PDA

View Full Version : Once again, please fix the collision model...



GR142-Pipper
11-25-2006, 06:00 PM
I realize that this has been discussed before but it's worth bringing up again. The collision model is not only innacurate, it's pathetic.

I was flying in a Yak-3 when a 190 came by and collided with me as he made a BnZ pass. His wing came off and nothing...absolutely nothing...was wrong with my plane. C'mon. In real life, both planes would have gone down.

I've been in similar circumstances where my plane lost a wing and the other guy went on his merry way.

Maddox, please take the time to fix this. It's nothing less than gross and detracts from your fine game.

GR142-Pipper

han freak solo
11-28-2006, 02:35 PM
I found this link with the accident report.

http://www.ntsb.gov/ntsb/brief2.asp?ev_id=20001212X2228...o=FTW01FA025B&akey=2 (http://www.ntsb.gov/ntsb/brief2.asp?ev_id=20001212X22287&ntsbno=FTW01FA025B&akey=2)<div class="ev_tpc_signature">

http://www.lssdigital.com/lwpilot/HFSsig47.jpg

Xiolablu3
11-28-2006, 02:44 PM
Do you really expect Oleg to go back and start changing base code in the Il2 engine for all planes at this late stage?

I dont see a big problem in the collisionmodel, I mean how often do you actually collide with someone and have these strange phenomena? Once a month? , Less?

Lets say the average pilot collides with someone once a week.

Out of those times say 1 in 10 you have a 'strange occurance', that means it comes up about once every 3 months.

If you are colliding with planes more often then maybe change your flying style a bit.

Its just not worth fixing such a tiny minor problem.<div class="ev_tpc_signature">

--------------------------------------------------------------------
"I despise what you say; I will defend to the death your right to say it."
-Voltaire

han freak solo
11-28-2006, 02:46 PM
Originally posted by han freak solo:
I found this link with the accident report.

http://www.ntsb.gov/ntsb/brief2.asp?ev_id=20001212X2228...o=FTW01FA025B&akey=2 (http://www.ntsb.gov/ntsb/brief2.asp?ev_id=20001212X22287&ntsbno=FTW01FA025B&akey=2)

Apparently, what I thought I saw and what was, are a little different in the details.<div class="ev_tpc_signature">

http://www.lssdigital.com/lwpilot/HFSsig47.jpg

BillyTheKid_22
11-28-2006, 03:13 PM
Originally posted by GR142_Astro:
Ok, this one is for you interweb experten. What has lag or packet loss got to do with the outcome of a collision, when:

a) ac1 strikes ac2, with ac1 losing a wing and ac2 losing a stab.

b) ac1 strikes ac2, with ac1 losing a wing and ac2 having its wing ruined or "holed"

c) ac1 strikes ac2 with ac1 losing a wing and ac2 flies on unharmed

The computers and game are obviously acknowledging that both ac1 and ac2 are striking each other. I can fully agree that it would be difficult to accurately represent the varying degrees of damage that would result from collisions, but something is wrong when it is more the exception than the rule that BOTH ac go down in flames.

Even incidental contact between fast-moving ac is most often fatal to both machines. I suppose in simplistic terms, it should be possible to increase the probability values that lean more towards the destruction of both ac.

http://www.mustangsmustangs.net/p-51/survivors/pages/picfiles/44-74536/44-74536_31_jd3.jpg http://forums.ubi.com/images/smilies/25.gif<div class="ev_tpc_signature">

http://members.cox.net/bkid/pacificfighters/PacificFightersSignature.JPG

.................................................. ..............

"All I got was a bellyful of English Channel."

Capt.LoneRanger
11-28-2006, 03:17 PM
I didn't say you are the one causing them. http://forums.ubi.com/images/smilies/10.gif<div class="ev_tpc_signature">

greets
Capt.LoneRanger

http://img.photobucket.com/albums/v411/Arcadiac/Lone1copy.png

BillyTheKid_22
11-28-2006, 04:31 PM
Originally posted by Capt.LoneRanger:
I didn't say you are the one causing them. http://forums.ubi.com/images/smilies/10.gif http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif<div class="ev_tpc_signature">

http://members.cox.net/bkid/pacificfighters/PacificFightersSignature.JPG

.................................................. ..............

