PDA

View Full Version : random flak?...i don't think so



PaleRidur
11-26-2005, 12:07 PM
I hardly ever post cause life is too short to waste it on forums. But this i just have to ask/post here. I often build small coop missions and i'm often annoyed by the mission editor in general and the lack of a randomize option. Anyways...i started placing friendly flak/aa near the enemy airfield hoping this would result in a random number of enemy aircraft surviving takeoff so the mission would be slightly different each time. But man was i wrong. The same things kept happening again and again...same plane got shot down...same damage..etc..etc. So i ended up with a tiny test mission and you can run this mission as many times as you like but the same things will happen like clockwork. The exploding shells near the plane will go of at exactly the same spot and time. The plane will get damaged in exactly the same way each time. While the plane tumbles down a few other parts will get shot off...always the same parts.
Now one might argue, this is one plane versus quite a few 20mm flak objects, so it's logical it gets shot down. I agree that the chance of getting shot down is high in this case...but still this should not result in something which is as predictable as this is.
http://www.xs4all.nl/~wllm/samples/notrandomattall.mis

Cadet_Bobo
11-26-2005, 01:03 PM
Since AI runs on routines, I suppose it is possible that the same gun will shoot down the same AI plane every mission if they do exactly the same thing without any varience. The routines begin at the same time and run like clockwork http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif If you could randomize the routine it would be different every run. Just my thoughts, I don't really know anything http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Bobo

LEXX_Luthor
11-26-2005, 05:21 PM
Try changing the flak gun positions each time. The script differs if the position changes.

I first noticed this in old AI vs AI dogfight tests made in FMB. Each time I ran the missions and watched from External View, the exact same thing happened...same exact moves and shots taken, and naturally the same exact results.

I used only a start and end waypoint in these simple tests. However, if I slightly moved the start or end waypoint of the planes, the dogfight would be entirely different. Even changes as small as a few meters makes things come out different by the end. Try moving the flak a few hundred meters until you get the results you desire.

The problem you face is that the aircraft take off in a very slow sequence one after the other. So, all aircraft will get shot at. If they are not destroyed, they can be damaged and this includes fuel leak (white vapour trail) and they run out of fuel after a few minutes. If you use the open terrain takeoff than the planes take off in formation. Be aware they you need a big flat area for they are widely spread apart when they spawn on the ground.

Chuck_Older
11-26-2005, 07:22 PM
Originally posted by PaleRidur:
I hardly ever post cause life is too short to waste it on forums.

Good intro for asking for assistance! http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

"Well, you guys are losers for spending time here, but could you help me out anyway?" http://forums.ubi.com/images/smilies/88.gif

PaleRidur
11-27-2005, 03:05 AM
You're paranoia chuck, but your reply is exactly the reason why i spend little time in forums. Do you see me asking for assistance?...i was pointing out something which strikes me as rather important. The only reason i mentioned the time i spend on forums was to prevent people like you from going "he has only 7 posts so he must be a total n00b".


Originally posted by Chuck_Older:
Good intro for asking for assistance! http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

"Well, you guys are losers for spending time here, but could you help me out anyway?" http://forums.ubi.com/images/smilies/88.gif

PaleRidur
11-27-2005, 03:37 AM
Originally posted by LEXX_Luthor:
Try changing the flak gun positions each time. The script differs if the position changes.

I had tried this before posting here, but it changes little.
With flak in position X -> plane 2 would get shot down the same way with the same damage each and every time. Change flak to position Y -> plane 1 would get shot down each and every time. So if you are after a fixed result you can mess around with positions/waypoints...but if you want a random result it is of no help.

I will try the wide spread take off pattern but i expect this will not change anything in terms of resulting in planes getting hit/shot in a random pattern.

And in general, how can something be shot down in exactly the same way with same damage, same flak hits and so on if the flightpath of the AA shells is supposed to be different each time. To me this basically proves that shells, at least for the AA, are not calculated "in flight". Instead the engine seems to do a pre calculation going: "The target will get hit here here and here..it will get this kind of damage..i will generate explosions here here and here at these points in time/space". Which in my opinion is not great.

