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

[windmill]

I wouldn't expect any other reply from Sony NL - it's not their problem.

Not there problem and PD does not answer emails of you cannot call them so its nobody's problem except then the players problem who payed for a not playable game .

Maybe an idea for Sony for GT6 do NOT put Sony logo's on the disc of casing !!

 

[Speedy6543]

Made an official complaint a week ago with official scee number yesterday after asking by telephone what was happening a Sony official on the telephone gave me that answer .

Ok

Also, if you want to answer mutliple posts, try using the 'Multi-Quote This Message' button

instead of triple posting... 👍

 

[esh]

I kid you not! It was simply unbelievable how slow the game was running.

Made it really easy to spot the speed differences. I already knew the that using a G25 wheel made quite a difference, but in this race I also noticed the DS3 users without mics seemed to have more pace than anyone. Once I changed to the DS3 and turned off my mic I could cruise past other DS3 users still using mics.

Free run before the race (with all 14 cars on the track) seemed perfectly fine.

We bailed on the rest of the MX5 races and just ran some shuffle events for the majority of the evening - limiting them to 10 competitors at a time while the other 3 or 4 drivers took turns sitting them out just watching instead. Whether or not the update was complete at that point is another story but the content of the update probably had already been decided upon.



I don't have much hope 2.03 will fix this issue Jordan.

This bug has been around for quite a while. Well before the Spec 2.0 patch (although it does seem to have gotten worse somewhere along the line) It's either too hard to fix or PD simply aren't aware of it.

I'm hoping it's the latter and we just need to draw as much attention to it as possible.

It's too early to say PD are not aware of it as the update cycle is always going to be 1-2 patches behind, meaning the screens we saw for 2.03 that were leaked and then official were dated Dec. 16th, before 2.02 even came out on Dec. 20th. Whether or not the update was complete at that point is another story but the content of the update probably had already been decided upon.


I would not be concerned that it isn't fixed in this update and possibly the next although the fact that a timeframe hasn't been mentioned for an update after 2.03 besides the generic monthly cycle Kaz talked about @ SEMA.

 

[chukabilly]

I've been looking for an over an hour, but I can't find the original post. The point was: PD gave preference in its netcode for the visual engine to take priority over the physics engine. So during on-line races, the game is constantly trying to make everything LOOK perfect. Meanwhile, the physics engine is trying to compensate for a lack of computing power, and when FFB and geography comes into play, the physics break down.

The visual component has the power, but doesn't get the input it needs because the physics aren't sending input fast enough. So the result is the visuals give you whatever it gets, but that input is "staggered", resulting in frame rate issues.

As this is happening, any driver input is momentarily muted by the conflict between physics and visuals, resulting in a "coasting" effect, not to mention a lack of steering input at critical times. The result is a lap that looks better but is much slower.

I am not sure, Speedy, if you were in that thread or not. Maybe you remember, or maybe it was not you. I just know it was a WRS racer who was having the conversation with said programmer. He also mentioned that the PS3 is already maxed out by GT5, but PD could have found a better solution than what they have now. That seems pretty obvious!!!

I will search again later, but for now I'm off to work.

 

[esh]

I wonder if the audio update in 2.03 is pushing things over the edge which is why we're starting to see more people experiencing these issues, not just wheel users.

 

[Nanaki]

A quick hello from France and my community GT5rs (for those here who remember us).

Since a few weeks, this problem is ruining all our competitions, it's unbelievable ! Our races have been suspended until the problem is fixed. We have done many tests like those you are making here. We cannot race safely above 9 or 10 drivers. I experienced myself a 10-seconds-per-lap loss on some races. This is ****ing ridiculous.

We cannot compete anymore, this means the death of the GT5 community if the problem is not solved quickly !

 

[Sail IC]

It's too early to say PD are not aware of it as the update cycle is always going to be 1-2 patches behind, meaning the screens we saw for 2.03 that were leaked and then official were dated Dec. 16th, before 2.02 even came out on Dec. 20th. Whether or not the update was complete at that point is another story but the content of the update probably had already been decided upon.