"All I got was a bellyful of English Channel."

GR142_Astro
11-29-2006, 06:21 AM
Originally posted by Capt.LoneRanger:
I didn't say you are the one causing them. http://forums.ubi.com/images/smilies/10.gif

http://forums.ubi.com/images/smilies/10.gif

I know, and I cannot control the actions of an over-zealous Butchered Bird Person who finally has a clear shot on me while I'm dueling with somebody.<div class="ev_tpc_signature">

http://members.cox.net/kinetic/SigImages/LockheedLightningMed.jpg

____________________________

"If Adolf Hitler flew in today, they'd send a limosine anyway." ~TheClash~

Capt.LoneRanger
11-29-2006, 07:11 AM
Clear shot in a dogfight?

Well, I understand your pain, you know, but really, this happened to me like 20 times max, since the original IL2 and 18 times of which while I was sitting in a bomber and 99% with both planes going down.

Since this is a problem for you and how you address it, it seems like this is happening like everytime you get online!?<div class="ev_tpc_signature">

greets
Capt.LoneRanger

http://img.photobucket.com/albums/v411/Arcadiac/Lone1copy.png

GR142-Pipper
11-29-2006, 12:51 PM
Originally posted by WWMaxGunz:
Here is reality vs unreal. IRL in a very close call or nudge if your plane does not get broken
you say 'whew!' and some other words. Unreal, you get upset and complain We're not talking about a "close call". We're talking about collisions. In real life, in most cases both planes are lost.

It's a game flaw. If the game can detect one aircraft was involved in the collision (and it can) then it an certainly establish that both planes be lost.

GR142-Pipper

BillyTheKid_22
11-29-2006, 01:35 PM
Originally posted by GR142_Astro:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Capt.LoneRanger:
I didn't say you are the one causing them. http://forums.ubi.com/images/smilies/10.gif

http://forums.ubi.com/images/smilies/10.gif

I know, and I cannot control the actions of an over-zealous Butchered Bird Person who finally has a clear shot on me while I'm dueling with somebody. </div></BLOCKQUOTE>

http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif http://forums.ubi.com/images/smilies/clap.gif<div class="ev_tpc_signature">

http://members.cox.net/bkid/pacificfighters/PacificFightersSignature.JPG

.................................................. ..............

"All I got was a bellyful of English Channel."

Capt.LoneRanger
11-29-2006, 01:53 PM
In real life, in most cases both planes are lost.

That is not really true. I can show you several pictures of catastropic collisions, where one of the planes survived very well.

The most famous is probably the B17 with an almost cut through fuselage. It was hit by a Bf109, who's pilot didn't make it out after the mid-air, but the B17 and the crew made it back home safely. It is assumed the negatve accelleration of the crash probably killed the Bf109-pilot.

I also got a picture of a documented crash between a P40 and a Japanese plane (don't remember what type) - the P40 lost it's stabilizer almost completely, the Jap had his wing-tank ruptured and exploded shortly after the accident.
I also read a story about a P51B colliding with his leader. The leader crashed, the wingie was able to fly, but bailed too early and behind enemy lines.

Still, I don't know if solving this problem really is worth the effort.<div class="ev_tpc_signature">

greets
Capt.LoneRanger

http://img.photobucket.com/albums/v411/Arcadiac/Lone1copy.png

GR142-Pipper
11-29-2006, 04:22 PM
Originally posted by Capt.LoneRanger:
[QUOTE]In real life, in most cases both planes are lost.


That is not really true. I can show you several pictures of catastropic collisions, where one of the planes survived very well. If you read my post I said "in most cases" both planes are lost. This is particularly true of fighter aircraft that are engaged.

