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

[Ealirendur]

It was a reasonably close 5 lap race, so I have no reason to suspect that the 'phantom lag' was present in that race. But why did my replay 'smooth out' with repeated viewings?

I can deal with the whole online/offline being different. I can deal with that fact that because of that, the online and offline laptimes will never match. I don't care.

What really riles me is that in online racing, some of us are going like our cars have 3 wheels while the others have 4. There is no consistency amongst cars/timing online. How can we race like this?

Well, another thing to test. I'll have to try exporting a saved replay, watching it/running it multiple times, then comparing it to the other one after importing it to another account. Trick is, of course, you would have to export it before viewing if the file is being modified for interpolations when viewed.

We can only make the best of it. Once we get to the point where we can say more definitively "Doctor, Doctor, it hurts when I do this!" "Don't do that then". (Until the next patch..... )

Being a bit careful with testing methodology will hopefully eliminate the "Doctor, it hurts when I press here, here, and here" "Your finger is broken." class of problem - for example, we might have to check those stopwatch/timing issues again over several runs on the replays. Well spotted, if it's true.

 

[Jubby]

:
We can only make the best of it. Once we get to the point where we can say more definitively "Doctor, Doctor, it hurts when I do this!" "Don't do that then". (Until the next patch..... )

That's an awesome outlook, seriously!
It really makes me laugh thinking of an issue with my car at a certain speed (35-50mph)... my buddy said, "Don't go in that speed range then!".

It seems as though this issue may be hurting online the worst and unevenly. Since we know they'll be tinkering with online in the next update, hopefully they'll fix this too.

 

[CSLACR]

http://www.online-stopwatch.com/full-screen-stopwatch/
If someone can verify this timer is NOT bonkers mad, I'm roughly 25 seconds ahead with this timer then the in-game timer, only 39 minutes into a 4 hour race at the Nurb.
OMG, imagine the luck, that timer is also faster then my phone, now I have to start timing again.

 

[MTLclllutch]

does anyone here play battlefield 3 for ps3. Its also full of strange errors and bugs that the devs cant seem to work out. VOIP issues, connection drops, input lag[xbox version has none]. Im starting to beleive this is a hardware problem.

 

[mrkevans]

does anyone here play battlefield 3 for ps3. Its also full of strange errors and bugs that the devs cant seem to work out. VOIP issues, connection drops, input lag[xbox version has none]. Im starting to beleive this is a hardware problem.

I prefer bad company 2 than bf3 but both have bugs, it's a programing issue, I think with big games its more apparent, gta4 is another one.

 

[windmill]

Well MGR, you have just raised my eyebrow with mention of the Replay and I will go that far and say how Replay *could* be the main reason for this glitch.

It fits perfectly into picture, because from what we may presume, Replay file have to be *recorded* somewhere, and that somewhere is probably on the console/player with the highest bandwidth.

Because of the P2P nature of the connection, poor Host probably hqve to take all the distribution over himself AND record a replay. And since there is +12 cars of Premium quality - notice difference in damage modelling as a most important difference factor between P&S cars for real-time data, everytbing else could go prebaked - it could easily be the main reason for the choke.

Also, I have to say how issue was probably less noticeable once when "Visual Damage OFF" was introduced, but I can't be sure on this one.

Replay saving could easily be the main reason why the host is bottlenecked by bandwidth. And it could also explain why I can play my replays perfectly, while others have a slowdowns and twitches with the same file that was recored after the race under the same conditions. Replay could very easy be the key to this issue.

@ amar
I have a friend with a g27 wheel who also expeirences a heavy slamming from the wheel in corners without touching kerbs have you heard that before ?

 

[Ealirendur]

does anyone here play battlefield 3 for ps3. Its also full of strange errors and bugs that the devs cant seem to work out. VOIP issues, connection drops, input lag[xbox version has none]. Im starting to beleive this is a hardware problem.

That's an awesome outlook, seriously!
It really makes me laugh thinking of an issue with my car at a certain speed (35-50mph)... my buddy said, "Don't go in that speed range then!".

It seems as though this issue may be hurting online the worst and unevenly. Since we know they'll be tinkering with online in the next update, hopefully they'll fix this too.

I should also mention again, probably redundantly:

Nothing will save even well written netcode from bad links, dropouts, and the tyranny of distance - past a certain point, it can only fail, gracefully or otherwise. I've written [post=6270975]before[/post], and others have explained with less verbiage [post=6349386]elsewhere[/post], for example.

