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

  • Thread starter MGR
  • 1,131 comments
  • 90,706 views
My league has had best results in open lobby, fixed host with mics. disabled.

When slow lap times are present any 1 car is not lagging, skidding, smoking or pulsing around the track.

If any 1 driver is lagging slow laps times are not present. This has not been tested for consistency.

The screenshot below is from lap 2 of a 16 driver race. Open lobby, fixed, mics. disabled. The pit crew is floating as if a car has left the pits, they should be in the ready position. The pit crew was not seen floating by anyone during the race. Replays should not be corrupt but it seems they are. I do now replays show all cars doing strange things except for the car of the replay host. Is it possible the replay timing is bad as well?

ecfaaf06.jpg
 
I just watched this replay, but I did not come to the same conclusion. To address your last point first, yes Gooners can corner better, when everyone is in the same car then cornering speed is the only thing that determines the winner. Would he be making the time in a straight line? It is natural for cornering speed to decrease with a persons increase in lap time. Also he is harder than anyone else on his front tyres indicating that he is getting more grip out of them, thus he is faster. After just two laps there is a lot a wear while others back in the pack are still a little cold. It is his driving style, not a physics thing.

As for the time inaccuracy, it is very obvious as soon as the replay loads. I was quite surprised actually. But it is not the clock that is sped up, the whole replay looks that way to me. The cars move way to fast, plus the sound is high pitched.

Surly if the race was like this someone would have noticed? Didn't anyone mention how fast the clock was going? I bet not because I think it would have been running normally and its the reply that is corrupt.

EIDIT: watching that replay again... The speed is comical in places lol. There is NO WAY the actual race was like this and nobody noticed. Watch it for yourselves guys.
But gooners is very clearly not the best driver of that group, no offense to him, he's not even close, even he'll tell you that.
I had a feeling you'd say that though, I was surprised you didn't earlier.

But once again, you haven't addressed the actual issue. Why is the better part of the field running 3-5 seconds slower for their best lap in the race vs qualifying?
You don't have an answer for that, nor do you care to find one, you're looking for reasons to say it's not happening.
Make a Youtube video and show us NTSC folk this "comical" replay.
 
But gooners is very clearly not the best driver of that group, no offense to him, he's not even close, even he'll tell you that.
I had a feeling you'd say that though, I was surprised you didn't earlier.

But once again, you haven't addressed the actual issue. Why is the better part of the field running 3-5 seconds slower for their best lap in the race vs qualifying?
You don't have an answer for that, nor do you care to find one, you're looking for reasons to say it's not happening.
Make a Youtube video and show us NTSC folk this "comical" replay.

I am not surprised they could not match their best times. The pressure of a race is much higher than that of qualifying.

In qualifying you have lots of chances to set your best lap time. In the race you get 1 lap, that is the second lap after your last pitstop. That is the only time in the race you are light on fuel and have good tyres.

But why put down gooners? I watched his lines and driving, it was great, better than anyone else I watched in the race. I am watching him now hitting apex after apex and being very tidy. Why should that be slow?

Oh and one last thing... It is convenient that nobody has mentioned that gooners is one of only 2 drivers on racing softs (on the lap I am watching).... But yeah its a huge mystery why people on racing hards can not match their qualifying times on racing softs. Cornering speed too.. No idea how someone on racing softs can corner faster than someone on racing hards... Must be the network data physics... Must be.

My god. You all changed tyres throughout the race to different compounds but expect all the times to match? Please people think a little.
 
Last edited:
I am not surprised they could not match their best times. The pressure of a race is much higher than that of qualifying.

In qualifying you have lots of chances to set your best lap time. In the race you get 1 lap, that is the second lap after your last pitstop. That is the only time in the race you are light on fuel and have good tyres.

But why put down gooners? I watched his lines and driving, it was great, better than anyone else I watched in the race. I am watching him now hitting apex after apex and being very tidy. Why should that be slow?
Edit: Rephrase - Oh really?

