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

[MOTORTRENDmitch]

I have came to the conclusion there is a definitive Bug online.. On Mazda Laguna Seca, I find it happens more then anywhere else, Turn 1 and Turn 2 of that track are Notorious for this, I run 500PP there alot, I use all sorts of cars, but In Particular the Shelby Series 1, and the Elise.. Both handle superbly. I was following an M3 E46 in my Elise one race, Turn two comes along, I do my normal turn in, except for the fact that the car goes straight, off the track, and doesn't slow down at all until I slam the wall, I was doing around 50mph (estimate) and even though when I started understeering I was on the brake a little and completely off the gas, it didn't slow down in the least, and when I was on the sand, which is known to slow the car down rapidly, I was on the brakes HARD, and It only slowed down to 46MPH throughout the whole Fiasco until I slammed the wall. There Is absolutley no reason for a Car on Racing Soft Tires, Warmed up, and only on lap 2 ( Hardly Worn At All ) to lose overall grip to the point your car is on Black ice. I hope PD can fix this issue, because it has ruined more races then I can recall, and It makes me look like a HORRIBLE driver if I'm in a room with people I don't often race with.
Glad I could put my 2 cents in.

 

[nealcropper]

seems like this is bigger than we thought eh?

 

[MOTORTRENDmitch]

You're darn right, The only way I can describe it is when you are putting excessive force on the tires, they suddenly turn into Comfort tires and you are outta luck.

 

[Sail IC]

MGR and other,

I'm in a competitive closed group racing league, and I can also suscribe to this issue. We have experienced it but not to the same level as some of ther other posts, but in the 1-2 seconds level. Most if not all are using streering wheels, so I don't know about that part, but I agree to the other "reasons" causing it.

We normally race with a maximum of 14 drivers and we first discovered it when we for some reason were 15 drivers. As some people got pissed about lack of pace in the race and quit, things improved in the room.

Unfortunatly we don't have a fix or what causes this but I just got to think about something.

Before races we run a 20 min qualification session. In the "15 driver" race, nobody later affected in the race had any problem in the qual. Given the shortish qual session, all cars will be on the track at the same time. So thinking loud ..... what is different between the qual in "free run" and the race itself.

In the race, at least initially the cars will be much closer bunched togehter putting higher load on the graphic system on the PS3. Having said that, if its 12 or 15 on the starting line might not big difference on how many cars the graphic system has to draw at the beginning of the race.

The other thing I just thought about is that ..... and free run, the PS3 does not record what is happening on the track, but in the race it records what is happening to every car.

Does anyone know if there is a way to "turn off" the record function. This might be a solution to the problem.

 

[SkengD]

I think this issue has to do with internet connection and the # of ppl in the room.

What usually happens is other racers get faster and then others get slower .

I have only experienced this in rooms 14- 16 ppl.

We figured out this was due to racing in lounges and we started racing in fixed ownership lounges and this seemed to fix the issue .

What happened last night with Neal was a first because it was only 4 of us in the room .

Oh and this is not a steering wheel issue glitch/bug , its a general problem but as I said it doesn't happen all the time so I think its something we will have to live with for now .

 

[GODfreyGT5]

From my first online race, there has been a marked difference between Free Run and Race. Less grip and slower lap times, this always seemed to affect everyone as I will usually finish by a predictable gap behind the fastest qualifiers .

However, +10 seconds is a much bigger differential than I've witnessed (carnage aside).

In my experience, everyone else runs equal or faster times than in quali during the race, except for the one or two cars affected.