It could also be something not so wonderful in the PS3's network software (wireless, anyone?). Certain APIs/libs won't actually be PDs, particularly relating to PSN protocols, but likely a great deal more. The race quality setting assumptions might have to do with expected minimum timings or (upload) bandwidth, and go south quickly if they're not met. None of this means that said code can't be better, but try finding me some successful FPS players with Satellite connections.

The problem in the 3D3 races could easily have just have been ..... me. Putting an Australian in a 15-player room full of EU and NA continent players might just be a Bad Idea now matter which way you slice it. But some latency issues are not so easy to guess.

Discovering the limits, evidence of flaws and mitigating factors is useful, however, since PD will never reeaaally consider putting a "You Sure? I can only barely make 14 players on two continents work with High Quality now - DON'T HIT OK, you inconsiderate aussie git!" button on lobbies.

 

[SuperSic]

I experienced this for the first time last night.

Open lobby,all was good for about 2 hours of racing with a steady core of 6-7 drivers and 2-3 popping in and out. By the time we became 13 exactly we started to have some regular lag from some cars during Free run. We started the race and on lap 1 nearly everyone had the issue and it was pretty clear who was affected and who not, but no regular lag as far as i could see(no erratically moving cars),only huge FPS drop. People started backing out because of it and as they were leaving things went gradually back to normal with 3rd and 4th laps being completely normal in terms of laptimes.

I can't explain what happens exactly, but i can't call it loss of grip as some are describing it here. I believe that when FPS drop massively, we become unable to nail braking points and apexes as we have very limited visual information for any minimum range of time. So it's not the physics that change, it is just the time flow that is different for each driver. We just interpret in like loss of grip because of our racing instincts.

 

[CJSpencer77]

myself and 22 other friends run a league, ALL of which have had this for the last few months, and a lot of us are at our wits ends, we dont have any ropey connections, my connection is a 50mg fibre optic line, one of the guys has 100mg fibre line, we all get this problem, for an example our next race is a inter league race with another league, the cars are the premium Mazda Roadster RS, 325bhp on sports softs at Grand Valley, whilst practicing ive been hitting 2:01.5's give or take a few tenths, when we raced last night i couldnt get out of 2:03.0 due to the lag, i get some frame skip but it feels like the physics engine skips with the frames which makes no sense at all, basically the overall control input of the car disapeers, at the end of the day its only a 325bhp Mx5 and when this lag happens its nigh on impossible to hit apexes, how our 2 divisions in our league are going to run 12 player races is beyond me at the moment seeing as thier was only 9 of us last night

 

[Ealirendur]

I can't explain what happens exactly, but i can't call it loss of grip as some are describing it here. I believe that when FPS drop massively, we become unable to nail braking points and apexes as we have very limited visual information for any minimum range of time. So it's not the physics that change, it is just the time flow that is different for each driver. We just interpret in like loss of grip because of our racing instincts.

That's what I'd call perceptual lag. Along with framerate drop might well come false or slow FFB information, which would really mess with your instincts.

 

[nasanu]

I can't explain what happens exactly, but i can't call it loss of grip as some are describing it here. I believe that when FPS drop massively, we become unable to nail braking points and apexes as we have very limited visual information for any minimum range of time. So it's not the physics that change, it is just the time flow that is different for each driver. We just interpret in like loss of grip because of our racing instincts.

This makes some sense however.. speedy? Can't remember the name (guy who posted the replay) said everything was normal for their race, no significant loss of frame rate. Both of your experiences can not be representative of the bug.

Working further on this post.

Good news!
As I do a livestream every week of the race, which is viewable here, you guys can all see the difference between my quali grip and race grip.
Also, take note that the GT5 in race timer, in this livestream video, is a 100% accurate. Which wasn't the case when I reviewed the replay of this race........
You can all take out your stopwatch and FFWD to 21'00.0min and see my second lap which I already timed in the replay and being 14 sec off then.

Sorry I didnt get around to looking at this earlier, but.. What is that video of? It is not the race you uploaded a replay of. The race winner is not even in it.

 

[chuyler1]

It's been a while since I have experienced this but I can also confirm it existed pre-spec-2.0.

Slow motion is the best way to describe it. One time it happened at Madrid with a full grid. I had the pole and by turn 1 I dropped several positions. Going around that first hairpin I ran the inside line at close to max grip. Other drivers went around me like I was a caution cone. By the time I reached the next braking point I was in last place.