JtD
11-27-2005, 06:04 AM
The AI will do exactly the same as long as there is the same input. This is how the trk tracks work. This also means that the AAA will go through the same script over and over. So the same hits occur, and the pilots react the same way.

Chuck_Older
11-27-2005, 07:43 AM
Originally posted by PaleRidur:
You're paranoia chuck, but your reply is exactly the reason why i spend little time in forums. Do you see me asking for assistance?...i was pointing out something which strikes me as rather important. The only reason i mentioned the time i spend on forums was to prevent people like you from going "he has only 7 posts so he must be a total n00b".

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Chuck_Older:
Good intro for asking for assistance! http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

"Well, you guys are losers for spending time here, but could you help me out anyway?" http://forums.ubi.com/images/smilies/88.gif </div></BLOCKQUOTE>


Three things.

1) Yes. Yes I DO see you asking for help. Read your second sentence for yourself

2) How many little smiley laughing guys would you like me to post next time to prove to you I'm not being hostile?

3) I am not now, nor ever have been, mean to "n00bs" and I think post count is BS as a way to value an opinion. Here's the last time I talked to a "N00b":
http://forums.ubi.com/eve/forums/a/tpc/f/26310365/m/6941066083

That's how I treat "N00bs" even though "life is too short to waste it on forums"

Now take you for example. You have a low post count. That means nothing to me. What I do is look at what you post. I read it. So my opinion of you, for example, is of somebody who is very awkward in communicating, because he assumes a lot, thinks he knows the rest, doesn't ask for clarification if he doesn't understand, and also doesn't realise the impact of what he posts. You don't know that's an odd way of asking for things, and you don't know me, and you don't even know you asked for anything. Not my fault, any of it. You assume way too much. If you need help, I'll give it, and I won't be mean about it. But I'll set you straight right now: you're the one acting "paranoia". Being paranoid means you think everyone's against you. That's what this:

"your reply is exactly the reason why i spend little time in forums. Do you see me asking for assistance?...i was pointing out something which strikes me as rather important. The only reason i mentioned the time i spend on forums was to prevent people like you from going "he has only 7 posts so he must be a total n00b"."

means, after all. You're afraid that everyone's going to jump all over what you post. You see my reply and assume that the big attack has come. That's paranoia, my friend.

Deadmeat313
11-28-2005, 03:57 PM
Paleridur. Check out Flying Nutcase's work-in-progress in Mission Builders.

http://forums.ubi.com/eve/forums/a/tpc/f/50910533/m/3401024833

Regards

T.

BfHeFwMe
11-30-2005, 10:35 PM
Clashaholic, it's showing again. http://forums.ubi.com/groupee_common/emoticons/icon_rolleyes.gif

Flying_Nutcase
12-01-2005, 07:07 AM
PaleRidur, I feel you pain.

As linked a couple of posts up, I'm working on a software utility caleld IL2 Fog Of War with the basic features doing exactly what you're looking for.

For example changing the number of aircraft in a flight, changing the number and position of static obejcts such as AA positions, static aircraft at an airfield etc etc (etc). Oh, including random enemy CAP flights so you'll always need to look over your shoulder. http://forums.ubi.com/images/smilies/59.gif

Failing any calamity, it should be around - at least in a basic form - by Christmas.


~S~

P.S. I've also done some testing and seen the exact same occurence and reoccurence of an event, so one feature of the non-stationary object module will be to have a 'jiggle parameter which will change waypoint positions by a user-defined amount - say 10 meters. That should be enough to avoid reoccurrences like you've experienced.

PaleRidur
12-01-2005, 10:03 AM
Flying_Nutcase, very much looking forward to what you come up with. I'm really craving for some form of randomness.

Cheers

Waldo.Pepper
12-01-2005, 10:20 AM
Originally posted by PaleRidur:
Flying_Nutcase, very much looking forward to what you come up with. I'm really craving for some form of randomness.

Cheers

Have you tried (in addition to moving the placement of the flak guns) rotating them in place? Perhaps away from the target airplanes...so that before they commence their hellishly accurate fire they first have to rotate...a variable amount of degrees?