 

[MOTORTRENDmitch]

I might also add, This only seems to happen when there is a lot of network traffic in the room

 

[JaySin85]

I just realized this might have happened in another race as well. It was a strict clean room. I was doing really good for many races. The car I was in was an Opel Speedster Turbo. Then one race, I couldnt get the car to drive right anymore. To the point where I bumped another player, then in the same lap (I think it was even the first lap) understeered so bad that I turned away from the corner so I didnt hit the person in front of me, in hopes that it wouldnt get me kicked. Well, they saw it and kicked me anyways. Even though they saw me driving very clean every other race. At the time I was very confused as to why I suddenly lost my ability to control the car. I figured I had just been at it for too long, so I took a break. Now with this thread, it seems more likely that I ended up experiencing this problem.

There were also several people using mics throughout my entire time in that room.

 

[Lambob]

I'd like to share that this very last update is a minor stepback in ways of handling of cars with front toe and front camber on racing softs. I think something severly broke in handling characteristics, where as soon as you add even 0.01 +/- Front camber or toe, or both, in some hopefully "positive" handling trait, the car just drives worse. The cars ted to push all over the corners with front camber between 0.1 and 1.5, whereas they grip happily with 0.0 front camber. (Front only) I now use factory suspension setting mimicked into the fully cust suspension, and cars drive best with 0 front settings, and super soft springs and dampers on the Nurburgring.

I didn't test out racecars per say, perhaps front camber isn't good for "a street car"

Something is out of whack, although I appreciate all efforts by PD to make the Nur a harder and or more realistic approach, but I think with the "undocumented change" of tire wear, this could be a beta test for all we know. What if this new tire wear, is a bug? lol

Happy New year all.

Oh, and online lap times were always always a tad slower than offline real practice sessions, I just accept this fact. But cars are handling somewhat less confidently than before this last update. G25 wheel user, FFB at level 6 (I find 7 - 10 way too harsh now, and hampers steering)

 

[seanneedscar]

I havent noticed this before but i never ever race with more than 8 or so people, because of the lag and contact.

 

[dudeondacouch]

It's been established there's two different physics engines operating in GT5 between online and offline.

100% agree that the physics is different, but I don't think there's actually a different model or engine operating.

I think the connection, number of participants, and room quality setting controls how frequently the computer gathers data about each car's physics, so higher settings/better connections essentially increase the "resolution" of the physics engine. I believe that when this "resolution" gets below a certain point, for whatever reason, it leads to the effect of significant grip loss.

This is, of course, never an issue offline, and if you go to your own lounge with a high speed internet connection and "very high" quality setting, you'll find that the handling is very very close to offline modes.

I have no data to support my hypothesis, except my own personal experience. I run in a private league and everyone knows to do their tuning for each week in their own lobby at the "high" setting, because that's what we use on race day. We found that cars tuned in A-spec or Practice mode DO NOT perform properly in the online lobby, and as many have said, understeer/oversteer are exaggerated.

Perhaps the use of a wheel leads to the engine having to handle more data and contributes to those users experiencing the low "resolution" effect before DS3 users.

 

[dsgerbc]

It has a name - phantom lag. Don't know if it affects only people with wheels, but definitely happens frequently in rooms larger than 12. It's nothing new and I had more than a few instances during last year. If you race in one-make room with no engine/gear tuning and see people overtake you like you're standing from the get-go, you better brake for turn one waaaaay early, cause otherwise you'll under-steer into a wall. After that, despite cornering speeds showing nothing different you'll be 3-4+ seconds a lap slower.

 

[r2dc2]

This topic has finally prompted me to sign up with GTplanet. I can confirm I get it, all the time online.

Furthermore I've noticed that get it on both large and small rooms.

And on large and long races of more than 10 cars with decent drivers and more than 8 laps, I have noticed that occasionally the pace problem becomes very obvious. The race pack seems to split into 2, with the fast lot up front, and a very slow pack of maybe 3-4 cars, loosing about 2-3 secs a lap consistently (for example, at suzuka) to the lead pack.

Needless to say, I'm always stuck with the slow pack. We all seem to be driving consistently and reasonably competitively within the slow pack, but when I observe during recorded replays, all of us in the slow pack have very obvious and irregular traction issues.