I would not be concerned that it isn't fixed in this update and possibly the next although the fact that a timeframe hasn't been mentioned for an update after 2.03 besides the generic monthly cycle Kaz talked about @ SEMA.

For planed enhancements i agree, but serious bug fixes might be on a shorter schedule. If the bug we are discussing here will be well spread in the gaming community, it will be very bad publicity for PD and perhaps even Sony as well who owns PD. This might put a fix on a different timeline.

"GT5 makes the PS3 platform to go on its knees" .... That would be a bad headline.

On your second post (don't know how to multi qoute), I agree. In my view, PD has done a superb job adding features to an already release product. But in doing so, something has made the game crippled online. Only guessing, one quick fix would be to remove the feature and go back to a previous feature, redo it and put it back on later. As an example, when the new slipstream was released, I think a more powerful dirty air was incorporated. Lets say for discussions sake this is the issue, just go back to the old version, engineer it smarter, and put it back on again in 2.04 or 2.05.

I would say, if its not fixed in 2.03, the gaming community should put som pressure on PD which at the same time will make Sony to put pressure on PD to fix it too.

 

[RUFin]

I've been looking for an over an hour, but I can't find the original post. The point was: PD gave preference in its netcode for the visual engine to take priority over the physics engine. So during on-line races, the game is constantly trying to make everything LOOK perfect. Meanwhile, the physics engine is trying to compensate for a lack of computing power, and when FFB and geography comes into play, the physics break down. The visual component has the power, but doesn't get the input it needs because the physics aren't sending input fast enough. So the result is the visuals give you whatever it gets, but that input is "staggered", resulting in frame rate issues. As this is happening, any driver input is momentarily muted by the conflict between physics and visuals, resulting in a "coasting" effect, not to mention a lack of steering input at critical times. The result is a lap that looks better but is much slower.

This would mean its constant,which it isnt.It can hit you after a lap or two.Happened to me in races against fast fins.Suddenly you just lost 5sec lead with slowdown and poor 'physics'??? Its MADE to slow you down by PD.

 

[esh]

For planed enhancements i agree, but serious bug fixes might be on a shorter schedule. If the bug we are discussing here will be well spread in the gaming community, it will be very bad publicity for PD and perhaps even Sony as well who owns PD. This might put a fix on a different timeline.

"GT5 makes the PS3 platform to go on its knees" .... That would be a bad headline.

On your second post (don't know how to multi qoute), I agree. In my view, PD has done a superb job adding features to an already release product. But in doing so, something has made the game crippled online. Only guessing, one quick fix would be to remove the feature and go back to a previous feature, redo it and put it back on later. As an example, when the new slipstream was released, I think a more powerful dirty air was incorporated. Lets say for discussions sake this is the issue, just go back to the old version, engineer it smarter, and put it back on again in 2.04 or 2.05.

I would say, if its not fixed in 2.03, the gaming community should put som pressure on PD which at the same time will make Sony to put pressure on PD to fix it too.

I can't speak to how other PS3 games are patched as I only play GT5, but I have experience with PC games, well 1 or 2 games in particular, and the windows between updates/patches/etc is far, far greater than 30 day cycles (years in some cases). Don't take this the wrong way as I suffer from this issue nightly in larger rooms but the game is far from unplayable. There are workarounds as have been discussed throughout this thread that mitigate the described issues online. If there was a bug crashing everyone's PS3 left and right or saved information was being lost then yes, that is a critical issue and would require immediate attention. This is more of an annoyance or an inconvenience and I would rather have them take their time to fully sort out the issue, streamline code and/or processes to ensure that it is fixed proper and we don't experience bug regression further down the road with future updates. And if that takes another 30-45 days from 2.03, so be it. Just my opinion....

 

[chuyler1]

There has been talk about the issues not happening in free run before the race. Can anyone verify that 16 cars sitting on the track in a mock-free-run-race won't cause the same lag issues? Do these network bugs occur in free-run and drag-race lobbies?

