Patch 2.0 oversteer bug still present! (?)

[David Brooks]

Yes, so I see. Thanks.

Same question then, why does locking the diff (not allowing the wheels to spin at different speeds add so much stability mid corner?
I'd like to understand the logic.

Locking the diff inherently leads to understeer (i.e. stability). Its just then about tuning the diff to give the right degree of stability.

By the way, have you ever ridden an old school 3 wheeler farm motor bike with a fixed rear axle? The ultimate demonstration of understeer. Doesn't really matter which way the front wheel is pointing. It's going to push straight ahead. If you want to go around a corner you have to lean out to get the inside rear wheel to lift off the ground.

 

[redi]

Source?

Or is that speculation?
Gran turismo had a different physics model for online. It's not unheard of.

I've been part of pCARS1/2 development the past 6 years. If there were different physics/tyre models, we (WMD1/2) would have known.

 

[Bealdor]

Locking the diff inherently leads to understeer (i.e. stability).

In off throttle situations, yes. During acceleration, not necessarily.

 

[NLxAROSA]

Gran turismo had a different physics model for online. It's not unheard of.

It is unheard of outside of PD. Absolutely insane they did that, definitely not the norm for any other dev.

 

[David Brooks]

In off throttle situations, yes. During acceleration, not necessarily.

Agreed, but the question was specifically referencing the coast setting and mid corner stability.

 

[LPlates]

Thanks for all the info/clarifications.

 

[twitcher]

In off throttle situations, yes. During acceleration, not necessarily.

How does locking the diff cause oversteer during acceleration? That's counter to the basic principles of how a differential works.

 

[LPlates]

Diffs make my head hurt.

I totally accept that locking a diff reduces the ability for the wheels to turn at different speeds and therefore causes understeer.

What I don't get is why unlocking the diff would cause (massive) oversteer?

 

[RomKnight]

Good info here (take an aspiring first though )

http://forum.projectcarsgame.com/sh...ifferentials&p=1367069&viewfull=1#post1367069

TLDR;
http://forum.projectcarsgame.com/sh...ifferentials&p=1404805&viewfull=1#post1404805

 

[Bealdor]

How does locking the diff cause oversteer during acceleration? That's counter to the basic principles of how a differential works.

Because locking a diff sends more torque to the wheel with more grip (outside), which will result in said wheel losing traction and therefore your car spinning out.
A high acc diff lock will lead to understeer during low acceleration but to power oversteer during hard acceleration.

Diffs make my head hurt.

I totally accept that locking a diff reduces the ability for the wheels to turn at different speeds and therefore causes understeer.

What I don't get is why unlocking the diff would cause (massive) oversteer?

During coasting the weight is transferred forward which makes the rear "light".
Less weight at the rear means less grip for the tires. You need some amount of diff lock to work against that in pretty much every car.

 

[OdeFinn]

Diffs make my head hurt.

I totally accept that locking a diff reduces the ability for the wheels to turn at different speeds and therefore causes understeer.

What I don't get is why unlocking the diff would cause (massive) oversteer?

Don't worry, of above answers aren't either "correct" ones, differential locking/unlocking causes both understeer and oversteer.
People tells one way only and forget other, first way is when there's traction on tires and other when there isn't, oversteer changes understeer and vise versa on there.

 

[Gr8_Lakes]

I own a Jeep with selectable locking differentials front and rear. Until you've actually driven with fully locked differentials, you'll never understand how much it changes the drivability of a car. And I'm using it for off road traction mostly, but I experiment on road for better understanding. Amazing how much a locked differential can "push" until you finally get some wheel slip.

 

[MarleyMoo]

Because locking a diff sends more torque to the wheel with more grip (outside), which will result in said wheel losing traction and therefore your car spinning out.
A high acc diff lock will lead to understeer during low acceleration but to power oversteer during hard acceleration.

Yes, it sends more torque to the outer wheel. The higher torque outside vs lower torque inside means the outside pushes forward harder, therefore rotating the car into the corner, creating oversteer. This effect works in all of AWD, FWD, RWD. Note that some high powered FWD use front locking diff to fight power understeer.

No, it doesn't result in the outside wheel losing traction earlier. The diff only sends torque from the faster inside wheel to the slower outside wheel. The inside is less loaded and spins faster, therefore it loses traction earlier.

 

[Bealdor]

Yes, it sends more torque to the outer wheel. The higher torque outside vs lower torque inside means the outside pushes forward harder, therefore rotating the car into the corner, creating oversteer. This effect works in all of AWD, FWD, RWD. Note that some high powered FWD use front locking diff to fight power understeer.