One point of view is that the slow pack guys don't know how to drive or tune their cars, but we can't all be driving at the same slow pace and be consistently slow together, lap after lap, can it? It certainly seems like the slow pack guys are all experiencing the phantom lag issue and are similarly slowed.

 

[Brock1Samson9]

I've played GT5 online a lot and only ever with a wheel and can't say I've ever experienced this or heard about it either. Sounds very unusual

 

[dsgerbc]

^ do you run in a lot of spec. series with qualifying, where you spend quite a bit of time practicing a particular car on a particular track before the race? Otherwise you won't know what laps you could've been running w/o this lag.

 

[MGR]

MGR and other,

I'm in a competitive closed group racing league, and I can also suscribe to this issue. We have experienced it but not to the same level as some of ther other posts, but in the 1-2 seconds level. Most if not all are using streering wheels, so I don't know about that part, but I agree to the other "reasons" causing it.

We normally race with a maximum of 14 drivers and we first discovered it when we for some reason were 15 drivers. As some people got pissed about lack of pace in the race and quit, things improved in the room.

Unfortunatly we don't have a fix or what causes this but I just got to think about something.

Before races we run a 20 min qualification session. In the "15 driver" race, nobody later affected in the race had any problem in the qual. Given the shortish qual session, all cars will be on the track at the same time. So thinking loud ..... what is different between the qual in "free run" and the race itself.

In the race, at least initially the cars will be much closer bunched togehter putting higher load on the graphic system on the PS3. Having said that, if its 12 or 15 on the starting line might not big difference on how many cars the graphic system has to draw at the beginning of the race.

The other thing I just thought about is that ..... and free run, the PS3 does not record what is happening on the track, but in the race it records what is happening to every car.

Does anyone know if there is a way to "turn off" the record function. This might be a solution to the problem.

I think you're onto something here....

I totally overlooked the fact the PS3 does not record any of the free run mode as a replay. Of course, come race time that all changes with the console now required to keep a log of every car and its position, current gear etc. That would definitely explain why the problem is not present in free run mode.

This coupled with the additional work required to provide FFB effects may be the reason why wheel users suffer so much from this issue. I can handle framerate drops when the screen gets busy but to have it effect the physics/game clock is absolutely unforgivable.

In my opinion this is easily the biggest problem with the game. It basically means that competitive online racing is pointless.

 

[BWX]

In my opinion this is easily the biggest problem with the game. It basically means that competitive online racing is pointless.

Exactly, but I doubt PD have any clue the problem exists. If they do know, I wonder if they even care?

PD's attitude about huge bugs/issues in GT5 is the real problem, yet they work on other pointless stuff constantly. Off-line is still good, but online is frustrating as hell because of horrible net-code and inconsistencies in grip. Oh well.

G25 wheel user, FFB at level 6 (I find 7 - 10 way too harsh now, and hampers steering)

Interesting.. G25 user and exact same settings now- 6. Before, 7 all the way to 10 felt good.
BTW, Spec 2 physics in general, I thought were great.. 2.02 physics are messed up again. They should learn to quit while they are ahead sometimes. Maybe focus attention on huge glaring problems like net-code?

 

[jon-]

Not sure if anyone has linked this yet OP but you'll be interested in this thread:

https://www.gtplanet.net/forum/showthread.php?t=222711

We basically found the same, but it wasn't 100% limited to wheels, just mostly. Seemed worse on high draw circuits like SSr5.

The most interesting thing seemed to be when playing back the replays of people affected the stopwatch vs in game timers were out of sync.

 

[eran0004]

Could it have anything to do with the fact that the physics are different in free run (qualify) mode and race mode? For instance, there is no drafting in qualify mode, maybe there are some other aspects of the physics that are changing as well (like more wind resistance when racing compared to when qualifying/free run?).

As I understand OP it seems like this glitch is not only about changing from offline to online, but also from free run to race in an online room, correct?