GR142-Pipper

Popey109
11-29-2006, 08:14 PM
I think its kinda cool...A world where anything can happen http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif Once ran out of amo, so I cut his tail of with my prop http://forums.ubi.com/images/smilies/winky.gif made it home with a bent nose. Have lost many a vertical stab in head on passes. Chok it up to the weeeerd world of war...So be carful out there http://forums.ubi.com/images/smilies/1072.gif

Badsight-
11-29-2006, 10:28 PM
Turkish & Greecian F-16's recently collided whilst doing an aggressive mock-DF

Maj.Kaos
11-29-2006, 10:29 PM
What's the problem with the damage model?

http://www.pacificstorm.net/Articles/US_sea_aircraft/image/14.jpg

I doubt the pilot of this plane wrote to Grumman to complain about the plane's "unrealistic" flight model.

Just be happy. It's a great game!!

I've been playing since MS 1, on the apple 2. Flew the Sopwith Camel up to 30,000 feet till out of gas, then glided in for a landing. That tiny square flight world looked pretty small from 30k. Totally unrealistic, but I nonetheless enjoyed it.

Fork-N-spoon
11-29-2006, 10:56 PM
I've always wanted to say this, "Got a track?"

God, the things I would like to do to people that say that!<div class="ev_tpc_signature">

Blondes are for nancy boys. Stop the cruelty adopt a brunette today

http://img.photobucket.com/albums/v358/bolillo_quemado/brunette.jpg

GR142Pipper
11-30-2006, 12:29 AM
Originally posted by Fork-N-spoon:
I've always wanted to say this, "Got a track?"

God, the things I would like to do to people that say that! Actually, I don't. However, collisions with the loss of just one aircraft (often with no damage whatsoever to the other aircraft) occur frequently enough given the number of players that I'm sure it's been widely experienced. This is one of those issues that's not a Red or Blue matter. It's a game flaw that on the surface of it should be easy enough to correct by altering the code to result in the loss of both aircraft...which is what occurs in the vast majority of fighter aircraft collisions in real life.

GR142-Pipper

Vipez-
11-30-2006, 01:37 AM
Originally posted by GR142-Pipper:
I realize that this has been discussed before but it's worth bringing up again. The collision model is not only innacurate, it's pathetic.

I was flying in a Yak-3 when a 190 came by and collided with me as he made a BnZ pass. His wing came off and nothing...absolutely nothing...was wrong with my plane. C'mon. In real life, both planes would have gone down.

I've been in similar circumstances where my plane lost a wing and the other guy went on his merry way.

Maddox, please take the time to fix this. It's nothing less than gross and detracts from your fine game.

GR142-Pipper

Funny thing online is the slower plane usually gets away unscratched. So in this case you as a slower plane (B&Z-victim) and nothing happens. And for the attacker with higher airspeed he looses his wing.. happened to me hundeds of times.<div class="ev_tpc_signature">

__________________________


http://personal.inet.fi/cool/tmu/kuvat/sig.jpg

A.K.A LeOs.K_Vipez

Badsight-
11-30-2006, 01:41 AM
Originally posted by Vipez-:
Funny thing online is the slower plane usually gets away unscratched.
Complete

Lies

you can be slower , faster , higher , lower , turning , travelling straight , firing guns , guns silent , power on , power off . . . . & any other damm variable YOU care to name , & STILL BE THE ONE WHO FACE PLANTS

it is beyond understanding how the collision modelling works in this game because it has no reason behind it

R988z
11-30-2006, 06:39 AM
The Beer Run

On June 13, 1944, (D-Day plus seven) number 412 (Falcon) Squadron, along with the others comprising 126 Wing gathered for a briefing by W/C Keith Hodson at our Tangmere base.

We would get details of our now regular Beach Patrol activities, only this one had a slight variation.

The Wingco singled me out to arrange delivery of a sizable shipment of beer to our new airstrip being completed at Beny-sur- Mer.

The instructions went something like this ??? "Get a couple other pilots and arrange with the Officers Mess to steam out the jet tanks and load them up with beer. When we get over the beachhead drop out of formation and land on the strip. We're told the Nazis are fouling the drinking water so it will be appreciated."

"There's no trouble finding the strip, the Battleship Rodney is firing salvoes on Caen and it's immediately below. We'll be flying over at 13,000 so the beer will be cold enough when you arrive."

I remember getting Murray Haver from Hamilton and a third pilot (whose name escapes me) to carry out the caper.

In reflection it now seems like an appropriate Air Force gesture for which the erks (infantrymen) would be most appreciative.

