- 2,073
- Newcastle, NSW, Australia
- mrvajj
Lemme guess - Gandalf was in the boot?
When is it a feat - and when is it a cheat? [The Suspend/Resume Bug]
- Thread starter Famine
- 79 comments
- 12,869 views
- 1,105
- NC/NoVA
- dudeondacouch
If that were the case, there wouldn't be much cause for completing 157 laps, would there?
- 283
- California
- RoadRocket77
I'm gonna agree with the guy that said it was a loading glitch with the Save game feature. That sounds like a GT5 glitch... they lost the GTAuto Creeper in the last one, so it sounds like it was a screwup in the Save feature for endurance races.
- 84,449
- Rule 12
- GTP_Famine
Heh, no. That would affect all players in all endurance races - and you can see my AI are notably slower than the two examples, despite my AI having a completely clear, dry run.
That would also have affected the dude using the R35 TC. Incidentally, that car is a beast at any PP level.
I only used 4WD because I anticipated rain and I know that you don't need much car on wet tyres to beat the 905 and R10 on wets. It didn't rain though
You can block one car indefinitely on the Ring. You can't block more than one, particularly if they have a 50mph straight line advantage. Six... no.
There's something with those numbers...as if after 4h the A.I. changes the pattern.
And why is the R10 ahead the 905, and both behind the McLaren? Both are superior than the McLaren...right?
...I dunno, you got to give us more evidence Famine. Maybe by combining all of this previous guesses from other members.
Oddly, the order is just about as I would have expected from a wet race. The 905 and R10 are hopeless in the wet and the McLaren and Pagani are reasonable. It didn't rain though, so you're right that the order is a bit screwy. Not a cause of the cheat, just a symptom.
No, though that's a really good call.
In timed enduros, when the timer runs out the cars are classified by distance. This means that any car that finishes, say, its 156th lap and finishes is briefly classified ahead of any car that has finished its 156th lap but is still running on lap 157. This lasts until those cars finish - but of course if it's the player car classified ahead, you just need to mash "X" to be awarded the win.
But you need to be on the lead lap - you won't be classified ahead of cars that have finished 157 laps and are running a 158th if you finish 156 and finish. Plus that doesn't explain why the AI are more than an hour slower than they should be...
I actually experienced some lag - amazingly - when my router fell over and I was logged out of PSN. Ridiculous.
Yes. Spot on.
Yes - I did so after 5 hours of my R34 TC race. That's how I discovered it.
I put on my wizard's hat and robe...
Interesting...forget about that feature.
Let's assume that the A.I. is loaded correctly into the track in the same position and location on the track, BUUUT, pit in after that?
Now we're getting to it - though yes "feat" means "accomplishment" not "feature", Husky. I should point out though that it didn't rain at any point in my race - the race screen showed 0% surface moisture at all times.
It is very specifically caused by an oversight in this part of the game. There's a number of bugs at play, but they're unified by a single thing the player can do. My own term for it is "the Suspend/Resume exploit"...
- 1,105
- Bakewell UK
- dunkrez
- spunwicked
I'm intrigued, come on Famine, spill the beans before I have to go to work 
- 5,283
- South West
- KAYTEAM73
I had noticed that my half a lap lead would become a 1 and a half lap lead after continuing a suspended save....
Years ago, playing Formula 1 2005, you could 'save' at any point on the track at any point in the race.
Loading the game - all the AI were where you left them.
I was surprised that GT5 didn't have this function.
Is a race suspension a save? Does it save the 'gameplay' at that very moment?
I don't think it does...
Years ago, playing Formula 1 2005, you could 'save' at any point on the track at any point in the race.
Loading the game - all the AI were where you left them.
I was surprised that GT5 didn't have this function.
Is a race suspension a save? Does it save the 'gameplay' at that very moment?
I don't think it does...
- 84,449
- Rule 12
- GTP_Famine
Okay okay! I think we're close enough to class it as having got it.
The Suspend/Resume Exploit
This affects all endurance races with variable weather - that's the Suzuka 1000km and the two 24 hour races. There's a number of bugs at play here, some of which we're already familiar with, but ultimately they can be triggered by one thing - suspending an endurance race and then resuming it later.
When you start a weather-change race, the weather is randomly-generated from a large field but it will most commonly settle on 100% wet and raining or 0% wet and clear. A preview will load, the AI cars will tootle round on a demonstration and they will select their tyres accordingly - you, of course, can select your tyres as you choose.
When you pit, you are given the option to Suspend the race. This saves your position as you enter the pits, along with the AI positions and quite a lot of other information (it looks like it adds about 2Mb to your gamesave) and allows you to quit the game.
When you load the game next, it will load up a new preview screen that allows you to Resume where you left off or quit. Here's where the bug surfaces. The new preview screen also randomly loads weather, the AI also tootle round on a demonstration and they also select their tyres based on this - bizarrely, one of the pieces of information that isn't stored in the gamesave is "current tyre grade".
