GT5's Game-Breaking Online Flaw (OP Updated: 11 Feb)

  • Thread starter MGR
  • 1,131 comments
  • 96,669 views
Well done guys for all of the hard work with the data on this problem. At least if PD are looking at this they have some great info to work with.

I hope the 18th update brings some improvement, if not I am sure the next update should address this problem; to what success I don't know but fingers crossed.

Our season should have kicked off yesterday but we have postoned till next week to see if the update sorts things out.
 
The way I see it, the server(or rather it's CPU and/or RAM) is getting overloaded with data that needs to be processed. In other words, the server does all the math for speed, distance, FFB etc. and then sends the info to the clients. That could explain why the timer is slower.

That would also explain why there are about 12 cars in single player.

Note that I don't have a PS3 and I'm basing my experience on multiplayer games for PC ,so I could be wrong. If i'm wrong, please explain were
 
I think what we are seeing here is related to the entire online vs. offline dual-physics model flaw. The online and offline physics differences have been a huge problem since GT5 was released!

Now, I do realize that there appears to be a new bug on top of that which seems to be exacerbating the problem as there are more and more cars on the track. And it is very nice to see that finally this flaw has some renewed attention and some further testing. It is especially interesting to see that this flaw may be specific to ffb wheel users. That explains some of the earlier disagreement over the severity of this issue. But the underlying flaw itself is in no way a "new" bug and I'm convinced that what we are seeing is somehow related to the fundamental differences in online vs. offline physics. My guess is that this newest bug is amplifying the dual-physics model issue to the point where suddenly a LOT more people are discovering it. It's been there all along and it has NEVER been an acceptable part of the GT5 experience. If PD ever fixed a single thing this should have been it: the offline and online handling physics should be identical, period.

Offline and Online handling of the Alpine 1600s

Stock Yellowbird Steady State Sweeper issue in online mode

And here is the original thread where we all started trying to figure out what the heck was going on. I was one of the people going about it from a relatively scientific point of view, so I did a LOT of testing regarding the issue, which varies in severity considerably from car to car and track to track:

Original Online vs. offline physics changes thread

I think it is important to draw attention to this not necessarily being a new problem, even if it has suddenly gotten worse for busy online races and ffb wheel users.

I'm really happy to see this problem/problems finally getting a bit more attention, as for me this has been the killer-flaw since release.
 
I copy-paste here from news page:

"Maybe GT5 can not skip frames correctly and slowdown in FPS due high computational load causes also real slowdown of lap times?"

Edit: Thanks panjandrum for the videos. Watch those PD!
 
Last edited:
jon-
Likewise the game seems to make 2 people arch rivals so when one joins the room the other gets kicked.
Click to expand...

We had that problem at ICGTC too, it was fixed when everyone put their ps3 to a fixed ip and port forwarded the ports the game manual tells you to open.

It seems you can get NAT2 even if you don't deserve it. It's not NAT3, but still it gives you this loading bar under someone's name in the lobby and guys can't see each other on the track. We call it NAT2,5 ;)

As a example I put my router in DMZ mode for my racing PS3 and it tells me I'm on NAT2. But I have another PS3 in my living room for casual gaming and movie watching. This is on the same router and are because of this blocked. Still it gets NAT2.

Also I disabled UPnP as I couldn't get it to work with those 3 routers I tested, I needed to port forward manually for it to work properly. And after everyone did the same thing the loading bar and people disconnecting seems to have been history.
 
ErikAndre71
We had that problem at ICGTC too, it was fixed when everyone put their ps3 to a fixed ip and port forwarded the ports the game manual tells you to open.

It seems you can get NAT2 even if you don't deserve it. It's not NAT3, but still it gives you this loading bar under someone's name in the lobby and guys can't see each other on the track. We call it NAT2,5 ;)

As a example I put my router in DMZ mode for my racing PS3 and it tells me I'm on NAT2. But I have another PS3 in my living room for casual gaming and movie watching. This is on the same router and are because of this blocked. Still it gets NAT2.

Also I disabled UPnP as I couldn't get it to work with those 3 routers I tested, I needed to port forward manually for it to work properly. And after everyone did the same thing the loading bar and people disconnecting seems to have been history.
Click to expand...

Thanks, very helpful post. We'll give it a try.
 
panjandrum
I think what we are seeing here is related to the entire online vs. offline dual-physics model flaw. The online and offline physics differences have been a huge problem since GT5 was released!