No, it doesn't result in the outside wheel losing traction earlier. The diff only sends torque from the faster inside wheel to the slower outside wheel. The inside is less loaded and spins faster, therefore it loses traction earlier.

It's a bit of both really. The inside wheel will always want to slip first because of the lower load (no matter the diff setting).
However, with a high differential lock, as soon as the inside wheel starts spinning, more (read: too much) torque is transferred to the outside wheel, making it lose traction almost at the same time.
Since both wheels are losing traction almost simultaniously, the rear breaks away more suddenly (some people call it snap oversteer).

 

[mattikake]

Just had it again! EXACTLY like at Fuji. Except this time at a track in race conditions I know much better.

This time at Algarve in the 650 GT3. I can do 1:44s easy in race setup FP or online, loads of fuel etc. Nice and stable setup that I can push easily. But this one online race, of maybe 5 so far, the car was all over the place. Even oversteering in sections I can take flat comfortably. It was like the wrong setup was loaded into the car. Once it decides it's going to spin there is nothing you can do no matter how early you try to catch it. Nothing at all like other times where you can catch oversteer moments.

Interestingly, the quali was wet before the dry race. So I'm really wondering if there is something up with live track. Even if it was a really green track because of the rain, I believe the races always start with a green track anyway?

Something odd is definitely going on. No video this time but I have got a replay so maybe I can put some proof together.

 

[mattikake]

To add; what also occurs was in the wet quali the car was much less of a problem. It would understeer and struggle for traction as you would expect but no weird oversteer issues at all. Even in the corners where I couldn't go flat in the dry I was flat in the wet. So that confirms it for me. This is, yet another problem that MUST BE FIXED. 100% certainty.

 

[05XR8]

I did a couple drives in the GT-B Escort(stock settings) @Ruapuna, dry, 29/10/17, Spring, Light cloud 4 slots, 11am, XBONE, Controller.
Tight turns were fine. The longer radius turns(such as the final turn), I would spin at corner exit.
1. I maintained a constant throttle from apex to exit. Spin.
2. Coasted through turn, pit up throttle just at exit. Spin.
3. Steer with throttle from apex to exit. Slightly drift, but I got into a tank slapper. Spin.

I haven't spent much time setting this Escort, but have noticed the suspension is much lower than in PC1. I love it. However, the car seems very sensitive.

 

[twitcher]

I did a couple drives in the GT-B Escort(stock settings) @Ruapuna, dry, 29/10/17, Spring, Light cloud 4 slots, 11am, XBONE, Controller.
Tight turns were fine. The longer radius turns(such as the final turn), I would spin at corner exit.
1. I maintained a constant throttle from apex to exit. Spin.
2. Coasted through turn, pit up throttle just at exit. Spin.
3. Steer with throttle from apex to exit. Slightly drift, but I got into a tank slapper. Spin.

I haven't spent much time setting this Escort, but have noticed the suspension is much lower than in PC1. I love it. However, the car seems very sensitive.

Haven't tried the racing escort, but I tried the RX Escort at Dirtfish boneyard. I've never driven a car in PCARS that had so much grip.

Traction from the rear was good. Wheelspin on the mixed surface like you'd expect, but it still tracks straight and is predictable.

The throttle, omg, it actually works!! When traction at the rear is broken, modulating the throttle actually creates engine noises lol...and the car responds to it. Not like so many of the slides I've had where once the slide starts, input from steering, brakes, and throttle goes completes dead.

The front grip is what really stood out though. It was phenomenal!! Even on the mixed surfaces, it was rock solid, stable, yet responsive.

Not to pump my own tires, but I'm a member of what was once one of the top drift teams in GT5 and GT6. I have hundreds upon hundreds of hours going door to door with some of the best video game drifters of their time...I know how a car in a game should kind of feel while sideways.

It took me the better part of 2 years to learn how to drift in GT5. I couldn't drift period in PCARS1 (I don't think anyone could drift in that game, testiment to the amazing physics), it was impossible.

However, in PCARS 2, with that Escort at Dirtfish, within the first 15 minutes, on the stock tune (pre 2.0 when they improved the stock tunes), I was throwing Scandinavian flicks and reverse entries like a pro. Compared to anything else I've driven in PCARS 1 or 2, it felt like a completely different game.

Contrast that with driving some of the race cars, especially when they go into those dead feeling 90' rotation spins (I call it the generic PCARS crash, it's almost the exact same every time), it's really tough to figure out what's going on.

If I get time and motivation, I'm going to test that Escort RX at a bunch of different tracks. I'm curious to see how into handles driving on the grass surrounding various circuits. At Dirtfish on the boneyard course, there's a corner of grass you can cut across, and there, the traction on the grass is still relatively high.