These two images are from two successive Resumes. One shows 0% moisture and not raining - the AI cars on their demonstration ran Racing Hard tyres - and the other shows 100% moisture and not raining - the AI cars on their demonstration ran Racing Wet tyres. In fact the moisture level can be pretty much any value (I've seen the whole spectrum - from 7% through 43% and 83%) but, like the race preview, 0% and 100% are by far the commonest.
When you then Resume the race, all cars are on the tyres they were running in the demonstration (this includes the player car, which has a demo AI run too, but it's irrelevant as you're in the pits...), though the wear levels are retained! So if you're in a completely dry race and Resume with 100% moisture, the AI are now running old Racing Wet tyres that they burn out in a few corners, scrubbing 3 minutes or more off their Ring laptime.
Now, this is player-fixable for the most part. If your race isn't raining and your Resume Preview shows 100% moisture, it will fall - and quickly. Wait for it to reach the right level, Resume and you're golden. No cheating. If your race is raining and your Resume Preview shows 0% moisture, it will rise. Wait for it to reach the right level, Resume and you're golden. Alternatively, quit the game and reboot to get a new moisture level. However, if you're at an unusual level of moisture - say a drying track at 27% or a dampening track at 52% - you're screwed. It is also easily exploitable.
There are variants of this bug - and I experienced several corollary errors during my race - but this is the most basic form. GTP user Chemenu noted something a little different which I can't reconcile easily, but I imagine it's based on a bug within a bug. Emptyone also noted an instance of this bug - which I've confirmed for him.
So the issues at play are:
Of course the problem is that now, long endurance races cannot be trusted. All it takes is a Suspend and you can win in anything. I Suspended and Resumed the race every three laps to get this quite apparent effect to win in a GT300 car. The cycle went like this:
Lap n: Resume; Leave pits; Pass stricken AI at Aremberg-Kallenhard; AI forced to pit behind me; Acquire 2 minute lead
Lap n+1: Fast lap
Lap n+2: Pit-in; Suspend; Resume and all AI are forced to pit behind me; lead of about a minute
Lap n+3: Leave pits;
Lap n+4: Fast lap; Hold AI up on Dottinger or allow them to pass and pit in their normal, 2 lap cycle
Lap n+5: Hold all AI up and allow past at Dottinger (so that they start a new lap); Pit-in, Suspend
Rinse, lather, repeat.
The Suspend/Resume Exploit
This affects all endurance races with variable weather - that's the Suzuka 1000km and the two 24 hour races. There's a number of bugs at play here, some of which we're already familiar with, but ultimately they can be triggered by one thing - suspending an endurance race and then resuming it later.
When you start a weather-change race, the weather is randomly-generated from a large field but it will most commonly settle on 100% wet and raining or 0% wet and clear. A preview will load, the AI cars will tootle round on a demonstration and they will select their tyres accordingly - you, of course, can select your tyres as you choose.
When you pit, you are given the option to Suspend the race. This saves your position as you enter the pits, along with the AI positions and quite a lot of other information (it looks like it adds about 2Mb to your gamesave) and allows you to quit the game.
When you load the game next, it will load up a new preview screen that allows you to Resume where you left off or quit. Here's where the bug surfaces. The new preview screen also randomly loads weather, the AI also tootle round on a demonstration and they also select their tyres based on this - bizarrely, one of the pieces of information that isn't stored in the gamesave is "current tyre grade".
These two images are from two successive Resumes. One shows 0% moisture and not raining - the AI cars on their demonstration ran Racing Hard tyres - and the other shows 100% moisture and not raining - the AI cars on their demonstration ran Racing Wet tyres. In fact the moisture level can be pretty much any value (I've seen the whole spectrum - from 7% through 43% and 83%) but, like the race preview, 0% and 100% are by far the commonest.
When you then Resume the race, all cars are on the tyres they were running in the demonstration (this includes the player car, which has a demo AI run too, but it's irrelevant as you're in the pits...), though the wear levels are retained! So if you're in a completely dry race and Resume with 100% moisture, the AI are now running old Racing Wet tyres that they burn out in a few corners, scrubbing 3 minutes or more off their Ring laptime.
Now, this is player-fixable for the most part. If your race isn't raining and your Resume Preview shows 100% moisture, it will fall - and quickly. Wait for it to reach the right level, Resume and you're golden. No cheating. If your race is raining and your Resume Preview shows 0% moisture, it will rise. Wait for it to reach the right level, Resume and you're golden. Alternatively, quit the game and reboot to get a new moisture level. However, if you're at an unusual level of moisture - say a drying track at 27% or a dampening track at 52% - you're screwed. It is also easily exploitable.