By the time I got down to 5,000 the welcoming from the Rodney was hardly inviting but sure enough there was the strip.

Wheels down and in we go, three Spits with 90 gallon jet tanks fully loaded with cool beer.

As I rolled to the end of the mesh runway it was hard to figure . . . there was absolutely no one in sight. What do we do now, I wondered, we can't just sit here and wait for someone to show up. What's with the communications?

Finally I saw someone peering out at us from behind a tree and I waved frantically to get him out to the aircraft. Sure enough out bounds this army type and he climbs onto the wing with the welcome . . . "What the hell are you doing here?"

Whereupon he got a short, but nevertheless terse, version of the story.

"Look," he said "can you see that church steeple at the far end of the strip? Well it's loaded with German snipers and we've been all day trying to clear them out so you better drop your tanks and bugger off before it's too late."

In moments we were out of there but such was the welcoming for the first Spitfire at our B4 airstrip in Normandy.

The unbelievable sequel to this story took place in the early 1950s at Ford Motor Company in Windsor where I was employed at the time.

A chap arrived to discuss some business and enquired if I had been in the Air Force. "Yes, indeed," I responded.

"Did you by chance land at Beny-sur-Mer in Normandy with two other Spitfires with jet tanks loaded with beer?" he asked.

"Yes for sure I did," I answered, "But how on earth would you possibly be aware of that?"

"Well I'll tell you," he said, "I was the guy who climbed on your wing and told you to bugger off."

We finished the afternoon reminiscing.

http://www.vac-acc.gc.ca/remembers/sub.cfm?source=histo...ondwar/berryman/brun (http://www.vac-acc.gc.ca/remembers/sub.cfm?source=history/secondwar/berryman/brun)

http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

NonWonderDog
11-30-2006, 09:43 AM
This doesn't happen offline, ever, so it's pretty easy to say it's due to packet loss and/or latency. The server knows that the one guy hit you but doesn't know you hit the other guy, or something equally kooky.

It's probably fixable, but the costs just aren't worth it. If we had full redundancy on every thing that happened online we'd no longer be able to play in real time. Code that executes only when one plane hits another to make sure both are damaged, on the other hand, would be necessarily inaccurate. It would probably end up such that any touch at all would result in an explosion. Anything else would cause a good deal of lag whenever there's a collision, since you'd effectively be asking the server to recreate the collision from two conflicting accounts.

Longpo
11-30-2006, 10:51 AM
http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

I have a book with some interesting pictures of the Spitfire, one of which is a Spit with two beer barrels under each wing, wonder if it???s related.

I will try scanning it, just hope the old scanner wont kill my pc like it did a few years ago.

LStarosta
11-30-2006, 10:55 AM
http://www.cybermodelers.com/articles/review/spitfire9/Beer%20Barrel.jpg <div class="ev_tpc_signature">

_____________________________

http://img214.imageshack.us/img214/5310/tagjimmyssw1.jpg

I_KG100_Prien
11-30-2006, 11:02 AM
My question is, have you had a lot of the same type of results happen while flying offline?

Every single time I've had a collision while flying offline against AI I have NEVER NEVER NEVER seen what you describe happening. Not once.. and I've done a boucoup load of offline flying since I havn't flown online since 4.05 came out. (Waiting on the '46 DVD).

However, I have had it happen a few times when I was actively flying online, which points to the issue of lag/packet loss. I also recall watching a track of one of my online sorties, and at one point I got shot down. The cool thing was, when watching the track.. the tracers of the plane that shot me down never even "Came close" to my plane.. they where whizzing about 200 yards away. However, that was from my computers point of view.. Obviously, on the other end and on the server end things were "good". The hits were there although I could not actually physically see them until my wing fell off.