Who knows...might introduce some variability.

PaleRidur
12-01-2005, 01:50 PM
Originally posted by Waldo.Pepper:

Have you tried (in addition to moving the placement of the flak guns) rotating them in place?

Who knows...might introduce some variability.

Yeah, didn't make a difference. I've now resorted to using werbelwinds instead of static flak and have them driving around. This does seem to have a slighty random effect even though all routes remain the same each time the mission is run. I still have to really see how random...but sofar i have seen different results a few times

Chuck_Older
12-01-2005, 03:36 PM
Originally posted by BfHeFwMe:
Clashaholic, it's showing again. http://forums.ubi.com/groupee_common/emoticons/icon_rolleyes.gif
Oh please. Here's an idea- mind your own business for a change

LEXX_Luthor
12-01-2005, 05:01 PM
Pale::
how can something be shot down in exactly the same way with same damage, same flak hits and so on if the flightpath of the AA shells is supposed to be different each time.
In your intro physics, when you input the same numbers into your trajectory formula, you get the same answer. You have to change the numbers you put in if you want a different answer. mmm, maybe also the damage hit boxes are really big, so you can move the hit spot but still be inside the same damage hit box so cause the same damage. Stuff like that

Flying_Nutcase
12-02-2005, 08:41 AM
I did some pretty methodical testing today and a small difference in height (e.g. 1 meter)or a small difference in speed (depending on the length of the waypoint, even 1 km/hr) or a small change in x and/or y-coordinates (a few meters?) resulted in a complete miss.

This is great news, as IL2 Fog Of War's non-stationary module will have a x/y jiggle, an altitude jiggle and a speed jiggle, thus eliminating all this predictability with AI behavior.

Hooray to that. http://forums.ubi.com/images/smilies/heart.gif

stubby
12-02-2005, 12:28 PM
Originally posted by Chuck_Older:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by PaleRidur:
I hardly ever post cause life is too short to waste it on forums.

Good intro for asking for assistance! http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

"Well, you guys are losers for spending time here, but could you help me out anyway?" http://forums.ubi.com/images/smilies/88.gif </div></BLOCKQUOTE>

I thought the same thing. PaleRidur dude lacks any and all skills in basic diplomacy. But then again, it's good troll bait.

Aviar
12-02-2005, 04:34 PM
Personalities aside, I don't think this is a troll thread in the least.

I noticed this aspect of the game years ago, even in the original IL-2. In the course of making missions, much testing is usually in order. One can easily run parts or all of a mission dozens of times in order to get a desired effect.

While doing this, I saw that missions (with very few exceptions) would literally run exactly the same way every time....basically like a movie. Plane 'A' would get hit by flak 'B' at exactly the same time and place. Plane 'C' would always attack target 'D' and would always either hit or miss by the same distance.

This is one of the shortcomings of IL-2...no true 'random' events. Here is an easy example that everyone should be able to understand.

Example #1:

Make a simple test mission with no human players. Place several AA guns on an airfield. Have an AI enemy plane fly a straight course between two waypoints right over the flak. Now, what actually happens is determined by the game itself. The point is that it will happen exactly the same way every time you run the mission.

Go ahead and run the mission again and again. It will always run exactly the same way. If the plane is hit with a flak burst 47 seconds into the mission, it will always be hit with a flak burst 47 seconds into the mission. If originally. the plane explodes from a flak burst 34 seconds into the mission, it will always explode from a flak burst 34 seconds into the mission. If originally, the plane makes it through without being hit by flak at all, it will always make it through without being hit by flak at all.

This is what PaleRidur was trying to explain in his original post.

One would think that if we were to place 'X' amount of flak at an airfield and flew 'X' amount of AI planes over that airfield, we would see varying amounts of damage to those planes every time we ran the mission. In real life, yes.....in IL-2, no. Once the mission is saved and run, all events will occur in the same manner. This is the main point of the original poster.