Now, I do realize that there appears to be a new bug on top of that which seems to be exacerbating the problem as there are more and more cars on the track. And it is very nice to see that finally this flaw has some renewed attention and some further testing. It is especially interesting to see that this flaw may be specific to ffb wheel users. That explains some of the earlier disagreement over the severity of this issue. But the underlying flaw itself is in no way a "new" bug and I'm convinced that what we are seeing is somehow related to the fundamental differences in online vs. offline physics. My guess is that this newest bug is amplifying the dual-physics model issue to the point where suddenly a LOT more people are discovering it. It's been there all along and it has NEVER been an acceptable part of the GT5 experience. If PD ever fixed a single thing this should have been it: the offline and online handling physics should be identical, period.

Offline and Online handling of the Alpine 1600s

Stock Yellowbird Steady State Sweeper issue in online mode

And here is the original thread where we all started trying to figure out what the heck was going on. I was one of the people going about it from a relatively scientific point of view, so I did a LOT of testing regarding the issue, which varies in severity considerably from car to car and track to track:

Original Online vs. offline physics changes thread

I think it is important to draw attention to this not necessarily being a new problem, even if it has suddenly gotten worse for busy online races and ffb wheel users.

I'm really happy to see this problem/problems finally getting a bit more attention, as for me this has been the killer-flaw since release.
Click to expand...
I've been wondering this for some time now.
It seems to have all the same characteristics in killing performance, right down to cars being slower in a straight line.

It's like the offline/online physics difference gets amplified in full "bugged" rooms.
Which also leads me to believing it's a core problem, that won't be solved anytime soon. Hopefully they can at least make it "good" like it used to be.
 
CSLACR
So it's essentially possibly a graph of the bug's work in progress?
Click to expand...
I'd have to do more testing before I can say 'probably'. Gonna try to record a race replay (or rather just the clock) right off TV screen with my only camera that is capable of 60fps and see if I can get cleaner pics.
 
Hi everybody

Haven't had much time in the last couple of days to add anything new (with the birth of our third child and all) but I would like to thank everybody for their contributions to the thread.

Massive thanks to Famine, Jordan and the rest of the GTP testers involved in verifying the bug so it could be posted on the front news page.

I've updated the OP and compiled some of my more useful data.

Thanks again for your help.
 
MGR
Hi everybody

Haven't had much time in the last couple of days to add anything new (with the birth of our third child and all) but I would like to thank everybody for their contributions to the thread.

Massive thanks to Famine, Jordan and the rest of the GTP testers involved in verifying the bug so it could be posted on the front news page.

I've updated the OP and compiled some of my more useful data.

Thanks again for your help.
Click to expand...

Congratulations :cheers:
And thanks again for your continued efforts.
Great job 👍
 
Congratulations on the baby :cheers:

So much more important than some random game's hopeless flaws.
 
I'd like to send out my thanks to everyone that tested the issues, and to the mods & staff for handling the issue really well and fortunately we now have a blog post on the front page :)

I hope we can all resume our respective GT5 championships soon :)
 
Arkano5
I'd like to send out my thanks to everyone that tested the issues, and to the mods & staff for handling the issue really well and fortunately we now have a blog post on the front page :)

I hope we can all resume our respective GT5 championships soon :)
Click to expand...

Resume! Just get creative. If you have 16 guys that want to race in a series use two lounges and a sprint/feature bracket system and limit each room to 8 racers. Top 4 advance to Feature A, bottom four to Feature B and adjust points accordingly. Or run some sort of pre-quali where you can divide up into divisions ahead of time.

Use the lounge options to turn off visible damage and set other settings to minimize the likelihood/impact of the "bug" and then just race. My experience shows you can have a very good race if you stay at 10 or lower. There is still a chance a driver or two will be "hit" by the bug, but there was a chance before they would get DC'd or experience lag or suffer from invisible driver issues so this is nothing new.

What you can't do is expect 14-16 car fields to function well, but that doesn't mean you can't race and compete and have fun while we are waiting for PD to address the issue. I can tell you the WRS-Online series plans to continue racing and adapt to the conditions with the racing formats we offer.
 
Last edited:
MGR
Hi everybody

Haven't had much time in the last couple of days to add anything new (with the birth of our third child and all) but I would like to thank everybody for their contributions to the thread.

Massive thanks to Famine, Jordan and the rest of the GTP testers involved in verifying the bug so it could be posted on the front news page.

I've updated the OP and compiled some of my more useful data.

Thanks again for your help.
Click to expand...

Congrats and thanks for your effort on the very detailed test results. 👍
 
Just now I remembered something that I really wanted to ask here but sadly forgot till now. I think it has something to do with the physics theory in the OP.