NAResults3.png

3D3_20111207_Event13_Results_NA.png
 
Last edited:
nasanu
I am not surprised they could not match their best times. The pressure of a race is much higher than that of qualifying.

In qualifying you have lots of chances to set your best lap time. In the race you get 1 lap, that is the second lap after your last pitstop. That is the only time in the race you are light on fuel and have good tyres.

But why put down gooners? I watched his lines and driving, it was great, better than anyone else I watched in the race. I am watching him now hitting apex after apex and being very tidy. Why should that be slow?

Oh and one last thing... It is convenient that nobody has mentioned that gooners is one of only 2 drivers on racing softs.... But yeah its a huge mystery why people on racing hards can not match their qualifying times on racing softs...

Our crew always hits their qualifying marks during a race. So do all the other top league leaders during a race. When the racing is clean and room is given by other drivers who are able to anticipate moves other than their own, fast laps are present in all races which match qualifying, even without the benefit of low fuel.
 
My league has had best results in open lobby, fixed host with mics. disabled.

When slow lap times are present any 1 car is not lagging, skidding, smoking or pulsing around the track.

If any 1 driver is lagging slow laps times are not present. This has not been tested for consistency.

The screenshot below is from lap 2 of a 16 driver race. Open lobby, fixed, mics. disabled. The pit crew is floating as if a car has left the pits, they should be in the ready position. The pit crew was not seen floating by anyone during the race. Replays should not be corrupt but it seems they are. I do now replays show all cars doing strange things except for the car of the replay host. Is it possible the replay timing is bad as well?
I think you're allowing nasanu to much room to side track the thread.
It doesn't matter if the replays are corrupt or not, there is clearly an issue where some people are lacking massive amounts of grip in online races, grip they had in qualifying.
We're here to figure out what's causing the loss of grip between free run and races.
Replays are irrelevant, the only thing knowing they're corrupt can help, is to not use them as a tool for sorting the actual problem we're after..
 
So you are all just going to skip over the fact that he was on different tyres for most of the race? Just skip over qualifying being on softs and the race mainly run on hards and mediums?..

You guys don't seem interested in the truth.
 
During a race, do any of you experience anything strange due to this time bug/glitch? Loss of grip (sometimes no grip at all) or extremely heavy steering? Or am I describing something completely unrelated? This has happened to me twice (first month the game was released and just a few days ago) and was just curious if both were related. Or do you figure the glitch was in effect while watching the replay only?
 
Last edited:
Without going into the whole thread, OP is 100% correct from my experience.

My community is experiencing this bug for a while now, under the all mentioned conditions:

- room with 12+ people
- Premium cars
- race only, free run is ok

However, I can tell it is not really wheel related, since we all have wheels. After many testing we concluded how this handicap reflects on person(s) with the highest upload/bandwidth speed.

We have been successfully resolving it with moving into new Lobby while recent option to force limitations for number of people in Lobbies made it less occurring.
 
So you are all just going to skip over the fact that he was on different tyres for most of the race? Just skip over qualifying being on softs and the race mainly run on hards and mediums?..

You guys don't seem interested in the truth.
Tires: Open, but all drivers must qualify and start on Racing Soft Tires (R3's)
Oh really?

During a race, do any of you experience anything strange due to this time bug/glitch? Loss of grip (sometimes no grip at all) or extremely heavy steering? Or am I describing something completely unrelated? This has happened to me twice (first month the game was released and just a few days ago) and was just curious if both were related. Or do you figure the glitch was in effect while watching the replay only?
I haven't encountered it personally more then a 2 second difference once or twice. I have experienced some extreme frame rate drops though, in which case I run terribly slow laps.
Whether that's the same issue others are experiencing I do not know.

Without going into the whole thread, OP is 100% correct from my experience.

My community is experiencing this bug for a while now, under the all mentioned conditions:

- room with 12+ people
- Premium cars
- race only, free run is ok

However, I can tell it is not really wheel related, since we all have wheels. After many testing we concluded how this handicap reflects on person(s) with the highest upload/bandwidth speed.

