PDA

View Full Version : Very strange finding: Structural damage related to packet loss !?



FatBoyHK
01-10-2005, 07:44 AM
Someone here may have read the thread I started yesterday, about the structural damage modeling here (http://forums.ubi.com/eve/forums?a=tpc&s=400102&f=23110283&m=7161043162)...

Tonight (when WC reboots) I perfomred a series of high G pulling tests on a Mustang in offline QMB... The result is very strange.... I couldn't break my wing unless I tried *very* hard... Basically I can't possibly do any damage when I fly below 400 MPH IAS.... and the test results are quite consistant over all altitutes

It contrasts with my online experience.... There I have broken my wing in 350 MPH and with just 50% stick deflection.... and it is kinda random, sometimes I get through it, sometimes I can't.

and IIRC, if I keep breaking wing in one particular night, I also experience bad gun effectiveness at that night, for example I can't damage someone that I caught standing still on top of his zoom, within 50m...

I started to fear that the breaking wing issue is also related to packet loss, just like the 0.50cal problem... may be the server "loss track" of my plane when I am pulling some G, and some miniseconds later, it receive my newest position, and to make up this loss of data the server make me "wrapped" through space (which is the UFO effect we all have seen before). Now think about it, a 10G turn in 20 msec is similar a 20G turn in 10 msec.... and here goes a breaking wings!!!

and I fear that, just like the 0.50cal problem, there will be no solution and we have to get over with it.... well, I can get over with a missed shot that should have blasted someone into piece, I can try again in the next pass... but I really can't get over with a randomly-broken wing, once it strike on me I am dead immediately...

FatBoyHK
01-10-2005, 07:44 AM
Someone here may have read the thread I started yesterday, about the structural damage modeling here (http://forums.ubi.com/eve/forums?a=tpc&s=400102&f=23110283&m=7161043162)...

Tonight (when WC reboots) I perfomred a series of high G pulling tests on a Mustang in offline QMB... The result is very strange.... I couldn't break my wing unless I tried *very* hard... Basically I can't possibly do any damage when I fly below 400 MPH IAS.... and the test results are quite consistant over all altitutes

It contrasts with my online experience.... There I have broken my wing in 350 MPH and with just 50% stick deflection.... and it is kinda random, sometimes I get through it, sometimes I can't.

and IIRC, if I keep breaking wing in one particular night, I also experience bad gun effectiveness at that night, for example I can't damage someone that I caught standing still on top of his zoom, within 50m...

I started to fear that the breaking wing issue is also related to packet loss, just like the 0.50cal problem... may be the server "loss track" of my plane when I am pulling some G, and some miniseconds later, it receive my newest position, and to make up this loss of data the server make me "wrapped" through space (which is the UFO effect we all have seen before). Now think about it, a 10G turn in 20 msec is similar a 20G turn in 10 msec.... and here goes a breaking wings!!!

and I fear that, just like the 0.50cal problem, there will be no solution and we have to get over with it.... well, I can get over with a missed shot that should have blasted someone into piece, I can try again in the next pass... but I really can't get over with a randomly-broken wing, once it strike on me I am dead immediately...

BBB_Hyperion
01-10-2005, 07:49 AM
I fear that your research is not complete therefore wrong.

FatBoyHK
01-10-2005, 08:14 AM
I haven't come to any conclusion yet, but I am honest with my finding.

To further my testing I did some more experiment on WC tonight... It has just been rebooted, so it should be running at its best performance, and, more favoribly, not much people was there. I didn't see a single "Cheating" message for the whole period of my experiment.

During the experiment, I applied max stick deflection at 410 MPH IAS, at 15000 ft, 12000 ft, 8000ft and 4000ft, no damage occurred. Then I tried once more at 4000ft and 420MPH, and I broke my wing this time.

I will try another test tmr, when I see more people online, and proably with a worse network performance.