 

[Blind829]

There has been talk about the issues not happening in free run before the race. Can anyone verify that 16 cars sitting on the track in a mock-free-run-race won't cause the same lag issues? Do these network bugs occur in free-run and drag-race lobbies?

I've been in a lobby with 14 premium cars practicing online before a race with no issues, not 16 though. Once the race was started the bug was present though. Everyone could just run their races in free run with no draft or replays...

 

[Highlandor]

A quick hello from France and my community GT5rs (for those here who remember us).

Since a few weeks, this problem is ruining all our competitions, it's unbelievable ! Our races have been suspended until the problem is fixed. We have done many tests like those you are making here. We cannot race safely above 9 or 10 drivers. I experienced myself a 10-seconds-per-lap loss on some races. This is ****ing ridiculous.

We cannot compete anymore, this means the death of the GT5 community if the problem is not solved quickly !

Hello / Salut NANAKI

For those that don't know, GT5rs is a prominant French Gran Turismo website, they run (very well) a highly prestigious inter forum series called "Fusion" - sponsored by a major motor industry company.

For them to have to postpone their races is quite serious, as well as many of the "little guys" who are doing the same thing (organised racing), Polypony really need to sort this problem out ASAP.

As Nanaki says, it's killing online racing - connection issues over the past year has seriously damaged Polyphony's / GT5's reputation - this latest issue, this is the straw that broke the camel's back.

New Year - start of new season for many organisors / race series, oh, not any more....

Just read on the "Kaz tweets" thread - he's going to a motor show in Japan the weekend to have a discussion about the 'ring.

Nice - 2 things though:

1) we / everyone all know about the ring anyway, surely his time would be far better spent kicking some a** at Polyphony and fixing this quite serious problem.

2) Kaz is going to the motor show with one of the Subaru bigwigs (so it say in the thread) - AT LAST, now we know why the EVO's are so much slower than the Scoobies in GT5 - been wondering about that for the last 12 months.

Pffffft.....

 

[chukabilly]

This would mean its constant,which it isnt.It can hit you after a lap or two.Happened to me in races against fast fins.Suddenly you just lost 5sec lead with slowdown and poor 'physics'??? Its MADE to slow you down by PD.

I've thought about that for a few hours, and I don't agree. The relationship between the physics engine and the visual engine is not static, but rather constantly changing. It is in this constant state of flux, where PD decided to assign priority to the visual, rather than the physics, where these problems arise.
As we continue to hone in on the triggers for this problem, it's keeps looking to me that the balance between the two engines is so fragile, and there are so many variables involved in a simultaneous, global on-line event, that any number of factors could instantaneously cause a change in any race at any time.
I believe that the only solution that would give long term stability to our racing lounges will involve isolating the rooms from the open environment, or cloud, if you prefer. In other words, we need to be able to close a room from any outside influences, such as someone checking in who just wants to watch the race.
I don't think PD deliberately programmed a random "slow you down now" function into the game. I do think PD tried to do too much on a platform that is not capable of handling such a large amount of variables, and I would not be surprised if we discover, at a later date, that part of our problem is PD and Sony tying to reconcile their new security apparatus with a game that was made before the new security protocol was put in place.
Jake, I tried to do better with the paragraph thing!!

edit: Besides people jumping into a room, very hard crashes seem to instigate the problem too. Again, the physics engine failing. Even if you weren't the one that crashed, the whole room can be affected, or only random individuals.

 

[dsgerbc]

^While the issue isn't new, it wasn't too bad until recently. So I doubt it's just the priority of visual over physics. Something clearly made it worse, so it could/should be feasible to undo/fix.

 

[BrandonW77]

^While the issue isn't new, it wasn't too bad until recently. So I doubt it's just the priority of visual over physics. Something clearly made it worse, so it could/should be feasible to undo/fix.

Think about it though, they are constantly improving the graphics which could be using more and more resources. As the graphics prioritize more resources the physics get worse, which is what we've seen.