We have been successfully resolving it with moving into new Lobby while recent option to force limitations for number of people in Lobbies made it less occurring.
So you quit if someone is falling far behind or...?
I'm just confused since it only happens in races, how you would know to switch to a new lobby.
 
I've PM'd Gooners to do the same test as I did with timing the laptimes with a stopwatch.
As soon as I get his results, I'll post them here and the we will have a better view on how different each replay is...

Oh, BTW, that Tokyo race had a 1 lap qualifying...

In qualifying you have lots of chances to set your best lap time. In the race you get 1 lap, that is the second lap after your last pitstop. That is the only time in the race you are light on fuel and have good tyres.
 
So you quit if someone is falling far behind or...?
I'm just confused since it only happens in races, how you would know to switch to a new lobby.

Basically yes.

We all have headsets pretty much and it is very easy to *detect* this *glitch* since persons(s) experiencing it are just too slow from the very start.

Lousy acceleration, low grip, no way to catch anyone, etc.

We all quit, everybody restarts routers, clears cache (we have the *procedures* from that since the beginning) and than move to another persons Lobby as soon as the "yellow icon" is established.

However, as I said, it probably have to do with something related to way how server is distributing bandwidth once the race starts. We noticed it happening only to the members with the highest upload/bandwidth so there must be some issue regarding *bottlenecking* and distribution of the connection on the server-side.
 
During a race, do any of you experience anything strange due to this time bug/glitch? Loss of grip (sometimes no grip at all) or extremely heavy steering? Or am I describing something completely unrelated? This has happened to me twice (first month the game was released and just a few days ago) and was just curious if both were related. Or do you figure the glitch was in effect while watching the replay only?

That's it. Same thing.

All sorts of weird things start happening even though it might look all normal and peachy at first glance.

My issues:
1) mid corner grip becomes inconsistent and slick. Unaffected drivers will be able to drive right around the outside of you like you were standing still, or stuck on black ice. Happens to me most on Suzuka's sector 1.

2) Game's recorded lap time will not correspond to the real world lap time.

3) The guys up front seem really, really, really fast and you (plus others in the room similarly affected) loose ground by the bucket load. This seems to happen both ways for me.

I regularly race in a 500pp room hosted by a Singaporean, with entrants from all over the world. Australia, USA, UK, Greece, Taiwan. There is one particular Taiwanese guy who is always extremely, unnaturally fast, whatever car he drives. Detractors will say that the guy knows how to tune while everybody else does not, but his rate of acceleration is so blazingly quick that something is obviously not right.

And then rarely, I find myself in the winning position so far ahead that I can cruise and still stay 20 secs ahead of the pack. The first time it happened it was in a Spa race. I was so bewildered that I saved the replay to watch. I thought everybody else was just cruising to, but no. The whole pack was going flate out, 5 cars separated by less than 3 seconds, far behind me, while I was 12 secs ahead. I was cruising at maybe 85%, and they couldn't catch up.

Now, this is addressed to the guys who say that this has never happened to them: If it's a big race in an open lobby, take note of the last 3 guys in the race. Are they falling back quickly? It's easy to write them off as noobs, but perhaps save and watch the replay. They could be running ragged with buggy handling. It doesn't happen to you, but that's not to say that it doesn't hit others in the room. Some are genuine noobs, but some are just affected by the bug, and get really frustrated at why their cars are all messed-up, and just rage quit.

And yes it's obvious that there's an unconstructive troll hanging around here. The rest of us, please stay on track.

Whatever the cause, if PD can fix the rolling start bug, here's hoping that enough noise will make them look into this. Otherwise, online racing for some of us is just pointless.
 
So you are all just going to skip over the fact that he was on different tyres for most of the race? Just skip over qualifying being on softs and the race mainly run on hards and mediums?..

You guys don't seem interested in the truth.

Nasanu,please man,its very obvious that there is an issue here!!
As i wrote i also never experienced this,but you cannot deny after reading all these posts by well respected and experienced GT5 players,the "bug" is real!!