ZG77_Lignite
01-10-2005, 09:32 AM
FatBoy, I think its reasonable to assume that structural failure is calculated locally (meaning on your computer, just like offline), and the results (breakage, or not) is what is sent to the host to be distributed. Packetloss should not be able effect stuff that is calculated at your end (for example, how many bullets you have fired, even if the host only sees you fire 1/2 the bullets due to packetlosses, you still run out of bullets the same as you would offline).

Its great if you keep testing, but I believe its already been done, with the diagnostic tools of DeviceLink, and the conclusion was that all aircraft break at 15g limit (as stated by Oleg, simplified but more realistic than nothing). Its just that some aircraft have such a huge advantage in elevator authority that fantastic (and probably unrealistic to some degree) manuevers can be pulled.

One thing you may be experiencing, is 'stick-spiking' (from well-worn joystick).

FatBoyHK
01-10-2005, 10:13 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by ZG77_Lignite:
FatBoy, I think its reasonable to assume that structural failure is calculated locally (meaning on your computer, just like offline), and the results (breakage, or not) is what is sent to the host to be distributed. Packetloss should not be able effect stuff that is calculated at your end (for example, how many bullets you have fired, even if the host only sees you fire 1/2 the bullets due to packetlosses, you still run out of bullets the same as you would offline).
<HR></BLOCKQUOTE>

I actually think you are right.... but I just ran out of idea to explain why I would break my wing much much easiler online...

And, corect me if I am wrong, DM is done on server... just like when you shoot someone, it is the server to decide if it is a hit or not. If it is done on client, well, the issue we have with the online 0.5cal will not occur, (but we may open the door for hackers at the same time) and then, G-related damage is part of the DM, so it may be done on the server too....

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by ZG77_Lignite:
Its great if you keep testing, but I believe its already been done, with the diagnostic tools of DeviceLink, and the conclusion was that all aircraft break at 15g limit (as stated by Oleg, simplified but more realistic than nothing). Its just that some aircraft have such a huge advantage in elevator authority that fantastic (and probably unrealistic to some degree) manuevers can be pulled.
<HR></BLOCKQUOTE>

You points has nothing to do with my claims. Yes all plane break at 15G. and I am fine with it, absolutely no whine http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif.

However, what I am trying to prove (or dis-prove) is that, due to packet loss, the server may preceive a high G pull for a short instant, while the average G over a longer period remind constant..... and hence, while all of us are governed by a *raw* G-limt of 15G, you would have an *effective* G-limit lower than 15G if your network is not 100% perfect.

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by ZG77_Lignite:
One thing you may be experiencing, is 'stick-spiking' (from well-worn joystick). <HR></BLOCKQUOTE>

No, I have a modded Cougar (using Uber2 mod) which is working pecfectly.

BBB_Hyperion
01-10-2005, 11:22 AM
You can easily record tracks and check for g forces in devicelink while playing back.

Loki-PF
01-10-2005, 11:39 AM
you sure about that Hyperion? I've had several longtime members here, and also one's quite experienced with device link state that there is no way to get G's outta tracks with devicelink.....

BBB_Hyperion
01-10-2005, 11:54 AM
Hmm as you told it i checked playing back track and indeed only 0 returns as value . Wonder how i did get this g graphs from my older recordings .

Bull_dog_
01-10-2005, 01:55 PM
offline, I can regularly dive a mustang to 900km/hr TAS and 540mph IAS and not break up if I watch it...

Offline I've broken up as low as 650km/hr TAS...without blackout. The above numbers, I did gray and blackout.

Looks like a bug, smells like a bug...I think it is a bug http://forums.ubi.com/images/smilies/16x16_smiley-indifferent.gif

Mad_Moses
01-10-2005, 04:33 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by FatBoyHK:
Someone here may have read the thread I started yesterday, about the structural damage modeling http://forums.ubi.com/eve/forums?a=tpc&s=400102&f=23110283&m=7161043162...