 

[MONSTAR-1]

Someone should just send a tweet to KAz and ask him why we are having these problems. If he dosnt answer just keep asking, everyday until he either blocks you or answers you.

 

[Speedy6543]

Someone should just send a tweet to KAz and ask him why we are having these problems. If he dosnt answer just keep asking, everyday until he either blocks you or answers you.

This has been done already but I'm not going to spam him...
https://www.gtplanet.net/forum/showthread.php?p=6440621#post6440621
https://www.gtplanet.net/forum/showthread.php?p=6479212#post6479212

 

[CJSpencer77]

Its not even the fact that the game is now this broken, its the wall of silence anything negative about the game is met with by PD, if people only knew PD were on the case and fixing it they probably wouldnt care if it took a month, but a month of not knowing if all the work you put in building a league up around GT5 with a good number of core players invested in said league is going to be a waste of time or not is killer

 

[Blaze_409]

There has been talk about the issues not happening in free run before the race. Can anyone verify that 16 cars sitting on the track in a mock-free-run-race won't cause the same lag issues? Do these network bugs occur in free-run and drag-race lobbies?

The issues dont happen in free run. In the GTP ALMS event at Cot'e de Azur that was hosted last friday we had a full room(16) and no glitches bugs or anything in the free run mode. This is where everyone set the qualy times. Come race time however we suffered from catastrophic glitching with the laptimes racing from 5 seconds -15 seconds slower for every person. People driving the same car where affect differently like me who was 15 seconds off pace while person in same car as me was only 7 seconds off pace...


You know whats interesting. I think there is a ton of truth in the rumours about the PS3 being overloaded or simply not having enough power for GT5. I used to have a very slow crappy desktop and i would play this game much like the "snake" (see TRON movie) game where you run around making the lines for others to hit. Because the computer was so slow the game was incredibly easy even on the hardest mode. However i decide to play it on a much more powerful computer and suddenly everything was running smooth but the game was sooo much harder since everything was moving faster.. If GT5 is overloading the PS3 that could explain the Time dilation that is occuring....PS3 simply cannot play GT5 in real time, and when you have a wheel of course(the wheel overloads the PS3 even more as the PS3 has to output the FFb data which might be why so many wheel users are affected.) The result is the Ps3 is running producing images slower than real time?


What i dont get however is my PS3 does not overheat and never has. I even use it in a fairly warm room. Temperatures on my PS3 Slim give no indication that the PS3 is at max processing capacity

 

[K1LLD0z3R]

Brain storming... sorry if it has come up.

Can we blame prologue servers? I remember this, it's like deja vu.

 

[OwensRacing]

Brain storming... sorry if it has come up.

Can we blame prologue servers? I remember this, it's like deja vu.

So it was an issue in Prologue? If so and they are using the same servers. The server being implemented slowly and then fully recently. It would make sense.

Hmmm hope this solved and I get faster. LoL

 

[cunny17]

Its not even the fact that the game is now this broken, its the wall of silence anything negative about the game is met with by PD, if people only knew PD were on the case and fixing it they probably wouldnt care if it took a month, but a month of not knowing if all the work you put in building a league up around GT5 with a good number of core players invested in said league is going to be a waste of time or not is killer

+1 👍

 

[Sail IC]

A quick hello from France and my community GT5rs (for those here who remember us).

Since a few weeks, this problem is ruining all our competitions, it's unbelievable ! Our races have been suspended until the problem is fixed. We have done many tests like those you are making here. We cannot race safely above 9 or 10 drivers. I experienced myself a 10-seconds-per-lap loss on some races. This is ****ing ridiculous.

We cannot compete anymore, this means the death of the GT5 community if the problem is not solved quickly !

Good summary of the result of this issue. If we look at the "online competition" part of GTPlanet forums, many of the most committed racing leagues have no suspended their racing. Its unplayable for them (us).

GTPlanet is a great meeting ground for the most committed GT5 players and this issue is affecting these players in a devastating way. That's why I cannot understand the organizers of GTPlanet are so hesitant to take a more active role here, like posting it on the front page.

