Myth: Gran Turismo Sport and thus GT7 have unrealistic Front Wheel Drive Understeer. Status: BUSTED!

[HugoTwoWheels]

I don't think a motorcycle can be compared to a car in this instance because a small amount of slippage is beneficial to turning a car.

It's true for motorcycles as wheel. Rubber, as far as i understand the topic, has maximum grip when there's slight slippage. However, on street tyres, that beneficial slip range is small, so you either don't slip or skid off track.
If you watch MotoGP nowadays many of them slip slightly both in and out of corners. They do this into corners in supermoto as well. Good luck though on street tyres as a regular human.

 

[Scaff]

I don't think a motorcycle can be compared to a car in this instance because a small amount of slippage is beneficial to turning a car.

The exact same is true of motorbikes, they still run rubber tyres, they still have a contact patch, they still have to deal with slip angles and self-aligning torque.

You are to be blunt, now just making stuff up that you clearly do not understand. Actually putting your hands up and acknowledging that, rather than digging an even bigger hole would be the smart thing to do at this point.

 

[Magog]

Please describe what understeer and oversteer are in terms of actual vehicle dynamics, as you're entering into the realms of nonsense now.


That has bugger all to do with under or oversteer at all, but I once again challenge you to prove me wrong. Describe in detail, referencing standard texts such as Milikin and Miliken exactly why what you say is accurate (it's not, but this should be fun).

Here. I copied and pasted this for you.

"Oversteer is when the momentum of a car breaks its back tires’ traction and is visibly seen when the rear of your car rotates toward the front of your car. The most illustrative example of this is drifting, which is when a driver initiates oversteer to drift/slide the car around the apex of a corner and continue moving forward.

Oversteer is more common in rear-wheel-drive vehicles, though it can occur and be initiated in front-wheel and all-wheel-drive vehicles as well.

Understeer is when you turn the steering wheel too sharply and, due to conditions or too much speed, the car continues forward in a straight line.

Understeer is more common in front-wheel-drive vehicles, though it can occur in rear-wheel and all-wheel-drive vehicles, too."

Which is exactly what I've been saying and I'm not really disagreeing with any of your points. I'm sure they do design rear wheel drive cars to compensate for their natural inclination to oversteer and front wheel drive cars for their natural inclination to understeer but all other things being equal those are their natural tendencies. I'm not sure how anyone can disagree with that. A lot of the cars in GT are old over overpowered and aren't going to be able to compensate for those tendencies.

 

[Scaff]

Here. I copied and pasted this for you.

"Oversteer is when the momentum of a car breaks its back tires’ traction and is visibly seen when the rear of your car rotates toward the front of your car. The most illustrative example of this is drifting, which is when a driver initiates oversteer to drift/slide the car around the apex of a corner and continue moving forward.

Oversteer is more common in rear-wheel-drive vehicles, though it can occur and be initiated in front-wheel and all-wheel-drive vehicles as well.

Understeer is when you turn the steering wheel too sharply and, due to conditions or too much speed, the car continues forward in a straight line.

Understeer is more common in front-wheel-drive vehicles, though it can occur in rear-wheel and all-wheel-drive vehicles, too."

Which is exactly what I've been saying and I'm not really disagreeing with any of your points. I'm sure they do design rear wheel drive cars to compensate for their natural inclination to oversteer and front wheel drive cars for their natural inclination to understeer but all other things being equal those are their natural tendencies. I'm not sure how anyone can disagree with that. A lot of the cars in GT are old over overpowered and aren't going to be able to compensate for those tendencies.

I didn't want you to copy and paste it (from an un referenced source), I want you to explain it, in your own words, refereeing cited sources.

 

[Famine]

Here. I copied and pasted this for you.

From where? Don't just steal other people's words.

No part of it addresses anything you've been chucking out there, even the stuff you've moved onto after abandoning the earlier stuff and pretending it didn't happen.

I'm sure they do design rear wheel drive cars to compensate for their natural inclination to oversteer and front wheel drive cars for their natural inclination to understeer but all other things being equal those are their natural tendencies. I'm not sure how anyone can disagree with that.

Your own stolen words don't agree with that. They say oversteer is more common in RWD and understeer more common in FWD, but that both can occur in FWD, RWD, and AWD. They say nothing about "natural inclination", and for good reason.