There are variants of this bug - and I experienced several corollary errors during my race - but this is the most basic form. GTP user Chemenu noted something a little different which I can't reconcile easily, but I imagine it's based on a bug within a bug. Emptyone also noted an instance of this bug - which I've confirmed for him.
So the issues at play are:
- Tyre grade data not stored
- Tyre grade selected at Preview screen (exploitable as seen in the B-Spec Spa Seasonal)
- Preview screen conditions not tied to race conditions
Of course the problem is that now, long endurance races cannot be trusted. All it takes is a Suspend and you can win in anything. I Suspended and Resumed the race every three laps to get this quite apparent effect to win in a GT300 car. The cycle went like this:
Lap n: Resume; Leave pits; Pass stricken AI at Aremberg-Kallenhard; AI forced to pit behind me; Acquire 2 minute lead
Lap n+1: Fast lap
Lap n+2: Pit-in; Suspend; Resume and all AI are forced to pit behind me; lead of about a minute
Lap n+3: Leave pits;
Lap n+4: Fast lap; Hold AI up on Dottinger or allow them to pass and pit in their normal, 2 lap cycle
Lap n+5: Hold all AI up and allow past at Dottinger (so that they start a new lap); Pit-in, Suspend
Rinse, lather, repeat.
- 2,073
- Newcastle, NSW, Australia
- mrvajj
Considering the information that IS stored during a race-save, that's a pretty massive oversight!
- 921
- SwoopGTR
slighty confusing but since I did the spa seasonal using this "cheat" I reckon I have the jist of it.
- 84,449
- Rule 12
- GTP_Famine
It is very similar to that, only it happens every time you Resume and you can reboot if you're not happy with it. It also means you can turn a 3 minute advantage whenever you want.
- 5,283
- South West
- KAYTEAM73
Nicely spotted Famine.👍
Come on PD, GT6 can only be better, surely.
Come on PD, GT6 can only be better, surely.
- 880
- Chesterfield
- DEKOWOLF
Interesting...I've noticed the suspend option didn't save everything when doing the 24h Le Mans. My best lap was messed up.
I can only imagine how long did all of this took you to organize, nicely done!👍
Let's hope this gets fixed. I want to win the race, but...not like that.
I can only imagine how long did all of this took you to organize, nicely done!👍
Let's hope this gets fixed. I want to win the race, but...not like that.
- 1,137
Can I ask you what kind of lag you are referring too?
I did my 24h LeMans race last summer, the first batch of DLC cars was released, never logged into PSN because I was afraid of corrupting my save (I already raced 10h or so).
During that time, I had a big input lag. When I gave throttle, the game would gave throttle about half a second late. Same with steering and gear changes. This almost ruined my race.
Regarding the bug, when was the first time you noticed this?
During my 24H race, I saved every 2 hours and I always checked the weather after resuming the race. The weather was always correct at the beginning, but it always falled down/raised rapidly after resuming the race. I raced with a stock audi R10 and was about 2-5s per lap faster then the Bentley at 2nd place and the gap only increased alot when the bentley pitted to late and had no fuel etc.
Btw, I also saw a car which had no fuel, but just didnt pit anymore.
Last edited:
- 84,449
- Rule 12
- GTP_Famine
It was about 5-10 seconds of severely reduced framerate - it reminded me quite a bit of what I experienced testing the online bug with WRS staff and participants.
16th February.
That sounds similar to what Chemenu reported.
I suspect there is some "bug-stacking" going on too. I backed up my save game every 3 laps when I suspended it for this reason and I noticed some very odd effects creeping in at times - two of the AI suddenly gained a lap for no reason, splits started becoming highly variable (even accounting for the known splits-to-last-car-to-pass-it issue that you most often see in Seasonals - when your split isn't to the leader to begin with, because he started beyond that split), AI pit behaviour became unpredictable and suchlike. It seems that exploiting one bug can cause others to pop up for some reason...
I just restored a last-known-good gamesave and plugged on.
- 880
- Chesterfield
- DEKOWOLF
That situation your describing, seams to be the current problem of this fellow racer, who is attempting the Suzuka 1000KM race.
https://www.gtplanet.net/forum/showthread.php?p=6736481#post6736481
- 3,531
- London, Ontario
- Starky513
Well done Famine.. for pretty much tackling this bug, and getting it out there! Seems like a relatively intracite bug.. Hopefully PD gets wind of this, and fix the issue 👍
- 140
- pinoyem1
Finally people are starting to notice this. I've only utilized this feature when I was participating in the Nur 24 hour event. I was leading in a Schulze GT-R fighting against a Pescarolo-Judd in second place. I had a 6 minute lead and counting, due to his tires being gone by his 3rd lap. I suspended on the 18th hour and resumed. Once I had done so, my lead just kept on decreasing. No matter what I did, even beating my previous best lap, my lead kept on decreasing. Usually, my lead would decrease a bit on his first and second lap after pitting, but I'd regain a lot of that plus more on his 3rd lap. Not this time. It just kept on decreasing, as if he wasn't pitting or his tire wear suddenly became non-existent. Your findings would explain a lot. Although I wish I would've done this work-around instead of starting over. Thanks for the report. 👍
- 4,040
- UK
- ReTyger
Famine, this is fascinating stuff but unless those other race results you included in your original post are from people you know, i'm not sure you should be assuming they also used this exploit?