 

[mrkevans]

Maybe next update will address this issue, fingers crossed

 

[Famine]

GTPlanet is a great meeting ground for the most committed GT5 players and this issue is affecting these players in a devastating way. That's why I cannot understand the organizers of GTPlanet are so hesitant to take a more active role here, like posting it on the front page.

We're not.

What I don't get is why both of you have assumed that it isn't going to be, nor why you assume that personally insulting me or my work will encourage me to do what you want. I also don't understand why you, Rodrigo, are so adamant about featuring this bug on a blog you consider so irrelevant.

Before continuing, though, let's take a look at a few things that you're ignoring, either by convenience or - hopefully, for your character's sake - ignorance.

• The overwhelming majority of bugs, ranging in severity, have been discussed exclusively in the forums, and an overwhelming number of them have been resolved without any mention whatsoever on the News blog. A blog post is clearly not a pre-requisite to resolution of an issue; a fact that's directly supported by what I've been told privately by an SCEA/PD employee.
• The blog collects and reports news about a wildly successful video game franchise. Unsurprisingly, the majority of that news is going to be 'positive'. That said, here's just a sampling of 'negative' stories I've posted about, which I strongly doubt Sony, Polyphony Digital, or Kazunori Yamauchi were particularly excited or proud of:
  • This Gran Turismo 5 Bug Can Kill Your Game
  • "Break Out" of GT5's Nurburgring 24h Course
  • Gran Turismo 5 v1.08 Update Fixes "Money Glitch"
  • Cheaters Ruin Official GT5 Prologue Competition
  • GT Academy USA Round 2 Marred by Corner Cutting
• Despite your outrage at not seeing this on the blog yet, not one of you has sent me a private message or an email about it. I saw the thread when it was originally posted, but dismissed it as one of the usual mis-informed complaints about lag. Only yesterday, did one person attempt to update me on the growing severity of this bug.

The forum has and always will be several steps ahead of any blog, because thousands of people constantly contribute to it, every hour of every day. There is no way I can possibly 'compete' with that, and I don't try to. If you're closely monitoring these forums all day, I'll be the first to tell you that you don't need to read the blog. For the rest of the site's more casual visitors, I hope they find it useful.

I created GTPlanet's blog to analyze, digest, and present the most notable information from the forums in a simple, chronological way - something I could be doing right now, were I not spending time here defending my character and my reputation. If you have any constructive feedback about what you'd like to see more of on the blog - other than complaints and malcontent - I would appreciate if you shared it with me.

When (not if) this complex and confusing bug is posted on the blog, it will be presented in a concise, precise, organized, and easy-to-understand format, backed by as much hard data and evidence as is available. Meanwhile, you can all help by continuing to gather and share as much relevant data as possible.

I appreciate your apology, Rodrigo. Now that these concerns about the blog are out of the way, hopefully this thread can return to sharing information about the problem. I'll be referencing useful posts in this thread, so please, everyone, share what data you have.

We are trying to determine specifically what circumstances cause the issue and how to avoid it. Once we've properly analysed the issue, we'll be posting about it. The help the GTP community can provide is not asking why we're "hesitant" to post it on the blog (we aren't) or when we're going to (when we have sufficient information) but by providing more and more relevant data.

 

[cicua]

In the WSGTC pre-seasonal race yesterday, we boldly tried a full room. This just ended up confirming that you can only have 12 drivers.

We also had 1-3 spectators but they seem to have little or no effect.

Breakdown of stopwatch check:

Lap 1 - 2 (15,14 drivers): 4s lost per lap
Lap 3 - 18 (13 drivers): Almost 1s lost per lap
Lap 19 - 43 (12 drivers): 1s in total lost over 25 laps

This matches the post-race reports very well, several people reporting screen tearing / frame rate issues until we got down to 12.

I recorded some of the race on video, including a realtime clock in the picture, to compare the running game clock against real time. They match exactly.

This means that my first lap was really 2:41, although it only lasts 2:37 when replaying. And the whole race really did last 1:13:34, although 21 secs are missing in the replay.