Road cars are - with very few, highly specialist, low volume exceptions intended for non-road use but with road legality - designed to understeer first.


Apropos of the thread title, the only "myth" that's "busted" is that you have any idea what you're talking about, though no-one's floated that myth because it's kind of apparent.

 

[Silver Arrows]

Again, I remind people that this thread was spun off because OP was getting worried that his defending of GTS' physics, and how they would apply to GT7, was rightfully getting lambasted, and so he decided to move the battlefield to somewhere where he could feel safer.

And even then, he's still getting absolutely punked by people who actually have some sort of idea on vehicle dynamics, and are still correct in GTS, and likely GT7's physics with regards to FF vehicle dynamics being nowhere near reality.

Absolutely cucked behavior.

 

[Igano]

@Magog
Motorcycles (RWD of course) also always understeer on acceleration, on asphalt, unless you accelerate beyond available grip and start drifting (IRL on street that would be crashing).
Motorcycles are an example of a simplified car in this discussion, I'd say.

Supermotard actually turn often by oversteering.

 

[Dave A]

Supermotard actually turn often by oversteering.

Do they run on asphalt? I thought they tended to run on dirt and gravel. Looser surfaces.

 

[Igano]

Do they run on asphalt?

Both asphalt and gravel!

 

[Wolfe]

^ Supermotard asphalt drifting is one of those things you witness for the first time and think, "wait, that works?"

It's unclear whether it's the actual physical weight transfer that's wrong, or that the rubber compounds don't alter grip appropriately according to the weight on them. I'm dubious about it being either of those though, because the first is basic high school physics and the second falls directly out of even the most basic tyre model. If Polyphony haven't got either of those right in 25-odd years then there's no hope.

Personally if I had to pick one thing, my guess would be the dynamic contact patch simulation. I'm not sure it exists. The contact patch not changing size in response to weight transfer would go a long way to explaining the inability to get cars to move around as you load up the front or rear tyres. This affects all cars, but it affects FF cars most because they rely entirely on weight transfer to modulate the available grip at the rear tyres. FR and MR cars to some extent can get away with using power to slip the tyres, although this would go some way to explaining some of the odd behaviours that MR cars have as well.

This would be my guess simply because tyre pressures aren't available as a modifiable option for the player. That was weird in 2010, and it gets weirder every game it's still not there. It feels like a big red flag for a simulation. If the tyre pressure was in the simulation as a variable at any level, then it would be relatively trivial to expose it to the player for tuning. But if it's not in the simulation at all, then potentially a bunch of things that would also rely on tyre pressure aren't dynamically modelled either, like a dynamic contact patch.

It's not beyond belief that these things are hard-coded as constants. It would just be very, very stupid.

If GT still has weight transfer wrong to such a degree, judging from all the comments*, I think the contact patch is an inadequate explanation by itself. Weight transfer effects can still occur with a fixed contact patch area, just from variable friction -- mu times weight. A dynamic contact patch is a good detail, but FWDs should already oversteer when provoked to do so before you have to add that.

Here's how weight transfer effects should manifest: if you apply tire forces at an offset from the center of gravity, your physics engine should produce torque on the chassis. Like you said, @Imari, high school physics. With this torque, the car should roll, squat, and dive naturally, restrained by the reaction of the suspension. The force value of that reaction from the suspension is the weight on the tire, by Newton's third law.

It could be...

• Tire force vectors are not applied at the contact patch, beneath the center of gravity
• The center of gravity is artificially low, a band-aid fix for improving behavior
• The chassis is not inherently subject to torque from the tire physics, and roll/squat/dive is a hard-coded effect

As it happens, a variable contact patch producing reasonable values does not need to be terribly complex. From a conversation on F1technical.net, I believe making it a function of weight load to the power of ~1/3 gets you in the ballpark. Put the pressure variable in a divisor, and lower pressure will produce more variability, and vice versa.

* -- I had a few races in GT Sport's public beta; it felt familiar enough for me to have an idea of what's up.

 

[HugoTwoWheels]

And technique is different on loose surfaces, where you always slide more or less.

Of course you're not sliding around in turns when riding casually on asphalt on street tyres...

 

[HugoTwoWheels]

Have you tried this week's race A? I can feel and see the weight transfer perfectly. You can also pick a view where you can watch the suspension springs compress and decompress when weight transfer occurs, which correlates perfectly to what I feel in my wheel.