Now, before emotions get any further out of hand here, there is one key element that needs to be understood here. All of the examples above are concerning missions or AREAS of a mission with NO HUMAN PRESENCE. Once a human controlled plane (or planes) enter an area, all bets are off. In other words, certain events may or may not change, depending on certain factors. Here is an example:

-------------------------------------
Let's look at Example #1 from earlier. Let's say the AI plane always gets hit by a flak burst 47 seconds into the mission (no human-controlled planes in the mission). If we run the mission again, it will always get hit by flak 47 seconds into the mission.

If we introduce a human-controlled plane into the area, let's say flying alongside the AI plane, things may very well change. For instance, the flak gun that originally hit the AI plane may now lock onto the human-controlled plane and the AI plane will not be hit 47 seconds into the mission.

In subsequent runs of this mission, if we have the human-controlled plane flying in random patterns (as humans are want to do...http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif) over the airfield, we will obviously get different results.
-------------------------------------

Now, since most missions obviously have at least one human-controlled player, it may seem that the point of the original poster is moot. Not so.

It is important to realize that a human presence must be in a certain area OR somehow affect certain AI units (possibly further along in the mission) for seemingly 'random' events to occur.

Unfortunately, this is where the tainted aspect of 'no random events' in IL-2 rears it's ugly head. In any particular mission, you may very well have numerous events taking place over many parts of the map. For instance:

---------------------------
a:A tank battle in grid A-1
b:A flight of bombers attacking a tanker in grid B-2
c:A Red AI flight attacking a Blue AI flight in grid C-3
d:A human-controlled plane leading his flight in grid H-5

So, here is the problem. Until (or if) that human-controlled plane and\or his flight somehow come into contact with an AI unit(s) in a, b or c, or somehow affect an AI unit(s) in those areas in some way, events in a, b and c will always be the same:

-The tank battle will always play out and end in the same way.
-The bombers attacking the tanker in will always play out and end in the same way.
-The Red AI vs Blue AI flights will always play out and end in the same way.
---------------------------

This is the problem of non-random events in Il-2. If we want to simulate some kind of dynamic battle going on in a particular mission, we really can't. The game engine, in particular the AI routines, are not programmed to do so.

In closing, we are obviously talking about running the actual mission, not viewing the track...as someone mentioned earlier that the track runs the same every time.

Sorry for the long post.

Aviar

WB_Outlaw
12-02-2005, 09:56 PM
Originally posted by Aviar:
In closing, we are obviously talking about running the actual mission, not viewing the track...as someone mentioned earlier that the track runs the same every time.
Aviar

Playing a .trk file back is exactly like running a mission except that your recorded control inputs are applied for you. In fact, aside from the mission itself, your control inputs are the ONLY thing that's recorded in a .trk file. All of the AI actions during playback are calculated just as if it were an actual mission. That's why .trk files are almost universally incompatible between versions of the game. It's also why there is no .trk "player". To play a .trk back requires the complete game engine.

This concept prevents the use of random events. If there were random events, the .trk files would be different each time. I believe this route was picked to minimize the amount of information that must be written to the disk when recording a mission.


-Outlaw.

PaleRidur
12-03-2005, 02:12 AM
Aviar, thanks for writing that long post. It's basically exactly what i was trying to explain.
For those who still think i was trolling when i started this thread...let me say again..i was not. I've been a beta tester for the flanker series up to lomac..and i've been involved with other products so i know my way around sims and testing. 18 months ago if fell off a roof and got very seriously injured...and on top of it all i'm 35. So my views on life have somewhat changed and i simply do not have the time and the patience for people like Chuck_Older who seem to be mainly concerned with increasing their post count and waving their banner. Which is why i avoid forums...

Anyways, back to the subject.
Lexx wrote: "In your intro physics, when you input the same numbers into your trajectory formula, you get the same answer"
This might be true but you are talking seriously basic physics. Agreed...in intro physics a shell which is thrown away with the exact same force in the exact same direction will travel the same trajectory all the time.
But(i know i'm stating the obvious here) in real life there's the weather, the charge in the shell which will always differ slightly per shell and so on and so on.
Even if you take into acount a seriously stupid (in my opinion) AI who keeps pointing the flak cannon in the exact same direction at exactly the same time and even pulls the trigger at the same moment...there should still be a difference in trajectory and thus a difference in hit or no hit.
Even if you programm this difference in table based form instead of dynamic real time trajectory calculations. Table based very roughly means that the engine would say "based on certain random variables this shell can land anywhere between these coordinates".
And even that would be very basic...but as far as i'm concerned way more acceptable then the totally static things which seem to be happening now.