Longpo
11-30-2006, 11:09 AM
Scanner plan went tits up, but the camera done a good job (probably better than the scanner would).

http://img.photobucket.com/albums/v726/calmo/Spit001.jpg

"Comforts for the troops: a Spitfire Mk IXE en route to Normandy with two 18-gallon beer barrels attached to its underwing shackles."

http://forums.ubi.com/images/smilies/16x16_smiley-tongue.gif

GR142-Pipper
11-30-2006, 11:29 AM
Originally posted by I_KG100_Prien:
My question is, have you had a lot of the same type of results happen while flying offline? I would agree that this points to being an on-line issue. Regarding difficulty to fix, who knows. The fact is that the game can recognize the collision, takes the wing (stab, rudder, etc.) off of one plane and nothing happens to the other and the wingless aircraft is recognized upon impact. That tells me it isn't packet loss because the collision is recognized and the participants identified. If it was pure packet loss then the collision itself may not be recognized and the program wouldn't know who to assign the "kill" to.

This is similar to the way to program can assign a friendly kill to the player that is ran INTO...very strange logic there. Although I may be wrong, this collision matter seems to be an area in which a nominal amount of coding could eliminate some very game strange behavior...for all players.

GR142-Pipper

T_O_A_D
11-30-2006, 11:42 AM
Well that makes for a popular pilot for a day. http://forums.ubi.com/images/smilies/16x16_smiley-happy.gif<div class="ev_tpc_signature">

<center>
Flying as BlitzPig_TOAD
My TrackIR fix, Read the whole thread (http://ubbxforums.ubi.com/6/ubb.x?q=Y&a=tpc&s=400102&f=49310655&m=15310285&p=1)
Charvel's Tutorial on setting up yout Track Ir (http://www.airwarfare.com/guides/tir_setup.htm)

<A HREF="http://blitzpigs.com/forum/index.php?sid=3b61cb4521d729767adc892dfd2ceddf" TARGET=_blank>http://home.mchsi.com/~t_o_a_d/T_O_A_D.jpg
</A>
<b style="color:black;background-color:#a0ffff">UBI[/b] Forums/IL2/FB/PF Moderator
I.O.C.L (http://www.gozr.net/iocl/)> Pictures of War (http://forums.ubi.com/groupee/forums/a/tpc/f/23110283/m/2291023914)>My AT-6 Flight (http://forums.ubi.com/eve/forums/a/tpc/f/23110283/m/6641084723?r=6641084723#6641084723)>

NonWonderDog
11-30-2006, 11:51 AM
Collisions are at least partially client-side, is all that says. One guy's computer thinks they've collided, the other guy's computer thinks they've just missed. Damage is applied to the client's plane locally, and then reported to the server. If there was no lag, no packet loss, infinite bandwidth, and the client computers were infinitely powerful, they would always agree that they collided in the same way. That's obviously not realistic. You can't double-check everything on the server because bandwidth is limited.

Making collisions server-side would probably require making a significant portion of the damage model server-side, or else they probably would have done it. Adding a collision check on the server that damages both planes if one client reports a collision is possible, but there would be no internet-friendly way to do it but "you touch, you go boom." You'd explode slightly after the collision, too, so you'd hit, be undamaged, and then explode as soon as the server updates.


The problem isn't just the amount of coding needed to fix this kind of thing, it's the server load and bandwidth it would entail. It has to work on 56k modems, as close to real time as possible. Only the barest minimum of data can be sent between the client and server. That means clients will occasionally disagree about the little things, and collisions were apparently deemed low priority.

And the more data you need to send between client and server, the more prediction each client has to do. If it takes a full second to transfer all the data about the battle, each client computer is on its own for that second. Every second, each client would update with the data on the server (this is why it gets jerky on slow connections). I think you can see, now, how the clients could disagree on collisions. There's no way to just flip a switch and fix these kind of things; any extra data sent for corrections makes the problem worse.

Fork-N-spoon
11-30-2006, 12:24 PM
Why do you think that Lockheed made 310 gallon drop tanks for the P-38?

I've read about units in the PTO that would fill their drop tanks with warm beer, climb to 25,000 feet, and by the time they landed, the beer was ice cold.<div class="ev_tpc_signature">

Blondes are for nancy boys. Stop the cruelty adopt a brunette today

http://img.photobucket.com/albums/v358/bolillo_quemado/brunette.jpg

stathem
11-30-2006, 12:42 PM
"So when will get this load-out?"