I'm only playing race C however, which also handles and feels excellent to me.

 

[ScottPuss20]

I think the F1 games explain oversteer and understeer best.

"Oversteer occurs when the rear wheels lose grip before the fronts"

"Understeer occurs when the front wheels lose grip before the rears"

This principle is a constant no matter what car you are driving. If you're within the grip limits and don't exceed the maximum slip angle, you will go around the corner just fine.

 

[05XR8]

Well, PD really will have to make changes because of the wheel offset and tyre width options.

In GT Sport, there's no difference after selecting a plus 1 and plus 2 wheel diameter, in Livery Editor. I think Forza do a good thing by at least showing wheel weight increase/decrease depending on the wheel brand/size and(I think) construction(one-piece, two-piece, three-piece).

This would be good to test.

 

[synon]

If GT still has weight transfer wrong to such a degree, judging from all the comments*, I think the contact patch is an inadequate explanation by itself. Weight transfer effects can still occur with a fixed contact patch area, just from variable friction -- mu times weight. A dynamic contact patch is a good detail, but FWDs should already oversteer when provoked to do so before you have to add that.

I agree, eg AFAIK Assetto Corsa models the contact patch as a single point.

It could be that PD models tires with low load sensitivity.

 

[Grand Prix]

^ Supermotard asphalt drifting is one of those things you witness for the first time and think, "wait, that works?"

It used to be a thing in cars as well:

"At the first bend, I had the clear sensation that Tazio had taken it badly and that we would end up in the ditch; I felt myself stiffen as I waited for the crunch. Instead, we found ourselves on the next straight with the car in a perfect position. I looked at him, his rugged face was calm, just as it always was, and certainly not the face of someone who had just escaped a hair-raising spin. I had the same sensation at the second bend. By the fourth or fifth bend I began to understand; in the meantime, I had noticed that through the entire bend Tazio did not lift his foot from the accelerator, and that, in fact, it was flat on the floor. As bend followed bend, I discovered his secret. Nuvolari entered the bend somewhat earlier than my driver's instinct would have told me to. But he went into the bend in an unusual way: with one movement he aimed the nose of the car at the inside edge, just where the curve itself started. His foot was flat down, and he had obviously changed down to the right gear before going through this fearsome rigmarole. In this way he put the car into a four-wheel drift, making the most of the thrust of the centrifugal force and keeping it on the road with the traction of the driving wheels. Throughout the bend the car shaved the inside edge, and when the bend turned into the straight the car was in the normal position for accelerating down it, with no need for any corrections." - Enzo Ferrari

Some footage from 1938:

"The Mercedes respond well to a new technique of using the throttle to steer them around bends."

Here is an example of Fangio using the technique at the 1957 French Grand Prix:

And here is an example of a modern driver using the 4 wheel drift in a Bugatti Type 35:



In addition to being able to adjust tire pressures, we also need the option of equipping period-specific racing tires - some cross-ply constructed tires available to use, as opposed to just lower grip versions of modern radial tires. Although I'm not actually sure what PD's "Comfort Tires" are supposed to represent, are they supposed to be cheap radial or cheap cross-ply? The entire tire model needs an overhaul anyway.

 

[phillgt2002]

Although I'm not actually sure what's PD's "Comfort Tires" are supposed to represent, are they supposed to be cheap radial or cheap cross-ply? The entire tire model needs an overhaul anyway.

Comfort stands for comfy…. Duh.

They should add Comfort+ tires for extra comfyness.

 

[05XR8]

Comforts are All season high mileage low road noise low grip eco treads… but not. It’s all about degrees of friction with PD. They can’t even do that correctly. Hopefully, they show improvement for GT7.

The feel of friction, tyre rubbing into a corner and literally hugging the road, isn’t there. We get the feeling of grip by controlling the car, but not that sensation of the friction. Like a granular disconnect.

 

[Grand Prix]

Comforts are All season high mileage low road noise low grip eco treads… but not. It’s all about degrees of friction with PD. They can’t even do that correctly. Hopefully, they show improvement for GT7.

The feel of friction, tyre rubbing into a corner and literally hugging the road, isn’t there. We get the feeling of grip by controlling the car, but not that sensation of the friction. Like a granular disconnect.