I recorded the network traffic of the entire race, 260MB, but I do not feel like going analyzing it because I do not expect it has anything that could help us. The only trigger for the problem seems to be the number of players, confirmed by everyone by now, and there is nothing we can do about that.

It is up to PD to fix this, now. The whole online community is going into suspend mode.

 

[Niop]

I ran a 40 lap NASCAR race at Suzuka East. We had 13 at the start, but one quit after one lap. The other 12 stayed for the remainder of the race. I lost at least 1.5 seconds per lap, and a total of 80 seconds over the course of the race.

In this race the bug only affected a couple of us. I use a G27 wheel and a USB headset.

 

[CSLACR]

In the WSGTC pre-seasonal race yesterday, we boldly tried a full room. This just ended up confirming that you can only have 12 drivers.

We also had 1-3 spectators but they seem to have little or no effect.

Breakdown of stopwatch check:

Lap 1 - 2 (15,14 drivers): 4s lost per lap
Lap 3 - 18 (13 drivers): Almost 1s lost per lap
Lap 19 - 43 (12 drivers): 1s in total lost over 25 laps

This matches the post-race reports very well, several people reporting screen tearing / frame rate issues until we got down to 12.

I recorded some of the race on video, including a realtime clock in the picture, to compare the running game clock against real time. They match exactly.

This means that my first lap was really 2:41, although it only lasts 2:37 when replaying. And the whole race really did last 1:13:34, although 21 secs are missing in the replay.

I recorded the network traffic of the entire race, 260MB, but I do not feel like going analyzing it because I do not expect it has anything that could help us. The only trigger for the problem seems to be the number of players, confirmed by everyone by now, and there is nothing we can do about that.

It is up to PD to fix this, now. The whole online community is going into suspend mode.

Good report buddy, thanks.👍

 

[jon-]

In the WSGTC pre-seasonal race yesterday, we boldly tried a full room. This just ended up confirming that you can only have 12 drivers.

We also had 1-3 spectators but they seem to have little or no effect.

Breakdown of stopwatch check:

Lap 1 - 2 (15,14 drivers): 4s lost per lap
Lap 3 - 18 (13 drivers): Almost 1s lost per lap
Lap 19 - 43 (12 drivers): 1s in total lost over 25 laps

This matches the post-race reports very well, several people reporting screen tearing / frame rate issues until we got down to 12.

I recorded some of the race on video, including a realtime clock in the picture, to compare the running game clock against real time. They match exactly.

This means that my first lap was really 2:41, although it only lasts 2:37 when replaying. And the whole race really did last 1:13:34, although 21 secs are missing in the replay.

I recorded the network traffic of the entire race, 260MB, but I do not feel like going analyzing it because I do not expect it has anything that could help us. The only trigger for the problem seems to be the number of players, confirmed by everyone by now, and there is nothing we can do about that.

It is up to PD to fix this, now. The whole online community is going into suspend mode.

Good work but you've confused me (not hard )

So

During race - real time vs game time = the same
During replay - real time vs replay time = replay time slower

So if a lap in the race took 2 minutes, and it took 2 minutes in real life, the replay would show a 2 minute lap actually taking 2.04 real minutes?

 

[MGR]

Good work but you've confused me (not hard )

So

During race - real time vs game time = the same
During replay - real time vs replay time = replay time slower

So if a lap in the race took 2 minutes, and it took 2 minutes in real life, the replay would show a 2 minute lap actually taking 2.04 real minutes?

Not quite.

If you time your laps using a stopwatch during the race both the stopwatch and game clock match perfectly whether you're suffering from the bug or not.

However only 'bugged' laps will exhibit a time difference to the game clock when watching the recorded replay - ie the stopwatch time will be faster.

There is a very strong correlation between stopwatch replay time and your game clock lap time in free run mode.

For example, if your able to punch out very consistent laps in free-run mode of 2:00 but in the race you're only able to do 2:04 - you can bet that the replay 'stopwatch' time will be 2:00.