... In SoW - (http://www.simhq.com/_air7/air_250b.html)

Probably.<div class="ev_tpc_signature">

http://homepage.ntlworld.com/griffnav/Gallery/MossiePRsig.jpg

Heavy_Weather
11-30-2006, 01:06 PM
and lets not forget how cheesy the smoke looks when flying behind a smoking aircraft. 2D

voyager_663rd
11-30-2006, 02:51 PM
http://www.directart.co.uk/mall/images/lr7.jpg

http://www.directart.co.uk/mall/more.php?ProdID=4073<div class="ev_tpc_signature">

my goal for my 1/8th scale RC Corsair 1D:

http://img.photobucket.com/albums/v37/voyager_663rd/corsair.jpg

Friendly_flyer
11-30-2006, 03:12 PM
Originally posted by Longpo:

http://img.photobucket.com/albums/v726/calmo/Spit001.jpg

It's quite possible the Spitfire is from 331 or 332 (Norwegian) Squadron. A very similar picture (possibly the same plane) is found in the book "Tally ho", about the Norwegian fighter squadrons flying with the RAF. The caption reads (translated):

The Spitfire could take tree bombs, two under the wings like in this picture, only this is not bombs but beer kegs! On the trip back from England to the forward bases on the Continent the Norwegian pilots where more than happy to hang kegs from the bomb-racks. The beer was thoroughly chilled on during the flight!

The picture is not necessary related to the Norwegian squadrons though. No squadron code is visible, and the text indicate it was a not a too uncommon event.<div class="ev_tpc_signature">

Fly friendly!

http://i8.photobucket.com/albums/a37/Friendly_flyer/WhirlwindforBoB-II.jpg

Visit No 79 Squadron vRAF (http://www.gazzamataz.com/79vRAF/)

Petter B??ckman
Norway

Low_Flyer_MkVb
11-30-2006, 03:54 PM
Friendly_flyer - the Spitfire 70 edition of Flypast magazine gives the serial of the Spit in the photo as MJ329. Don't know if that might inspire some detective work.<div class="ev_tpc_signature">

http://i109.photobucket.com/albums/n47/LFMkVb/1822.jpg

OD_79
11-30-2006, 03:58 PM
If I remeber right I saw a pictue of a Spit Mk I being used for the same role around the D-Day time. Must have been one of the older aircraft around then!

OD.<div class="ev_tpc_signature">

http://img.photobucket.com/albums/v240/tom1502/HurricaneSig2.jpg (http://www.79vrafhangar.co.uk)

GR142-Pipper
11-30-2006, 04:12 PM
Originally posted by NonWonderDog:
Collisions are at least partially client-side, is all that says. One guy's computer thinks they've collided, the other guy's computer thinks they've just missed. Damage is applied to the client's plane locally, and then reported to the server. If there was no lag, no packet loss, infinite bandwidth, and the client computers were infinitely powerful, they would always agree that they collided in the same way. That's obviously not realistic. You can't double-check everything on the server because bandwidth is limited.

Making collisions server-side would probably require making a significant portion of the damage model server-side, or else they probably would have done it. Adding a collision check on the server that damages both planes if one client reports a collision is possible, but there would be no internet-friendly way to do it but "you touch, you go boom." You'd explode slightly after the collision, too, so you'd hit, be undamaged, and then explode as soon as the server updates.


The problem isn't just the amount of coding needed to fix this kind of thing, it's the server load and bandwidth it would entail. It has to work on 56k modems, as close to real time as possible. Only the barest minimum of data can be sent between the client and server. That means clients will occasionally disagree about the little things, and collisions were apparently deemed low priority.

And the more data you need to send between client and server, the more prediction each client has to do. If it takes a full second to transfer all the data about the battle, each client computer is on its own for that second. Every second, each client would update with the data on the server (this is why it gets jerky on slow connections). I think you can see, now, how the clients could disagree on collisions. There's no way to just flip a switch and fix these kind of things; any extra data sent for corrections makes the problem worse. To be honest, I'm really not convinced that it's all that difficult to rectify to make it more realistic than what we have now. My initial purpose was to focus a bit of attention on this matter so that going forward we'll have a more real life representation of how collisions should be handled.

If an improvement gets implemented, great. If not, it won't surprise me.

GR142-Pipper