Yes, that is my suspicion as well. However, there are some instances where they will provide comfort tires as default for vintage race cars. What are these tires supposed to be? Are they meant to approximate the amount of grip the race tires would have had back in the day? Or are they meant to be modern road tires fitted to an older race car? Yes, they are probably just plain economy radials, but what are they supposed to be?

When they supply other vintage race cars with "Racing Hards" as default, those tires are clearly not the vintage racing tires, since they behave similarly to Racing Hards on other, modern cars in the game(s). Which has some interesting implications behind it.

For one, if you fit tires with much more grip than the car was designed to be used with in real life, you run the risk of warping or even breaking the chassis. This is partly why tire companies like Avon supply period-accurate tires; classic race cars can use tires like these safely for historic events, and they also simulate the grip of course. One could reasonably assume that PD doesn't care for this detail very much.

Secondly, it would imply that PD doesn't care all that much about recreating the grip these cars would have had in-period, which many of us would like to experience. Assetto Corsa on the other hand, provides several period-accurate options for older cars.

Bringing it back on topic: If they start caring enough about different tire construction methods, tire pressures, contact patches, deformation, etc., it just might encourage them to overhaul the tire model, among other things, resulting in less of that notorious, overarching understeer that has been present since at least GT4. Perhaps I am naive for thinking that.

 

[Scaff]

I think the F1 games explain oversteer and understeer best.

"Oversteer occurs when the rear wheels lose grip before the fronts"

"Understeer occurs when the front wheels lose grip before the rears"

This principle is a constant no matter what car you are driving. If you're within the grip limits and don't exceed the maximum slip angle, you will go around the corner just fine.

It's not a bad way of describing it, it just falls short when you dig into detail.

I prefer the approach that Millikin and Millikin and others take, which requires you to look not just at the balance of the whole car, but also the balance at each end and at each corner. You do this by defining understeer as any point at which the yaw rate at the front of the car is greater than the yaw rate at the rear of the car, and oversteer as being when the yaw rate at the rear of the car is greater than the yaw rate at the front of the car.

This definition has a number of advantages, the first being that it allows you to look at the natural inclination of the car before under or oversteer has occurred, as the yaw-rate doesn't need to be at a level that has exceeded the slip angle of the tyre, so its the cars 'inclination' towards under or oversteer for a situation 'before' you have even lost grip at either end.

Secondly it makes understanding how a car (as I and others have described above can transition from under to oversteer) much more straightforward. As initially the front yaw angle can be higher than the rear, with the front exceeding the tyres slip angle, resulting in understeer, however if you have sufficient torque (in a RWD cars) you can use it to generate enough rear yaw to exceed the rear tyres slip angle, but also to generate more rear yaw than front yaw. The result (if done correctly) is a nice transition from understeer to oversteer.

This definition also has the advantage of explaining four-wheel drifts (as described above), which the standard definitions of under and oversteer doesn't really work well for, as using yaw rates you have basically managed to exceed the slip angle just enough to be generating maximum grip on all four corners, while playing with the transition between under and oversteer using the throttle by balancing the front and rear yaw balance (which was much easier to do on old, skinny cross-ply's as they have a much wider window of slip-angle to play with).

This is a great book to look at it as a concept (as its around a 1/4 o fthe price of M&M's Race Car Dynamics and much, much more approachable as it doesn't dive into the heavy side of the physics).

 

[HugoTwoWheels]

This short video is relevant. How to drive a FWD car, explained by a professional rally driver. There are two other videos, where RWD and 4WD is explained in the same manner. And yes, he's driving on tarmac, not gravel.

 

[Scaff]

This short video is relevant. How to drive a FWD car, explained by a professional rally driver. There are two other videos, where RWD and 4WD is explained in the same manner. And yes, he's driving on tarmac, not gravel.

Great video (I've seen similar ones from Tiff Needell), perfectly in line with how most of us have been talking about FWD cars with LSDs and (surprise, surprise) nothing at all like FWD cars in GTS.

 

[phillgt2002]

Great video (I've seen similar ones from Tiff Needell), perfectly in line with how most of us have been talking about FWD cars with LSDs and (surprise, surprise) nothing at all like FWD cars in GTS.

I found FWD IRL and in sims hard to drive fast.

All one would have to do is watch BTCC - those cars are sideways most of the time.

 

[Scaff]

I found FWD IRL and in sims hard to drive fast.