Other times it has happened haven't been as obvious but usually I can race clean without noticing if I am alone on the track. Only when I look up to check lap times do I realize I am 3-5 seconds off pace from qualifying.

 

[MGR]

I can't explain what happens exactly, but i can't call it loss of grip as some are describing it here. I believe that when FPS drop massively, we become unable to nail braking points and apexes as we have very limited visual information for any minimum range of time. So it's not the physics that change, it is just the time flow that is different for each driver. We just interpret in like loss of grip because of our racing instincts.

Assuming you experienced the same bug being discussed in this thread then I completely disagree with your synopsis.

The car absolutely suffers longer braking distances, stunted acceleration and most notably lack of corner grip. The car simply will not turn into the corner at the same speeds achieved in free-run mode. You can of course still run clean lines but you must brake earlier and wash off more speed before turning into the corner.

Those not effected by the bug can quite often out-brake and pass you around the outside of a corner, such is the difference in grip.

 

[SuperSic]

Assuming you experienced the same bug being discussed in this thread then I completely disagree with your synopsis.

The car absolutely suffers longer braking distances, stunted acceleration and most notably lack of corner grip. The car simply will not turn into the corner at the same speeds achieved in free-run mode. You can of course still run clean lines but you must brake earlier and wash off more speed before turning into the corner.

Those not effected by the bug can quite often out-brake and pass you around the outside of a corner, such is the difference in grip.

As my experience with this 'bug' is pretty limited i can't be too confident on my asumptions so i'll have to come accross it again to be sure. I might be wrong, but this is my take at this point.

The fact that someone not affected can outbrake you and pass you from the outside doesn't conflict with my original theory of different time flow for each driver. If i were to summarize my theory in one sentence i would say:

''During every frame you miss, because of the FPS drop, you are unable to apply any kind of input. This means no throttle,brake or steering.''

Without further analysis,that explains how acceleration is slower,braking distances are longer,and handling is different. But it is not loss of grip and the physics are not affected at all.

PS. This is just a theory based on my brief experience.

 

[nasanu]

''During every frame you miss, because of the FPS drop, you are unable to apply any kind of input. This means no throttle,brake or steering.''

But that is not true. You can apply input, you just can not see the effect of your input till the next frame. The physics engine runs at something like 1000 times a second, if you could only input while each frame was being processed there you would no reason at all for the physics engine to run any faster than the frame rate.

Remember that most racing games run at 30fps, half the frame rate of GT5 and they run fine physics wise. Have a think about PC gaming also. There is no way to tell what the frame rate will be in most PC games, the physics engine has to run independently.

 

[esh]

If it affects only FFB wheels, how are people connecting these devices to the PS3? Directly or via a HUB? Is the hub powered by an AC adapter? I have experienced this pre and post 2.0 and have my wheel connected via AC powered USB hub but have also had it directly plugged into the PS3.

I think it's an issue with mesh topology, latency between clients, and bus being too slow from wheel to PS3, on top of clients having crap internet in the first place.

 

[chuyler1]

I really don't think a racing wheel has anything to do with it. I've gotten the same reports in my full lobbies from DS3 drivers.

The only reason you are making the connection to FFB is because most of the highly skilled online racers use some form of FFB and it is more noticeable.

The problem is network related. If just one driver has significant lag, he can wreak havoc with seemingly random people in the lobby, even the host.

I have never experienced this in a small lobby and I have never experienced it when there wasn't someone from another continent in my lobby.

I would venture to guess the clocks are synchronized with some sort of round-Robbin algorithm and if you are at the end of a 16 node chain, you have the greatest chance of experiencing the issue. ...and don't assume the chain starts with the host and goes in order of drivers in the list. It may be random or it may be related to the order drivers enter the track.

 

[Speedy6543]

Working further on this post.

Good news!
As I do a livestream every week of the race, which is viewable here, you guys can all see the difference between my quali grip and race grip.
Also, take note that the GT5 in race timer, in this livestream video, is a 100% accurate. Which wasn't the case when I reviewed the replay of this race........
You can all take out your stopwatch and FFWD to 21'00.0min and see my second lap which I already timed in the replay and being 14 sec off then.

...
Sorry I didnt get around to looking at this earlier, but.. What is that video of? It is not the race you uploaded a replay of. The race winner is not even in it.