 

[mattikake]

I did a couple drives in the GT-B Escort(stock settings) @Ruapuna, dry, 29/10/17, Spring, Light cloud 4 slots, 11am, XBONE, Controller.
Tight turns were fine. The longer radius turns(such as the final turn), I would spin at corner exit.
1. I maintained a constant throttle from apex to exit. Spin.
2. Coasted through turn, pit up throttle just at exit. Spin.
3. Steer with throttle from apex to exit. Slightly drift, but I got into a tank slapper. Spin.

I haven't spent much time setting this Escort, but have noticed the suspension is much lower than in PC1. I love it. However, the car seems very sensitive.

The bug is easy to spot. Start your ps4 from fresh and try the same conditions again. Load one of the standard setups. If it is this bug it is likely not to happen next time. From my experience this bug is happening about 1 time in 40-ish.

 

[JeffKill]

Trying out the 911 GT3 R tonight. The car feels amazing while tuning in private session or practice before the race. But as soon as I start a custom race, it’s like I’m driving on ice. Unbelievably frustrating.


Playing on Xbox with Thrustmaster TX. Even the ffb feels completely different when I start a custom race. I don’t get it.

 

[Animera]

Are the dates of your practice and custom race the same?
If not, track temps could be wildly different and affect your car's handling.

 

[BrandonW77]

Also remember that you have tire warmers on for the start of a race so your temps and pressures are probably going to be very different than what you had in practice/qualifying. My car/ffb always feel different (usually worse) in the race than in the preceding sessions, pretty frustrating and a bizarre decision to force us to always practice under different conditions than we race in.

 

[JeffKill]

Are the dates of your practice and custom race the same?
If not, track temps could be wildly different and affect your car's handling.

Everything was exactly the same. I tested on multiple tracks with the same result. In practice I did 10-15 laps to make sure my tires were up to full heat as well. Tons of grip and I can really push the car to it's limits. Get to the race in the same exact conditions and I'm on ice.

I tested this in private testing, then going to a custom race. And also setting up a custom race with a practice session first, and both were giving me the same result.

 

[Jack Dulmage]

I find that i encounter a distinct "blip" in my FFB and control... usually the ICM screen will come up as well when it happens, its almost like the car "resets" while im driving it... this usually makes the car undriveable for the next turn, super slippery like the rear wheels are on ice...

 

[JeffKill]

After searching around on some other forums and testing, it seems that this problem is directly related to how many AI opponents are on the track. I dropped it to 10 and the car has just as much grip as it does in practice now. The ffb still gets a little lighter during the race, but the grip is there and no more oversteer in every corner. I just turn the ffb volume up a bit right at the start of the race, and it’s good.

From what I’ve read now, this comes down to the standard Xbox just not having enough power to run everything. Really looking forward to trying this game on the one X.

 

[Racelow]

Have you even actually driven a car with drums all round on a track?

I would be more surprised if they were great than if they were a bit shoddy.

Now I've already said I will try it tonight and post up my findings. When will we get to see you driving it with your feedback?

Oh and you can cut the personal digs.

My thoughts on the Mustang, and its perfectly derivable regardless of corner.



Stopping distances are the only thing that needs a great deal of care, but that said it drives as I would expect a car of its era and type to drive.

This guy wants to say the Ford is perfectly fine to drive. Project2 cars have some problems with physics on various cars. I think it is fine to compare games to other games physics but when you stand an say it is perfectly like the real car and knowing it is not, I take problems with that. For most of the video he was driving the mustang way below the limits to look good but he still could not avoid that oversteer bug. At 2:25 the car does the strange oversteer but looking at his steering wheel input the car should have made the turn. This is not how the real car drives this is an insult to Ford if you claim it is. I think they try to make this game physics look perfect by ignoring the flaws. Take forza muscle cars and only using the controller you will see the physics is much closer to the real car, and they only simcade not trying to get perfect real car. I took this post to this one because I don't want to deal with any deluded fantisists who thinks there are no problems with this game. Every time that oversteer bug comes up the tell you to git gud. Well the guy in the video caught the oversteer bug not realizing this is the exact problem im talking about.

 

[mattikake]

After searching around on some other forums and testing, it seems that this problem is directly related to how many AI opponents are on the track. I dropped it to 10 and the car has just as much grip as it does in practice now.

From what I’ve read now, this comes down to the standard Xbox just not having enough power to run everything. Really looking forward to trying this game on the one X.

It's a real concern that ti's down to processing power as this will be difficult to overcome without optimisation for the consoles which would take a lot of work, possibly rewriting the physics which clearly take up a lot of that power. I would have no issues at all with the graphics suffering as a result, but maybe the way the graphics are not yet optimised may free up more space for processing the physics.