Tonight (when WC reboots) I perfomred a series of high G pulling tests on a Mustang in offline QMB... The result is very strange.... I couldn't break my wing unless I tried *very* hard... Basically I can't possibly do any damage when I fly below 400 MPH IAS.... and the test results are quite consistant over all altitutes

It contrasts with my online experience.... There I have broken my wing in 350 MPH and with just 50% stick deflection.... and it is kinda random, sometimes I get through it, sometimes I can't.

and IIRC, if I keep breaking wing in one particular night, I also experience bad gun effectiveness at that night, for example I can't damage someone that I caught standing still on top of his zoom, within 50m...

I started to fear that the breaking wing issue is also related to packet loss, just like the 0.50cal problem... may be the server "loss track" of my plane when I am pulling some G, and some miniseconds later, it receive my newest position, and to make up this loss of data the server make me "wrapped" through space (which is the UFO effect we all have seen before). Now think about it, a 10G turn in 20 msec is similar a 20G turn in 10 msec.... and here goes a breaking wings!!!

and I fear that, just like the 0.50cal problem, there will be no solution and we have to get over with it.... well, I can get over with a missed shot that should have blasted someone into piece, I can try again in the next pass... but I really can't get over with a randomly-broken wing, once it strike on me I am dead immediately... <HR></BLOCKQUOTE>

I noticed the same thing a while back... I guess I'm not crazy.

P-51D is just a hunk of **** now anyway. Speed is everything in dog fighting. A plane known for its great speed seems to be about the slowest of it's contemporaries since the latest patching. When you use an altitude advantage as your only way to generate speed it often results in the plane falling apart while other planes produced in dark caves with whatever materials that could be scrounged up hold together under much more stress.

I'm very disappointed in the Mustang's FMB now. http://forums.ubi.com/groupee_common/emoticons/icon_frown.gif

Regards,
Mad

FatBoyHK
01-10-2005, 07:19 PM
Mustang is not slow by any mean. Some times ago I think in your way, but after I learned about rad setting, and memorized those charts provided by IL2-Compare, I can outpace just everyone http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

but the structural damge, I really can't get over with it....

FatBoyHK
01-10-2005, 07:23 PM
I will start recording my online sorties from now on. No need to obtain the actual G value, I think just IAS (like compressibility, G-damage is related to IAS but not TAS, correct me if I am wrong) reading is good enough.... If you can break thing at 350 MPH, there must be some serious flaw....

Mad_Moses
01-10-2005, 07:28 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by FatBoyHK:
Mustang is not slow by any mean. Some times ago I think in your way, but after I learned about rad setting, and memorized those charts provided by IL2-Compare, I can outpace just everyone http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

but the structural damge, I really can't get over with it.... <HR></BLOCKQUOTE>

I'll jump into a server now and experiment with rad and prop pitch settings and see if I can squeeze some more speed out of it.

I flew it for a couple hours the other night (first time since the patch) and I was in shock.

Mad

FatBoyHK
01-10-2005, 08:55 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Mad_Moses:
I'll jump into a server now and experiment with rad and prop pitch settings and see if I can squeeze some more speed out of it.
<HR></BLOCKQUOTE>

I just leave the RPM at max setting, and if I need nothing but raw speed I would close the rad. You can overheat a Mustang for 5 minutes max.

Back to the topic. If you plan to do some experiment online pls do some G testing for me too. Pull your stick to the max in various IAS and alt, record the speed that you would break your wing. and pay attention to any abnormal result (i.e. wing break at low speed.)

FatBoyHK
01-12-2005, 09:25 AM
bump!

FatBoyHK
01-14-2005, 04:27 AM
today my net connection seem to be not ideal, so I went on to do some more experiment instead of DFing....

Vertical pullup with full stick defection at 2500M, on an icy map, online I can break my wing at as slow as 590kps IAS, or 659kps TAS....

Then I procced to do some "control experiment" at offline....In the same condition , the limit is about 620kph - 630kph....

I am still trying to catch something more dramatic which can show the problem more clearly. For example, those incident that I can break my wing with less then half stick deflection and a speed below 550 kph...

Diablo310th
01-15-2005, 10:05 PM
Hyperion.....were the trks you recorded g's on possibly offline?? DeviceLink will record such information offline jsut not online.