Sorry, I checked that link again and the significant gameplay of the race isn't there anymore...
I don't know how this sites go to work with which ones to save and which ones to delete after ~a week...
It was a video of my gameplay at that moment and you could see that the GT5 timer was running a 100% accurate...
+ you were able to see the difference between my qualifying grip and 'in race' grip.

 

[Yenadar]

Just adding in that we experienced this for the first time in our last league's race this past Monday night. It was also the first time we were over ~9 cars (14). Grand Valley Speedway, most of us were putting in practice lap times in the 1:48 to 1:50 range (575pp TC/RM cars). We started the race and literally after the first turn people were already making comments like "what the..." The fastest driver for lap 2 was a 2:01. 13 seconds slower than practicing. Most people were in the 2:02 to 2:10 range, and the guy leading the race was putting in practice fast times of 1:55 (still learning) and was even commenting "something is wrong, there is no way I'm suddenly faster than everyone else!"

We restarted the race, but got the same result. We then split the room, 7 people moving to a different lobby, and we immediately got our speed back and were turning 1:48-1:50 race laps again.

Edit:
Unknown if this is at all connected, but throwing it out there in case it might be. We all know tire wear was dramatically accelerated with 2.02. However, while we were practicing, most of us were getting around 5-6 reliable laps out of a set of Race Softs, but this was extended by 2 laps for everyone once we split the room and actually went racing. 7-8 reliable laps. If it was just "smoother driving" or whatnot it wouldn't be consistent with everyone. Even the guy that could more than 5 reliable laps on soft said he got 2 more laps than he was in practicing. From a coding angle, I know that tire wear is a function of perceived slipping on the pavement (just watch a replay from someone massively lagging to see their tire wear burning through tires over a few seconds of lag induced fake sliding, and their tire wear 'reset' when the car does before going off on another pinballing smoke fest melting through the tires until reset again). Perhaps the coding that is related to the smoothness/processing/framerate is also accelerating tire wear a bit more as well. Possibly as a result of the introduced grip issues. Reduce the room count and everything smooths out, including slightly better tire wear (Though understandably still much more than pre 2.02)

 

[MGR]

I really don't think a racing wheel has anything to do with it. I've gotten the same reports in my full lobbies from DS3 drivers.

The only reason you are making the connection to FFB is because most of the highly skilled online racers use some form of FFB and it is more noticeable.

No, I'm making the connection to FFB wheels because I have personal experience that on numerous occasions where this bug has become present that only the FFB wheel users in our lobby were affected. So often, and so consistent that I originally was convinced the use of a FFB wheel was the key contributor to the bug. It has since come to light that even DS3 users can suffer the same bug but based on my own experience and feedback from GTP members jon- and Speed6543 (as per the OP) the use of a FFB wheel greatly increases the chance of suffering from this glitch.

The problem is network related. If just one driver has significant lag, he can wreak havoc with seemingly random people in the lobby, even the host.

I have never experienced this in a small lobby and I have never experienced it when there wasn't someone from another continent in my lobby.

I would venture to guess the clocks are synchronized with some sort of round-Robbin algorithm and if you are at the end of a 16 node chain, you have the greatest chance of experiencing the issue. ...and don't assume the chain starts with the host and goes in order of drivers in the list. It may be random or it may be related to the order drivers enter the track.

How do you come to the conclusion this is just a network related bug based on guesswork?

Playing devils advocate here...

Assume it is a network issue. How can you explain why;

- this occurs far more often with premium (highly detailed) cars?
- does not affect free-run mode but does affect racing?
- it's more likely to occur while using a FFB wheel?

 

[nasanu]

Playing devils advocate here...

Assume it is a network issue. How can you explain why;

- this occurs far more often with premium (highly detailed) cars?
- does not affect free-run mode but does affect racing?
- it's more likely to occur while using a FFB wheel?

Why do you think it is more likely to effect wheel users? What exactly do you base that on?

It is logical that more wheel users will notice and report more bugs of any sort. After all wheel users are the most serious users and the most likely to be paying attention to lap times.

Anyway I've never encountered the bug and I've never played GT5 with a pad.

 

[bdubclub]

Just a thought, not sure if this has been mentioned or not, but could it be all of the paint chips and game cards(or whatever those things are) taking up too much memory?

 

[werty_56-1]

It happened a bug to me too when I was with my friends , started race , had black screen and when race started I was already in lap 2 with best lap 0:02:467

 

[jon-]

Why do you think it is more likely to effect wheel users? What exactly do you base that on?