I don't know who you are trying to fool with your ignorance,but you are not helping us by any means...so or just move along or try to help solving or understanding where this issue is coming from or what causes it or how we can eleminate it... !!!!

Thanks.



Spy.
 
nasanu
So you are all just going to skip over the fact that he was on different tyres for most of the race? Just skip over qualifying being on softs and the race mainly run on hards and mediums?..

You guys don't seem interested in the truth.

"I deride your truth-handling ability!"

Seriously, no. Firstly, ignoring the outlap, quali effectively only gave you one-two laps to hit a time in, which would constitute nearly as much pressure as the race situation, not that it matters.

Secondly, absolutely no-one was expecting to hit the same times on R2s or R1s, and some people had practiced on multiple compounds, although this was just after 2.02, so pit and tyre strategies varied. Everyone had to start on R3s, in any case, and most strategies involved more than one stint on them.

I assure you that history shows that several of the participants (not myself, so much) are very experienced, consistent racers, well able to handle pressure, and not infrequently better quali times in-race (especially given short stint quali, draft, and low fuel load late).

These were the largest fields we had yet had, with widely geographically distributed participants, with varying quality connections. I can tell you that my replay will not be exactly the same as Speedy's, but I have yet to examine it for timing issues. Y'know, quantitatively.

For someone claiming to want testing, data and large sample spaces, you seem to be worried about crows to the point that a rather large number of unnecessary straw men are required.

Feel free to actually suggest a concrete testing methodology, or even a single falsifiable test, rather than asserting that no one involved realized that changing tires would affect lap times.

Does anyone here have figures for the bandwidth involved per participant at the various quality levels? I don't have a hub or managed switch at home, but I could probably do so if sufficiently motivated.
 
Last edited:
Basically yes.

We all have headsets pretty much and it is very easy to *detect* this *glitch* since persons(s) experiencing it are just too slow from the very start.

Lousy acceleration, low grip, no way to catch anyone, etc.



However, as I said, it probably have to do with something related to way how server is distributing bandwidth once the race starts. We noticed it happening only to the members with the highest upload/bandwidth so there must be some issue regarding *bottlenecking* and distribution of the connection on the server-side.

I think you have a very important point there we are having same problems and also noticed that it happens only to people with high upload/bandwidth !!!
 
And how do you know this is not normal? Have you also tested races you think are fine?

Even worse, read the (my) posts before replying.

As I wrote, I fortunately only experienced one race when this affected me. When 15 racers there was a 4.2 sec delat between race clock and stopwatch but as people left the room the delta declined to almost zero. As I also wrote, I measured at three other races in the replay, and the delta was zero at all other occations.

What is your objective. If you don't have this issue, good for you, and move on.
 
I think you have a very important point there we are having same problems and also noticed that it happens only to people with high upload/bandwidth !!!

Now I'm not very with the the tech specifics, but I ran a connection test using the PS3's own built-in tester, and on my system I got Download speed: 4.9mbps, Upload speed: 298kpbs (!)

Considering that this bug is almost crippling my online racing experience, how does the connection speed sound to you guys?
 
I think you have a very important point there we are having same problems and also noticed that it happens only to people with high upload/bandwidth !!!

I am on 100/100 and never had the problem, just like spyrrari.

So "problem only happens to high-bandwidth users" is at least not the same as "high-bandwidth users always get the problem".
 
That bit is your imagination. It can not happen. The network has nothing at all to do with physics calculations.
Not physics, but with timing? If physics and timing are intertwined it could be inderect interference -maybe, if PD screwed something up bad enough. If there's some huge basic flaw with the way PD wrote the netcode, almost anything is possible, so I wouldn't ever say ''never''.

I don't understand how you think it all works. Games only send location data and very little else. Do you think the host PS3 is processing physics data for all players? Really how do you think it works?
It doesn't send the physics calculations over the internet, obviously.