Last year I did the 24h LeMans A-spec race. I choose my Audi R10 (roughly 300km mileage by that time), before the race i washed it and changed it oil. I always raced in 2 hour stints, then saved and restarted. During that time, GT5 was patched (it was the patch which came with the first dlc`s) and I was afraid that by logging into my PSN account while starting PS3, could force me doing the 24 hours race again (I did like 10 hours at that point). So I played offline till I finished the race.

Then something very strange happend: When I restarted the game at 14 or 16 hours, the car felt VERY strange. I didnt know what happened because I never had any problems of this kind with the R10, it felt like the car had some kind of input lag.
When I steered right it took the about half a second to do this. Same with braking or throttle (remember, we have to graph which shows the throttle/brake paddle position). I tested my DS3 and my DFGT, nothing changed. So I saved and tried again next day.... Everything was fine then.

I also remember that this accourd me also a few times while racing online. F.E. On some servers I was very competitive in qualifying, setting the fastest or 2nd fastest lap, but suddenly the cars physics changing in the race and me spinning out/driving very very bad.

I dont know if this helps, but this happened to me before the update which came with the first batch of DLC cars. And this phenomen could be explained by the "missing frame" theory by the OP.
 
Thank you everyone for publicising this, hopefully his will test if PD pay attention us really or not...
 
Thanks again MGR and everyone else who did research on this. We've run some 12 player races w/a mic disabled room and all damage off/fixed host. No evidence of clock issues.

Even before this came to light, it was really hit or miss in an open lobby w/ 14 or more on track. I stopped hitting random open lobbies because of so many connection issues. Perhaps this "bug" has always existed like has been mentioned. Sounds like an inherent problem w/ system overload.
 
ncrthree
Thanks again MGR and everyone else who did research on this. We've run some 12 player races w/a mic disabled room and all damage off/fixed host. No evidence of clock issues.

Even before this came to light, it was really hit or miss in an open lobby w/ 14 or more on track. I stopped hitting random open lobbies because of so many connection issues. Perhaps this "bug" has always existed like has been mentioned. Sounds like an inherent problem w/ system overload.
Click to expand...
It's actually been proven with links in this thread, it has been mentioned and semi-documented back to April of 2011.(at least)
It seems to have been exaggerated in one of the recent updates, specifically 2.03.
 
Interesting vid Speedy.

Did you take note of the bizarre readings on the speedometers of the other cars? you can see it especially well on the long straight. They seem to fluctuate in a way that seems impossible. (And is certainly not what you would expect to see, especially with WRS drivers.)

It seems like this may be evidence that a saved replay is only accurate for the organic user, and show how poorly the physics model deals with opponent cars on the circuit. Can anyone check a replay that they know NOT to be affected by the bug and see if the same thing occurs in regards to displayed speeds of other people's cars?
 
Hey, I just saw a comment on the first page and then reread the op, and it made me think about a potential cause.

In BermudaKid's LeMans series, we had 12 people racing at Sarthe. Laptimes were normal, and race quality was set to high.

Someone in the first page mentioned race quality, which hasn't been tested yet.

Perhaps race quality is the issue to this whole mess???
 
dudeondacouch
Interesting vid Speedy.
Click to expand...

Agree 👍

dudeondacouch
Did you take note of the bizarre readings on the speedometers of the other cars? you can see it especially well on the long straight. They seem to fluctuate in a way that seems impossible. (And is certainly not what you would expect to see, especially with WRS drivers.)

It seems like this may be evidence that a saved replay is only accurate for the organic user, and show how poorly the physics model deals with opponent cars on the circuit. Can anyone check a replay that they know NOT to be affected by the bug and see if the same thing occurs in regards to displayed speeds of other people's cars?
Click to expand...

Well spotted 👍

But I think it is just lag and not the same as the grip-loss bug. Lack of information about where the car actually is/was, so the game just makes something up to fill the holes. It would not look like that in the other drivers' replay.

I think you can see the same in all replays where there was any lag between you and the driver you are watching.
 

Similar threads

Famine
Forza Motorsport Photomode Bug Adds a Limited Home Area Free-Roam Feature
Replies
7
Views
2K
ShavedBadger
ShavedBadger
Falcon787B
B-Spec: AI Pits On Lap 1?
Replies
2
Views
2K
Falcon787B
Falcon787B
75S30
PS5 - GT7 Replay Storage Error
Replies
2
Views
1K
jrbabbitt
jrbabbitt
Sir Crashalot
Replay bug?
Replies
6
Views
1K
DJShadesUK
DJShadesUK
wlashack
How to report bug in GT7?
Replies
4
Views
2K
Famine
Famine

Latest Posts

Back