It is logical that more wheel users will notice and report more bugs of any sort. After all wheel users are the most serious users and the most likely to be paying attention to lap times.

Anyway I've never encountered the bug and I've never played GT5 with a pad.

In the first thread I put up a number of months ago, only wheel users were affected.

I might be one of the few players who can switch between wheel and pad and maintain speed (don't ask how I got to that situation, I'd much rather be faster with one!), so when we all try and replicate this bug I should be able to test out the theory that it impacts wheel users more.

 

[MGR]

Why do you think it is more likely to effect wheel users? What exactly do you base that on?

It is logical that more wheel users will notice and report more bugs of any sort. After all wheel users are the most serious users and the most likely to be paying attention to lap times.

Anyway I've never encountered the bug and I've never played GT5 with a pad.

I race with the same group of guys every week. We're all based in the same metropolitan area and run either ADSL2 or cable. I know Australia isn't the exactly leading the world in internet speeds but these connections combined with the correct router configuration (NAT2) should be more than ample to transmit the relatively tiny amounts of data required for online racing.

Each week our group size will vary from 10 to 16 players. Most of them use a FFB wheel but there's usually 3 or 4 DS3 users amongst us. There are a couple of decent DS3 racers but most of them you would consider ''average'' in our group, tending to run mid pack to rear of field.

Whenever this bug pops up every DS3 user seems to be immune. These 3 or 4 guys are suddenly the pack leaders. Even if they qualify dead last they race to the front and are fighting out for podium positions with their other DS3 brethren. While every FFB wheel user is left languishing down the field having their own slow paced, cruddy handling battle.

It's happened far too often for me to believe it's just a co-incidence. Although only jon- and Speedy have offered useful data I'm sure if proper testing and surveying is done with other users/lobbies they'll find a similar connection.

 

[nasanu]

I race with the same group of guys every week. We're all based in the same metropolitan area and run either ADSL2 or cable. I know Australia isn't the exactly leading the world in internet speeds but these connections combined with the correct router configuration (NAT2) should be more than ample to transmit the relatively tiny amounts of data required for online racing.

Each week our group size will vary from 10 to 16 players. Most of them use a FFB wheel but there's usually 3 or 4 DS3 users amongst us. There are a couple of decent DS3 racers but most of them you would consider ''average'' in our group, tending to run mid pack to rear of field.

Whenever this bug pops up every DS3 user seems to be immune. These 3 or 4 guys are suddenly the pack leaders. Even if they qualify dead last they race to the front and are fighting out for podium positions with their other DS3 brethren. While every FFB wheel user is left languishing down the field having their own slow paced, cruddy handling battle.

It's happened far too often for me to believe it's just a co-incidence. Although only jon- and Speedy have offered useful data I'm sure if proper testing and surveying is done with other users/lobbies they'll find a similar connection.

So EVERY wheel user gets affected and EVERY pad user is unaffected? If not they simple chance could account for what you have seen. With only a few pad users simple chance would see mostly or even only wheel users getting hit. I know people here will call that logic crazy, but the science is on my side. It does not mean you are wrong, but it can certainly not be said that you are right.

 

[chuyler1]

That is a very solid report...and I'd be inclined to retract my statement if I hadn't experienced numerous reports from DS3 drivers of the same bug.

I guess I personally cannot rule out the premium vs standard vehicle claim. Most of my racing involves a mix of vehicles, mostly standards, or mostly the same model vehicle. I mention the "same model" because if it is a graphics frame rate issue, there could be a correlation to the amount of data stored in memory...and having 16 drivers in the same vehicle could have a smaller footprint compared to 16 drivers in 16 different premium cars.

...but in the end, all my experiences have been linked to drivers from other continents entering my races and the issue disappearing after they were replaced with drivers from the same continent. There were a few issues with drivers on the same continent but eventually we narrowed their issues down to a bad router or network connection.

 

[dsgerbc]

Although only jon- and Speedy have offered useful data

So my report was useless then?

And everybody pretty much ignored my call to start discussing the set-up of the survey. And a random thought: wonder if unplugging the wheels helps at all.

 

[daymobrown]

The physics engine runs at something like 1000 times a second

Can you please explain a little more about what you mean?

 

[BWX]

It's very possible we're dealing with multiple unrelated bugs or code problems that cause the same type of symptoms. Some happen with GFX slowdown, some without.
I never experienced the GFX slowdown part, just felt like my car was slower than the previous race.