The game sends and receives location data, and mabye velocity, but your PS3 still calculates the physics for every single person on the track. The stream of location data is not fast enough to represent a smooth motion most of the time, so the PS3 fills in the rest of the frames with self generated physics data from one set of location data to the next. Generally in race sims, the cars closest to you are given the highest priority for the physics data as well as location data, that makes sense, since they are most visible and can affect you the most.

Maybe that's something else to look into.. people in close packs being more affected than someone on their own?

Given that GT5 taxes the PS3 pretty hard anyway, all those calculations could slow down your PS3 to the point that it causes problem with timing and yes, physics calculations -ESPECIALLY if PD made a huge mistake in the way they wrote the netcode or there's some huge bug in it, or both. Think about what they are giving CPU cycles first, or how they are generating the location data. There could be a flaw or bug in any number of steps. What if they did something silly to try to keep the GFX framerate up? That's my theory, could be wrong. Anything is possible, so it's best not to say something ''is not possible'' when it comes to computer software...
 
Now I'm not very with the the tech specifics, but I ran a connection test using the PS3's own built-in tester, and on my system I got Download speed: 4.9mbps, Upload speed: 298kpbs (!)

Considering that this bug is almost crippling my online racing experience, how does the connection speed sound to you guys?

Your speeds are pretty basic IMO as I have 19.3Mbps download and 2.9Mbps upload speed

EDIT: Also, my ping results mostly around 24ms.
This was tested through www.speedtest.net
 
Last edited:
I am on 100/100 and never had the problem, just like spyrrari.

So "problem only happens to high-bandwidth users" is at least not the same as "high-bandwidth users always get the problem".

I am not saying always but when it happens then only high-bandwidth users are effected in our groups of racers
 
I am not saying always but when it happens then only high-bandwidth users are effected in our groups of racers

Same here.

Tech-savy member of our community explained us how the issue is almost certainly on the server-side, where server takes the bandwidth of the most *capable* members and starts distributing the packages through their upload.

Because of the great discrepancy among players - for instance I have 5Mb upload, but majority of my friends are having only 512kb and lower - server just "suffocates" us on our side and basically creates a bottleneck because of that great discrepancy. Result of such "dividing" the upload is *suffocation* of high-upload players where their own bandwidth becomes a victim of other player's slow upload.

He was very dedicated in explaining how GT5 networking uses method of "adjusting" the overall bandwidth for the players with slowest upload. But it seems how that method produces crazy results when discrepancy among players-upload is more pronounced.

I really do not have deep knowledge about networking but it sounds plausible really.
 
Last edited:
I am not saying always but when it happens then only high-bandwidth users are effected in our groups of racers

I know 👍 - just trying to add some observation so we can get to the bottom of this.

Others have complained about low-bandwidth players getting hit more, that would also sound right intuitively. But your observation says the opposite, that is interesting.

I wonder if the PS3 can get bogged down in network traffic. About a year ago I had a loop in my network, 100000+ junk packets/second circulating. It lasted for hours, and happened while B-spec Bob was doing a 24h enduro, which ended up taking far more than 24 wall-clock hours! So the PS3 had become overwhelmed and as a result slowed it's clock and affected the game in a huge way.

Now this is an extreme case, you can get nowhere that amount of traffic on a normal connection, but if you say it mostly happens to high-bandwidth users I wonder if something like this could still be the case, on a smaller scale. I would like to know if these high-bandwidth users actually had a high traffic rate.

If so a solution could be to limit the PS3's connection (bandwidth limits in the router).

-- EDIT --

Tech-savy member of our community explained us how the issue is almost certainly on the server-side, where server takes the bandwidth of the most *capable* members and starts distributing the packages through their upload.

Wow, never heard of that, I thought each PS3 just had a connection to every other in the room.

This I would like to test. Not much time right now, but in a few days I will have my PS3's traffic running through a testing device so I can get some numbers on packet rates and jitter (unless I am treed to it).
 
Last edited:
Tech-savy member of our community explained us how the issue is almost certainly on the server-side, where server takes the bandwidth of the most *capable* members and starts distributing the packages through their upload.

Very interesting.

