedish "Gladiator" listed as J8A...in FB Gold...and...Aces Expansion Pack

"You will still have FB, you will lose nothing" ~WUAF_Badsight
"I had actually pre ordered CFS3 and I couldnt wait..." ~Bearcat99
"At the altitudes this community flies at, diving is not an option." ~Stiglr
"Gladiator and Falco, elegant weapons of a more civilized age" ~ElAurens
"109Z flew briefly, after being hit by a bomb. Go-229 also saw combat, when the factory was overrun." ~pingu666
:
"Where you did read about Spitfire made from a wood?
Close this book forever and don't open anymore!" ~Oleg_Maddox http://forums.ubi.com/images/smilies/16x16_smiley-surprised.gif

tagTaken2
10-24-2006, 04:01 AM
Not sure about the first, I only found the ground padlock while looking at view options (thanks, could never pick out the target http://forums.ubi.com/images/smilies/metal.gif).

I don't believe you can get full external- as in FB, Ctrl-F2.

If you're interested, the people at BDG might be able to comment, if they can be found... dead links. They worked on the game, and also on BoB2.

@Vidarts... If it's fifty miles away, who cares? Playability should come first and foremost. As far as movies go, I suspect that the camera will have a "bubble" around it when moved.

F16_Sulan
10-29-2006, 07:25 PM
Whats the big deal?? The bubble is neccesary to simulate the battles on a grand scale, when computers are up to speed I´m sure we´ll be able to extent the bubble.

I don´t see it as a problem for doing "what-if" scenarios either, because the historical statistics way of deciding the outcomes outside the bubble equals a perfectly historical flightmodel.

I like playing "what-if" scenarios, I´d like a way to tip the scales alittle, or have some influence on strategy to change the course of the war.

The bubble is a good thing if used well!!<div class="ev_tpc_signature">

-Gravity never looses, the best you can hope for is a draw

mazexx
11-01-2006, 01:35 PM
Well, I'm all in on bubbles. When running a game engine at 50 fps you don't have many milliseconds over for deciding if a Stuka 20 miles away actually did stall due to the harsh stick movement by the AI pilot that was nervous after seeing tracers passing his cockpit...

Where are those milliseconds best spent, on making your flight model feel right or his? If an unexperienced Stuka pilot is attacked by 6 Spitfires the AI could simply let the Stuka explode and make a quick "dice throw" to see if it by a calculated percentage got one (or even two) of the Spits with a lucky shot by the gunner before crashing. Battle over - a few CPU cycles spent. Instead it could calculate all the FM data for the battle and hopefully come up with the same conclusion. Millions of YOUR CPU cycles spent.

With the bubble version the CPU cycles could be spent on the planes in your vicinity, to make them consider protecting their wingmen, reacting credibly to tracers passing their cockpit etc... There could be hundreds or thousands of other planes in the air far from you that where statistically controlled - but with a rather uncertain "luck" factor. If you fly 20 miles north and arrive at the scene where the 6 Spitfires are just catching up the Stuka you could interact, and they are inside the bubble by then so the "flight AI" takes control of them.

In the "no bubble" solution there would be no one there as it would kill your fps to have a thousand planes fully AI controlled. If we would try to squeeze a thousand planes in, the AI would have to be really stupid due to the few CPU cycles available for each of them...

/Mazex

LEXX_Luthor
11-01-2006, 08:36 PM
Some good points.

But, why won't we be allowed to have any external view of AI aircraft outside the AI Bubble, if its all "okay?"

Something is wrong here. http://forums.ubi.com/images/smilies/51.gif<div class="ev_tpc_signature">

__________________
http://forums.ubi.com/images/smilies/10.gif Flyable Swedish "Gladiator" listed as J8A...in FB Gold...and...Aces Expansion Pack

"You will still have FB, you will lose nothing" ~WUAF_Badsight
"I had actually pre ordered CFS3 and I couldnt wait..." ~Bearcat99
"At the altitudes this community flies at, diving is not an option." ~Stiglr
"Gladiator and Falco, elegant weapons of a more civilized age" ~ElAurens
"109Z flew briefly, after being hit by a bomb. Go-229 also saw combat, when the factory was overrun." ~pingu666
:
"Where you did read about Spitfire made from a wood?
Close this book forever and don't open anymore!" ~Oleg_Maddox http://forums.ubi.com/images/smilies/16x16_smiley-surprised.gif