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

[dudeondacouch]

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.

I agree, but eliminating variables is job one when trying to isolate a problem.

I'm still of the mind that the difference between online and offline physics and this "bug" are not only related, but one in the same. There is some factor at work that causes it to be different, and while it originally only seemed to be exacerbated/amplified by certain cars, certain drivetrains, or certain tracks, it is now being amplified by the number of room participants and/or their peripherals.

Does the system handle replays better if there are fewer participants? If so, this could point to a hardware/software issue as opposed to a networking issue.

 

[NEPALII]

Amazing video. 👍 Felt sorry for the left bottom driver. Looks like the same issue I had experienced at WSGTC Preseason race. Well, now with 2.03 out, it's time to test the bug again. You should send that video to Kaz's twitter or something.

 

[Carlos Fandango]

Excellent update to the OP MGR - great summary 👍

 

[avens]

I made this video of our 'bugged' WRS-Online Event a while ago.
The race took place just after the 2.03 patch where nobody knew of the bug yet...

that's front page material

 

[cicua]

I agree, but eliminating variables is job one when trying to isolate a problem.

Spoken like a true engineer 👍

- and that is exactly the process we are in.

I'm still of the mind that the difference between online and offline physics and this "bug" are not only related, but one in the same. There is some factor at work that causes it to be different, and while it originally only seemed to be exacerbated/amplified by certain cars, certain drivetrains, or certain tracks, it is now being amplified by the number of room participants and/or their peripherals.

Does the system handle replays better if there are fewer participants? If so, this could point to a hardware/software issue as opposed to a networking issue.

I see where you are going, but can it be used to prove anything?

If a replay with fewer drivers shows less rpm fluctuation, it might hint at a local software/hardware issue. But it could be so many other factors unfortunately, most of all lag. And the game itself could also cheat and send engine rpm info often when there are few drivers, and cut down on it if there are many, we do not know.

Besides that, I totally share the notion that online/offline physics is related to this bug.

 

[NEPALII]

I don't believe it can be hardware issues as other games don't suffer from this issue. I play BF3, for example, and I don't have any of the issues online that I have on GT5. I'm assuming it has to be software related. Now, what is causing all of this is still a mistery to all of us.

 

[cicua]

I don't believe it can be hardware issues as other games don't suffer from this issue. I play BF3, for example, and I don't have any of the issues online that I have on GT5. I'm assuming it has to be software related. Now, what is causing all of this is still a mistery to all of us.

The hardware issue I am thinking of is ... it is too slow

 

[88c-v Racer]

Could this bug be related to this issue I had the past two days?
I'll explain.

I have a setup for my 908 HDI on High Speed Ring that allows me to run nearly flat out on the high line through the two banked 180 degree corners with fairly good stability. On Saturday at the High Speed Ring I ran the fastest lap in the Free-Run, got out in front at the start of the race and cruised to a victory. On Sunday (same PP, same direction) I again had the fastest lap in the Free-Run (an almost identical time), and in the race I got in the lead and got through the banked turns just fine on lap 1. I was leading by about 2 seconds at the line to start lap 2, went off into the first banked turn like usual and spun. From that point on the car was almost undrivable, it could not go around the high side without spinning. I finished dead last. Same story for the next race, which was the same track and everything, spun out and finished last. (FYI: no boost, no damage, no tire wear, and track grip wasn't set to Real).

I've also noticed in a group of fairly even cars, without any mistake, a car will suddenly get freight-trained by the whole pack. Is this also related to the bug?

 

[chukabilly]

Congrats MGR!

Thank you to also finding time to make so valuable a contribution of information!

 

[cicua]

Here is another weird replay bug I just found. Watch the tyre wear indicators as I come out of pit:



They behaved normally in the race (got it on video), they only do this in the replay!

I am not sure if this has any connection to the grip-loss bug we are discussing, but I found it interesting. It happened after both my pit stops and lasted for more than a lap.

 

[Blaze_409]

Here is another weird replay bug I just found. Watch the tyre wear indicators as I come out of pit:



They behaved normally in the race (got it on video), they only do this in the replay!

I am not sure if this has any connection to the grip-loss bug we are discussing, but I found it interesting. It happened after both my pit stops and lasted for more than a lap.

This has been happening since way back when. Prior to Spec 2.02 i would watch replays and the tyre wear indicator would go absolutely haywire!

Also there where cases where the replay showed that a certain car had damage when the person themselves reported that they had no damage and went on to win the race!