In the example I wrote on page 7, the only time where obviously our racing league got hit with this issue, the fist driver to quite in frustration was one having a fiber connection. Not much data to go by though.

I only been affected once and I have a fairly low bandwidth at 8/1mb but with good ping rate.

Couple of pages up there was a member showing a table with different drivers affected differently in the same race. It would be very interesting to know their interconnection speed. If correlation, we might have found the root cause of the issue.
 
Nasanu,please man,its very obvious that there is an issue here!!
As i wrote i also never experienced this,but you cannot deny after reading all these posts by well respected and experienced GT5 players,the "bug" is real!!

I don't know who you are trying to fool with your ignorance,but you are not helping us by any means...so or just move along or try to help solving or understanding where this issue is coming from or what causes it or how we can eleminate it... !!!!

Thanks.



Spy.

Sigh...

I am not saying there is no issue. I am just pointing out the stupidity of some of the theories on display here. First its all about graphics, then replays, then bandwidth etc...

I seem to be the only one here with the slightest clue how games programming works (things like this should be obvious to even non programers though) because many of the suggestions are absurd.

So far the only evidence that has been presented is only evidence of a replay bug. As for a slow driver winning, he got good start and was the only one to use softs after the first stint. If you get out in front while others are fighting for position its easy for slow drivers to create a lead. Then he was on faster tyres then others. Plus there is still zero evidence of the timing issue happening during live play.

You guys can believe what you want without a single scrap of evidence, believe the color car determines who is effected if you like (because that is just as valid as anything suggested here).

I don't understand how you think you are helping each other by discussing these idiotic theories. You will never work it out until you are able to listen to reason and logic.
 
Here's a copy (PAL) of the replay of this race.

event16_res_eu.png


To be found here on the forum.

I watched the replay again and timed the laps with my watch's stopwatch.
My conclusion here was that the racetimer in GT5 is incorrect.
Here are some compared laptimes all timed on lap 2 of the race.

Myself, GTP_Speedy6543 (DFGT)
  • Racetimer: 1'53
  • My stopwatch: 1'39
GTP_gooners (DS3)
  • Racetimer: 1'37
  • My stopwatch: 1'25
GTP_Hydro (DFGT)
  • Racetimer: 1'51
  • My stopwatch: 1'37

I now devided all times in GT5 by the real times to see how long one second is in this GT5 online race.
GTP_Speedy: 113sec/99sec = 1.141414sec
GTP_Gooners: 97sec/85sec = 1.141176sec
GTP_hydro: 111sec/97sec = 1.144329sec

Conclusion: GT5's in race seconds last for ~1.14 real life seconds.

So the timer is equally different for everyone.
However, when you will watch the PAL replay you will notice that the first driver, GTP_Gooners (DS3), is able to maintain higher speeds through the turns than anyone else on track. I think all other drivers where using a wheel. EDIT: GTP_MadMax86 is on DS3 too

So there must be something going on with the grip here.

Speedy,

Any chance you can find out the connection speed of the drivers above. I know its asking a lot from you, so if you can't its fully understandable, but if you can it might lead us to the answer.
 
Same here.

Tech-savy member of our community explained us how the issue is almost certainly on the server-side, where server takes the bandwidth of the most *capable* members and starts distributing the packages through their upload.

Because of the great discrepancy among players - for instance I have 5Mb upload, but majority of my friends are having only 512kb and lower - server just "suffocates" us on our side and basically creates a bottleneck because of that great discrepancy. Result of such "dividing" the upload is *suffocation* of high-upload players where their own bandwidth becomes a victim of other player's slow upload.

He was very dedicated in explaining how GT5 networking uses method of "adjusting" the overall bandwidth for the players with slowest upload. But it seems how that method produces crazy results when discrepancy among players-upload is more pronounced.

I really do not have deep knowledge about networking but it sounds plausible really.

But this would only explain lag, not a loss of grip. For this to be real there MUST be some sort of connection with the local driving model and the network traffic. Can you ask him about this?

Also this is how many driving games have worked over the years including GT5P. I don't remember anyone complaining of this then.
 
Back