LEXX_Luthor
12-03-2005, 03:09 AM
Pale::
But (i know i'm stating the obvious here) in real life there's the weather, the charge in the shell which will always differ slightly per shell and so on and so on.
I know I'm stating the obvious here. We are both talking about changes in input variables for a computer program (ie..."not real life" as we like to say).

Hopefully Oleg will allow variable inputs later, in BoB And Beyond. But not today.

Chuck_Older
12-03-2005, 09:04 AM
Originally posted by PaleRidur: i simply do not have the time and the patience for people like Chuck_Older who seem to be mainly concerned with increasing their post count and waving their banner.

Utter BS. You act as if you know me and you don't. You're the only one waving a flag, and the flag is "Chuck picked on me".

Hey look, you don't see me making you post things, do you? No. What you initially posted is backhanded and I thought it was funny.

So now I keep reading BS concerning you tossing mud at me. <span class="ev_code_YELLOW">You</span> tell <span class="ev_code_YELLOW">me</span> how you'd like me to respond to that, because what you're showing me is that you talk BS about me without bothering to know what you or I are about. You tell me how I should respond to the insult you just gave me. I explained this to you before and you didn't bother to read it apparently. You're out of line and way off base. So you tell me right now how I should reply to your cr*p http://forums.ubi.com/images/smilies/16x16_smiley-indifferent.gif It's your call

Flying_Nutcase
12-03-2005, 11:35 AM
Originally posted by PaleRidur:
...the totally static things which seem to be happening now.

Hang on to your hat. Seriously, this will soon be just a bad memory.

If I can have a little more luck with my free time, nice things (think Sturmi randomisation) should be here for Christmas.

Keep an eye on the the IL2 Fog Of War thread in Mission Builder's forum.

~S~

ManicGibber
12-03-2005, 12:37 PM
Rather amusing how you have to lick some peoples backsides the exact same way so that they don't get offended the exact same way before responding and end up cussing you out instead of helping out, which proves that there is a similarity with the original post, the lack of randomisation with flak burst's and real life on the Forums. If indeed you can call the Forums real life, another flame war be sure. Anyway great post. I was under the opinion that the aaa was controlled by an algorythm which would make it simple enough to alter start of firing time and rof of the aaa for x times a given mission was run?

Chuck_Older
12-03-2005, 12:50 PM
Gibber, the PM button is right there. If you want to tell "some people" off, but don't want to name them, you can PM them and nobody but they and you would know about it. Give it a shot, you might learn something

ManicGibber
12-03-2005, 01:05 PM
Originally posted by Chuck_Older:
Gibber, the PM button is right there. If you want to tell "some people" off, but don't want to name them, you can PM them and nobody but they and you would know about it. Give it a shot, you might learn something

It was amusing don't you think http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif?

Chuck_Older
12-03-2005, 03:22 PM
I've no idea what you're talking about

KG26_Alpha
12-08-2005, 01:27 PM
This thread is a mess I feel sorry for the thread starter.

No wonder he dont post much at UBI boards.

Hows that utillity doing Nutcase? need any testing help ?

Just to add my experience with "flak" problems, its seems that flak does have a routine but you can try diffrent types of artillery to see which give a more random effect try mixing up different calibre and even mixing different countries you may find some good results.

Jetbuff
12-08-2005, 01:53 PM
If anyone ever wondered how things can go South quickly this is a case study.

Threadstarter, starts off with a comment that could be misconstrued.

An otherwise well-meaning respondent makes fun of the comment.

Threadstarter is insulted and goes on the defensive.

Respondent is now really peeved and the two are avowed enemies on the forum unless one of them takes the initiative to clarify the original misunderstanding.

Mountain, meet mole-hill.