I have tried races with loads of AI and all I seem to get is a FPS drop... so far...

I came across the oversteer bug again last night. 650s GT3 @ the glen on default setup online public lobby. It was gracefully oversteering in the flat out turns 2, 3 and 4 as well as most other sections of the lap. These 3 corners should easily be within the limit of the car. I noticed a few other people were having the same issue, but others weren't. It wouldn't be so bad if everyone had the same issue but only half the field is a real problem for fair play. Again though, I wasn't recording. Must get into this habit!

 

[BrandonW77]

At 2:25 the car does the strange oversteer but looking at his steering wheel input the car should have made the turn. This is not how the real car drives this is an insult to Ford if you claim it is. I think they try to make this game physics look perfect by ignoring the flaws. Take forza muscle cars and only using the controller you will see the physics is much closer to the real car, and they only simcade not trying to get perfect real car. I took this post to this one because I don't want to deal with any deluded fantisists who thinks there are no problems with this game. Every time that oversteer bug comes up the tell you to git gud. Well the guy in the video caught the oversteer bug not realizing this is the exact problem im talking about.

Um, no, that wasn't a bug. He was a little late with his downshift which caused him to overshoot his braking point so he turned harder which caused the fronts to scrub and the rear to break loose, just as you'd expect a real car to do. Nothing weird about that at all, looked totally normal to me and I could see it coming.

 

[killerjimbag]

This guy wants to say the Ford is perfectly fine to drive. Project2 cars have some problems with physics on various cars. I think it is fine to compare games to other games physics but when you stand an say it is perfectly like the real car and knowing it is not, I take problems with that. For most of the video he was driving the mustang way below the limits to look good but he still could not avoid that oversteer bug. At 2:25 the car does the strange oversteer but looking at his steering wheel input the car should have made the turn. This is not how the real car drives this is an insult to Ford if you claim it is. I think they try to make this game physics look perfect by ignoring the flaws. Take forza muscle cars and only using the controller you will see the physics is much closer to the real car, and they only simcade not trying to get perfect real car. I took this post to this one because I don't want to deal with any deluded fantisists who thinks there are no problems with this game. Every time that oversteer bug comes up the tell you to git gud. Well the guy in the video caught the oversteer bug not realizing this is the exact problem im talking about.

You should put your keyboard away. I will come out and say this straight forward. I highly doubt you have driven a 60's muscle car. I own a 65 Impala. Its in my picture profile.I've owned it for 27 years. I have a 71 Chevelle that I've owned for 25 years. Having a car from that era that is pretty much stock, other than upgraded tires,headers and dual exhaust, I can tell you without a doubt, your dead wrong. I have driven a 65 Mustang, same damn car. I owned a 67. I've owned and driven more American Muscle cars than your age. I grew up on these cars. You have no clue what your talking about.

 

[Scaff]

This guy wants to say the Ford is perfectly fine to drive. Project2 cars have some problems with physics on various cars. I think it is fine to compare games to other games physics but when you stand an say it is perfectly like the real car and knowing it is not, I take problems with that. For most of the video he was driving the mustang way below the limits to look good but he still could not avoid that oversteer bug. At 2:25 the car does the strange oversteer but looking at his steering wheel input the car should have made the turn. This is not how the real car drives this is an insult to Ford if you claim it is. I think they try to make this game physics look perfect by ignoring the flaws. Take forza muscle cars and only using the controller you will see the physics is much closer to the real car, and they only simcade not trying to get perfect real car. I took this post to this one because I don't want to deal with any deluded fantisists who thinks there are no problems with this game. Every time that oversteer bug comes up the tell you to git gud. Well the guy in the video caught the oversteer bug not realizing this is the exact problem im talking about.

Nothing at all strange about the oversteer at 2:25, you are aware that the rear will step out on a car regardless of the steering angle applied if the rear yaw angle exceeds the front yaw angle?

However I have now posted my video up, I await yours, as this is now the third car you have claimed is undrivable withouit provifindg a shread of evidence from yourself.

I do also notice that you seem to have replied in a different thread, which I can only think was to try and avoid feedback from a member who you know full well has a massive amount of experience opf how these cars should behave in reality. Now that clearly didn;t work, but if @killerjimbag says its wrong, then his word I will take as evidence.

Again though, I wasn't recording. Must get into this habit!

Your playing on the PS4 I believe.

In which case you don't need to record, your PS4 is doing that for you all the time. Just get to the end of the race and hit the share button, then select capture video. Bingo, up to an hour of footage from the title you are in captured and read for editing or sharing.