 

[cicua]

This has been happening since way back when. Prior to Spec 2.02 i would watch replays and the tyre wear indicator would go absolutely haywire!

Have you seen it on your own car, or only on others'? This is my car in my own replay, no lag can explain this!

I have never noticed it, sorry if posting old hat info.

 

[Speedy6543]

Also there where cases where the replay showed that a certain car had damage when the person themselves reported that they had no damage and went on to win the race!

I noticed that too and I think it was due to having visible damage from the pre-race practice which was taken into the race.

 

[ajns73]

Here is another weird replay bug I just found. Watch the tyre wear indicators as I come out of pit:

[video snipped]

They behaved normally in the race (got it on video), they only do this in the replay!

I am not sure if this has any connection to the grip-loss bug we are discussing, but I found it interesting. It happened after both my pit stops and lasted for more than a lap.

Looks remarkably like the game chose to make them load indicators after the pit instead of wear indicators. That's the kind of animation I see when I am setting a car up in practice and using load indicators to make suspension adjustments.

 

[jon-]

Looking at the release notes I'm guessing when we test with 2.03 nothing will have changed...

 

[RodrigoDLL]

Here is another weird replay bug I just found. Watch the tyre wear indicators as I come out of pit:

<a href="http://www.youtube.com/watch?v=oLlfj2aSDaE">YouTube Link</a>

They behaved normally in the race (got it on video), they only do this in the replay!

I am not sure if this has any connection to the grip-loss bug we are discussing, but I found it interesting. It happened after both my pit stops and lasted for more than a lap.

Something tire related happened in our Gt500 race as well. The race director watched the replay and disqualified one of the drivers for using a second set of softs during his pitstop (regs only allow one set). The guy did not record his own replay, but another driver had, and on this second replay file the tire change was correct.

 

[chuyler1]

Also there where cases where the replay showed that a certain car had damage when the person themselves reported that they had no damage and went on to win the race!

This is usually caused by a typical lag spike on your side of the game. Even if you were no where near the driver a spike might cause his car to hit a wall on your screen. The replay saves it as damage.

Also note that damage will not be applied if you fast forward past an actual crash. If you don't watch the crash in realtime the damage doesn't display.

The replay system is horrible and often has little to do with the actual race unless cars are within a visible distance to you during the actual race. I've been accused of dirty passes by drivers that weren't there to witness it...and had to have the driver I passed save my reputation by explaining there was no contact.

In the end, I could care less if the replay is bad...just fix the actual gameplay first.

 

[Highlandor]

The video of the tyre wear indicators going "crazy" is very interesting, when myself and other in my lobby were affected by this "laptime bug", alot of the tiime, the front end just had NO grip. Cars felt like they had comfort hards on the front instead of sports softs or racing hards.

But, addmittidly, other times when this "laptime bug" has hit, the cars felt normal, normal braking points can be used and normal apex's easily reached, but the laptime is just different.

It's because there are these inconsistencies I feel this problem is alot deeper rooted in the game somewhere. It's not just the "bug" itself, but the different ways it affects different people - which in terms of solving it, is going to make things alot harder.

Poyphony should just cut their losses IMO - go backwards (reverse) create a version of the game that is BUG / GLITCH etc free, but take out as much unwanted faetures as possible, and try to rebuild the game more simply, more stable online and more consistent in terms of connection.

Like when something happens to your PC, you have a restore date you can always go back to. But make it so everyone's garage / car / ticket / paints etc etc data is transferable, but the game just runs smoothly and consistently once and for all.

Surely, by dropping unwanted features would create a more basic version of the game, easier to map and track issues and problems and sort them out once and for all (these are generally what most regular GT5 guys I know never use or want):

Museum cards
Horns
Some paints (500-1000 would suffice)
"Duplicate" cars i.e '96, '97, '98, '99 version of same model
Pretty "intro" scenes for games and challenges
Music
GT5 "mail"
Race suits and helmets
Tyre load indicator (maybe)

If loosing all of them meant the game was more simple and easier to "run" on the PS3 / PSN, resulting in a stable connection for online racing, myself and most of the guys I know would be so much happier.

All we want to do is hook up and race, clean, fair and connection free as reliably and consistently as possible - how difficult is that..??

I've been running race series for a year now, it's a complete nightmare when this should have been (video) game "heaven".

 

[shirakawaa]

[...]Surely, by dropping unwanted features would create a more basic version of the game, easier to map and track issues and problems and sort them out once and for all (these are generally what most regular GT5 guys I know never use or want):

Quit pushing for the removal of features that you don't like with the false pretext that without them the game would be easier to debug. 👎

As if things like dropping the number of paints or cars would make PD's netcode better...