All one would have to do is watch BTCC - those cars are sideways most of the time.

Have you ever driven one with an LSD? Makes a massive difference.

 

[Wolfe]

@Grand Prix -- The truth is that PD doesn't need a highly-sophisticated tire model to evoke differences in tire construction. Simpler tire formulas can and do offer variables besides just a friction 'μ' value. There is a "stiffness" variable in common Pacejka and Brush models, which affects the grip-slip transition. A higher stiffness value is more like modern racing slicks. If you also calculate a dynamic contact patch using a sidewall "rigidity" value, you can differentiate tire compounds like so:

• "Comfort" tire: Low stiffness, high rigidity
• "Race" tire: High stiffness, low rigidity
• "Vintage" tire: Low stiffness, low rigidity

I would also vary the per-surface friction between the three, so if a Comfort tire has decent tread while the Vintage tire is a racing tire, the Comfort tire should provide more traction on some surfaces -- helping differentiate the two some more.

 

[05XR8]

@Grand Prix -- The truth is that PD doesn't need a highly-sophisticated tire model to evoke differences in tire construction. Simpler tire formulas can and do offer variables besides just a friction 'μ' value. There is a "stiffness" variable in common Pacejka and Brush models, which affects the grip-slip transition. A higher stiffness value is more like modern racing slicks. If you also calculate a dynamic contact patch using a sidewall "rigidity" value, you can differentiate tire compounds like so:

• "Comfort" tire: Low stiffness, high rigidity
• "Race" tire: High stiffness, low rigidity
• "Vintage" tire: Low stiffness, low rigidity

I would also vary the per-surface friction between the three, so if a Comfort tire has decent tread while the Vintage tire is a racing tire, the Comfort tire should provide more traction on some surfaces -- helping differentiate the two some more.

Indeed. That’s why I’ve been hoping the track surface data, PD gathered, comes into play.

 

[ScottPuss20]

This definition has a number of advantages, the first being that it allows you to look at the natural inclination of the car before under or oversteer has occurred, as the yaw-rate doesn't need to be at a level that has exceeded the slip angle of the tyre, so its the cars 'inclination' towards under or oversteer for a situation 'before' you have even lost grip at either end.

Secondly it makes understanding how a car (as I and others have described above can transition from under to oversteer) much more straightforward. As initially the front yaw angle can be higher than the rear, with the front exceeding the tyres slip angle, resulting in understeer, however if you have sufficient torque (in a RWD cars) you can use it to generate enough rear yaw to exceed the rear tyres slip angle, but also to generate more rear yaw than front yaw. The result (if done correctly) is a nice transition from understeer to oversteer.

Aha! This is why Assetto Corsa and other simulators are better than GT!`You can use the throttle to steer the car and overcome understeer! Do it in GT Sport and you just get more understeer. If GT7 solves this issue, I'll be very happy.

 

[CBH]

Aha! This is why Assetto Corsa and other simulators are better than GT!`You can use the throttle to steer the car and overcome understeer! Do it in GT Sport and you just get more understeer. If GT7 solves this issue, I'll be very happy.

In 2018 PD did change the tire physics by an update because before the update the RWD cars did not oversteer much, but after the update they oversteered to much. So with the tire physics change it may have affected the FWD cars with to much understeer.

Is this right mate you mention awhile ago about the cars handling and something about the tires ?.

 

[Eunos_Cosmo]

Aha! This is why Assetto Corsa and other simulators are better than GT!`You can use the throttle to steer the car and overcome understeer! Do it in GT Sport and you just get more understeer. If GT7 solves this issue, I'll be very happy.

I would be really happy if you can drive an (older) 911 properly in GT7:

Throttle On: Understeer
Throttle Off: Oversteer

The only game I've played that seems to get this right is AC. Forza is just oversteer all the time. Throttle on, throttle off, steady state, whatever - it's like they just coated the rear tires in bacon grease. I do remember 911s in GT4 understeering quite a bit, but that game's handling model was pretty poor IMO. I am trying to recollect RR handling in GT5 and I'm not remembering anything. Probably because there were no premium Porsche/Ruf models so I didn't bother to try any of them.

 

[Scaff]

I am trying to recollect RR handling in GT5 and I'm not remembering anything. Probably because there were no premium Porsche/Ruf models so I didn't bother to try any of them.

It was all oversteer, even the low power / torque Alpine's.