 

[jon-]

Could it have anything to do with the fact that the physics are different in free run (qualify) mode and race mode? For instance, there is no drafting in qualify mode, maybe there are some other aspects of the physics that are changing as well (like more wind resistance when racing compared to when qualifying/free run?).

As I understand OP it seems like this glitch is not only about changing from offline to online, but also from free run to race in an online room, correct?

Correct, and an interesting theory.

Slipstream / drafting is the only physics difference I know between free run and race though, and if you're running on your own in the race then the physics engines should be fairly similar.

My vote would be load on the ps3 as it only appears on the more taxing circuits, with a lot of detailed cars (premium)

 

[MGR]

Off the top of my head I recall specific examples of the bug being present on Deep Forest, Spa, Autumn Ring, Cote de'azur, Sukuza and Trial Mountain.

I was going through my list of replays from our weekly one-make races. I recall one race on Tsukuba with a near full room that ran without a hitch:

8th September - Tsukuba - Honda Civic Type R (EK) '97 (Race Mod)
15 starters, 14 finishers.
No problems reported.

15th September - Laguna Raceway - Dodge Challenger SRT8 '08 (fully tuned)
Only 10 starters, 6 finishers.
I don't recall any issues.

22nd September - Special Stage Route 5 - Mitsubishi Lancer Evo IV GSR '96 (fully tuned)
Only 10 starters, 9 finishers.
I don't recall any issues.

29th September - Nurburgring GP/F - Mercedes-Benz C 63 AMG '08 (fully tuned)
13 starters, 12 finishers.
I don't recall any major issues, although this car destroyed tyres so lap times were not consistent.

6th October - Deep Forest - Fiat 500 1.2 8V Lounge SS '08
14 starters, 13 finishers.
Definitely major issues with this race.

The Fiat race was my first 'HOLY COW, this game is screwed!' moment. I suspect there may have been issues earlier but I wasn't 100% sure until this point.


Was there a patch released between Sep 8th and October 6th?

Where can I find a full change log listing all GT5 patches?

 

[nealcropper]

there was one release on the 8th of september

 

[MGR]

Never mind. I just googled 'GT5 Patch 1.10' and kept incrementing by .01 until if found them.

Patch 1.12 was released around the 3rd September. I remember this one because it fixed the FWD LSD problem and we we're going to race the Honda Civic later in the week, so that was a welcomed patch.

The next patch 1.12.572 (or 1.13 as it's commonly known) was released on 21st September.

http://us.gran-turismo.com/us/news/d5958.html

A new update has been released for Gran Turismo 5.
The main contents of the update are as follows:

- Prohibited driving areas on the Top Gear Test Track have been adjusted.
- The rankings of the current Mercedes Benz Driving Event has been reset, and the event duration has been extended.

As usual the patch notes are not detailed only listing the 'main contents'. However a GT Planet thread suggests that perhaps 2P replay mode was reinstated with this patch.

https://www.gtplanet.net/forum/showthread.php?t=223814&page=2

If that's the case I wonder if whatever they did to the replay system to allow recording of 2P races again has caused an issue recording the online replays?

 

[dsgerbc]

^I had this issue waaay before Sep'2011.

Correct, and an interesting theory.

Slipstream / drafting is the only physics difference I know between free run and race though, and if you're running on your own in the race then the physics engines should be fairly similar.

During affected race even on your own you won't be anywhere near your proper pace. So it's not draft physics. Turning visual damage off might help reduce the frequency of this.

 

[Sail IC]

The race we discovered this was on Sep'29th.

We run weekly races, and at this race the effect was big enough that it was obvious. But as been mentioned earlier, it might have been earlier but if the effect is not in the seconds its very difficult to know if its there or not.

 

[jon-]

^I had this issue waaay before Sep'2011.


During affected race even on your own you won't be anywhere near your proper pace. So it's not draft physics. Tuning visual damage off might help reduce the frequency of this.