 

[BrandonW77]

^ I was also trying to figure out how race suits and horns effected online game play. I was also not aware that having a '96, '97 and '98 model of the same car in my garage would cause this online bug to effect me.

 

[dudeondacouch]

Yes, please leave the complaining about certain features out of this thread. (Although, for the record, I agree that there's an awful lot of content that boggles my mind.)

 

[RUFin]

The video of the tyre wear indicators going "crazy" is very interesting, when myself and other in my lobby were affected by this "laptime bug", alot of the tiime, the front end just had NO grip. Cars felt like they had comfort hards on the front instead of sports softs or racing ha
But, addmittidly, other times when this "laptime bug" has hit, the cars felt normal, normal braking points can be used and normal apex's easily reached, but the laptime is just different.

It's because there are these inconsistencies I feel this problem is alot deeper rooted in the game somewhere. It's not just the "bug" itself, but the different ways it affects different people - which in terms of solving it, is going to make things alot harder.

All we want to do is hook up and race, clean, fair and connection free as reliably and consistently as possible - how difficult is that..??

Not difficult at all,BUT because GT5 is PDs first online "experiment",they need to have data on different variaisons.Its maybe so that they have picked reasonably good drivers already based on Prologues race results and started to add these slow downs and loss of grip things already in Prologue.They keep testing in GT5 too.I would say this is a good example of a BETA test!Things found here are used in next gen consol and that final GT.Then you find out how expensive a 'real driving' can get(wheel 500e,consol,GT with minimum content,DLCs $$$).Thats where this is amed.PD skam,discaid as a 'game'.Could be that what you see as youre settings(tires,pp,what ever)might be just that.WHAT YOU SEE.They really are something else. -I know,movie English,dont care

 

[ErikAndre71]

Well I haven't had the time to test it in race mode just yet, but when entering my lobby I had to set the lobby settings all over again. And when I got into my lobby they brought back the old leader board, and spectate mode is back to "normal". Could be a fix for all I know. But I can't really test race mode until Saturday so I'll be watching here if someone test it.

 

[cgr-10]

I've got an IndyCar race Friday so I hope this has been fixed by PD.

Anybody had a full lobby race since 2.03?

 

[Matty]

I've got an IndyCar race Friday so I hope this has been fixed by PD.

Anybody had a full lobby race since 2.03?

Read in the news section that it wasn't meant to be fixed by 2.03 👎

 

[cgr-10]

Read in the news section that it wasn't meant to be fixed by 2.03 👎

As if, pfft.

We'll just give it a go.

 

[Matty]

As if, pfft.

We'll just give it a go.

No harm in trying, I'm just not getting my hopes up.

 

[ErikAndre71]

Read in the news section that it wasn't meant to be fixed by 2.03 👎

I don't think anybody think this is a fix, just a observation, and the in the news section no one said it will not be fixed in 2.03, just that it was not likely. What PD decide, couldn't be assumed by anyone outside PD.

 

[WRP001]

I wouldn't expect some big announcement from PD that this bug was fixed (or will be). More likely they'll patch the code or make adjustments to alleviate the issue and just release with other public changes.

I doubt they are going to publicly acknowledge the bug was an issue with their code and publicizing a fix is going to be doing just that. Let's see if I'm wrong. Maybe they'll include something in a future change log like "Improved online racing performance under certain network conditions". That would be a good way for them to let us know they fixed it without really admitting there was an issue.

 

[RodrigoDLL]

I wouldn't expect some big announcement from PD that this bug was fixed (or will be). More likely they'll patch the code or make adjustments to alleviate the issue and just release with other public changes.

I doubt they are going to publicly acknowledge the bug was an issue with their code and publicizing a fix is going to be doing just that. Let's see if I'm wrong. Maybe they'll include something in a future change log like "Improved online racing performance under certain network conditions". That would be a good way for them to let us know they fixed it without really admitting there was an issue.

I also dont expect anything like that from them, but I have to admit that their ways are not what I would expect from a japanese company.

Japanese people are very closed, but are also very much in favour of publicly admiting their mistakes as a way of punishment. I can remember japanese authorities fromthe nuclear plants that broke down after the earthquake going public just to say sorry for their response to the disaster having not been up to public expectations. That was a very japanese way to do it.

I dont get why PD isnt more like that. It would really help their image to be more open to admit when they are wrong, because they are very sucessfull in general and people would support them when some small issues (compared to the big sucess of the game) arise.