I'd rate myself as an OK driver, but by no means a great driver, and I recently completed the 'Ring 4 hour endurance race. I also like a challenging race but must admit I don't really want to lose on my first attempt so I choose a car from the "typical opponents" list, but one I know is fairly powerful and handy (when I come round to doing the races again I will give myself more of a challenge!)
So I chose the Audi R8 Race Car (the one based on the production model), which although a little quicker than most of the AI cars (which were all tuner cars) isn't a great car by any means. I found it to be a little light and skittish, and it has a surprisingly low straight line speed (around 176mph on the back "straight")
However despite a few "moments" I still had about a lap in hand at the end, when I deliberately (and perhaps foolishly) slowed down to let most of the field past and briefly appear to unlap themselves (have to love how PD never fully fixed that, it just seems to figure out the total race distance after)
Anyway my point is that if I had a lap in hand after 4 hours, it wouldn't be too unusual to have 6 laps or more in the bag after 24 hours, because, let's face it, the AI drivers are just really slow =)
I'd rate myself as an OK driver, but by no means a great driver, and I recently completed the 'Ring 4 hour endurance race. I also like a challenging race but must admit I don't really want to lose on my first attempt so I choose a car from the "typical opponents" list, but one I know is fairly powerful and handy (when I come round to doing the races again I will give myself more of a challenge!)
So I chose the Audi R8 Race Car (the one based on the production model), which although a little quicker than most of the AI cars (which were all tuner cars) isn't a great car by any means. I found it to be a little light and skittish, and it has a surprisingly low straight line speed (around 176mph on the back "straight")
However despite a few "moments" I still had about a lap in hand at the end, when I deliberately (and perhaps foolishly) slowed down to let most of the field past and briefly appear to unlap themselves (have to love how PD never fully fixed that, it just seems to figure out the total race distance after)
Anyway my point is that if I had a lap in hand after 4 hours, it wouldn't be too unusual to have 6 laps or more in the bag after 24 hours, because, let's face it, the AI drivers are just really slow =)
Last edited:
- 4,040
- UK
- ReTyger
Also I dont think famine assumed we used the exploit after the spot the difference in the first post.
I just re-read it and that's still the impression I get, for the two pics which weren't his where Nissans won, the F1 still only completed 160 laps to his 156. I think it was meant as futher evidence of the exploit, not as a contrast.
But I can also see that whoever managed to win by that margin in the Skyline TC did so in an inferior car, so in fairness perhaps that would have only been possible with this exploit
What say you Mr Famine?
- 321
- New Zealand
R35 TC was tunedI just re-read it and that's still the impression I get, for the two pics which weren't his where Nissans won, the F1 still only completed 160 laps to his 156. I think it was meant as futher evidence of the exploit, not as a contrast.
But I can also see that whoever managed to win by that margin in the Skyline TC did so in an inferior car, so in fairness perhaps that would have only been possible with this exploit
What say you Mr Famine?
Have you seen post 39?
- 4,040
- UK
- ReTyger
Yeees I have, I still think you are missing my point. I think this exploit is an interesting find, and Famine has described it very well but I still get the impression from his first post that he is claiming the other two results screenshots are also proof of the use of this exploit (whether intentionally or not):
"The last few of that bunch are running pretty much of a muchness, but the lead AI are losing 2-6 laps (lest you think that trifling, it's 15 to 50 minutes)... How odd."
My point is that winning against the F1 etc. by several laps *without* the exploit is not so far fetched. Your confirmation that the R35 was tuned simply lends credence to this, as if it wasn't, I would certainly question such a large winning margin in such an inferior car.
The other result screen, 6 lap margin in a GT-class car which is superior to the stragglers in the field looks even more regular to me, and not evidence of this glitch/exploit. Clearly the exploit exists but I just wouldn't assume it came into play for these other drivers.
Last edited:
- 321
- New Zealand
I wouldn't call my GTR superior myself
I did have to do some work you know (Mine was not tuned, 518bhp, Racing Hards) If it weren't for the AI's suicide laps at the time it would have been a much closer race.I read it as the AI in Famine's race are not completing the number of laps they should be because of using the bug, so he compared the laps his AI did against others
Heres hoping PD fixes this soon. With this and the suicide laps fixed, I will redo this race.