Agreed, we've had it since around the middle of the year - we changed championship rules to run 2 different cars (same PP, didn't really work out so we're going back to 1 car) but the 2 different cars on track seemed to make it even worse.

By far the worst we had it was Special Stage 5 Reverse, with a premium ACR Viper v Lambo Murci LP670SV (stock) on sports softs.

Get 13+ people in a room and try and run your quali pace in the race. Nearly every wheel user was 3-4 seconds off quali pace and some of the pad guys struggled a bit too.

ETA Our first major issue was 11th Sept

 

[boshuter]

We have run into this issue a few times in our series. When it happens, it happens just as badly in free run before the race as it does in the actual race. Most of us in the series will run hundreds of laps during the week in practice for the race, we know exactly how fast of laps we can run, and we can run them consistently. When we get to the lobby or room on race night and start doing free run laps, it becomes obvious very quickly if we are having a problem. It affects DS3 users just as much as wheel users. We usually just all leave, open a new room, and the problem is usually gone.

 

[dudeondacouch]

^I had this issue waaay before Sep'2011.


During affected race even on your own you won't be anywhere near your proper pace.

Agreed on both counts.



I don't think this issue has anything to do with either drafting or replays. That just doesn't make sense.

I say it's clearly an issue with how often the system gathers and interprets the data about car locations. Whoever has the slowest ping/least bandwidth (whether due to connection issue, distance from host, car specifics, or the possible increase in data due to wheel use) suffers most.

This is evidenced by the fact that the physics change simply be being online. Go tune a car in practice mode and get a consistent laptime, then take it to your own lobby and try to replicate your time. Not only will you not be able to do it, but the lower the quality setting for your lounge, the further off you will be. I don't think there's a possibility of there being a different physics model (other than the axial rotation limit to prevent rollovers) for online play. So it has to be the "phantom lag" that dsgerbc mentioned.

 

[davelis]

I think it is the physics that are adjusted for all those people racing simultaneously.
Sometimes, (especially in NBurburgring Nordschleife) i have experienced a "lag", as if everything runs at 1/2 of the normal pace for a couple of seconds when the pack of cars is still tight.

 

[Sail IC]

This is a very good and interesting thread, at least for the people in online racing leagues. I don't think neither of us know the source of it error hence its difficult to find out how to avoid it or minimize the risk of it happening. I also don't think PD is aware of it.

Persumably a part of the system gets sort of overloaded. I guess it could be any of the graphic system, networking, communication, I/O, or computing part of the overall system.

Networking:
To mimimize lag we only allow European players in our league. As there is very little information actually communicated (only cars position and path) we been thinking response time is much more important than bandwidth. We all have measured ping rate between us and most have faster than 60ms with the slowest still below 100ms. I first thought long response time was the root cause but I'm not that sure any more. Sometimes there could be visible lag in the room but lap times seems normal. In the race we had the bigges effect we didn't have noticable lag.

Graphic system:
In the initial post it was mentionded this is more likely to happen with premium cars which have more details hence load the graphical system more. I've been thinking that if reducing the resolution on the HDMI output would improve it, but as its so difficult to know if this is happening or not I don't know. But it would be interesting to know what display resolution the people affected the most are using.

I/O system:
Its been stated its more likely to happen to people with steering wheel. I've been affected and are using the steering wheel settings simulation, FFB=4, power steering=off. Maybe some other setting such as using profissional/amature instead of simulation could fix it. Also, if someone knows a way to turn the replay off in race mode I like to know that. Or if there is a way to speed up the disk I/O like de-fragment the HDD.

Some in other threads have reported that only people with the slim PS gets this but I'm not sure.

As previously stated, this is really bad. I know I only been affected once at the seconds level, but I do often have problems reaching qual lap times in the race so maybe I always get this in the tenth of seconds in each race which is a lot in a competitive racing league. It really pisses me off not knowing anything about it.