Flaws of the tyre model simulation

[Lord Protector]

Hi there. This thread is a result of GTS' horrible tyre model simulation. This is a simulation of how the tyres gain and lose grip (not to be confused with actual driving of a car). GT Sport's tyre simulation is the worst simulation of all current gen simulators and simcades. PD has improved his a little throughout the updates but it's still shameful.

What the tyre model affects: acceleration, deceleration, understeer, oversteer.
What it does not affect: basically any kind of driving until tyres start to lose grip.

Acceleration: this is the easiest way to test the tyre model via 0-100 kph test (0-60mph), which in GTS results in totally unrealistic times. FF cars are most affected, followed by FR cars, MR and RR cars and 4WD cars respectively. 4WD cars are less affected but still not realistic.

The easiest way to test the tyre model is 0-100 kph test (0-60mph), which in GTS results in totally unrealistic times. FF cars are most affected, following by FR cars, MR and RR cars and 4WD cars are naturally not so far from reality.

@Scaff also did a pretty good video comparing real life, GTS and other games' tyre model. Please post the video here, I could not find it anywhere.

Examples of some 0-100kph tests: (I used TCS1 because that gives the fastest results, all cars stock, with stock tyres equipped)


Dodge Charger SRT Hellcat:
Real life time: ~3.4 seconds
GT Sport time: ~5.8 seconds - ridiculously slow



Honda Civic Type-R:
Real life time: ~5.8 seconds
GT Sport time: ~7.2 seconds - again, totally off



BMW M4 Coupe:
Real life time: ~4.0 seconds
GT Sport time: ~5.5 seconds - once again, way off


Conclusion: tyre model in GT Sport is ridiculously wrong. Generally speaking the more powerful a car is the bigger is the gap between reality and GT Sport. Accelerating in GT Sport feels like the car drives on ice rather than tarmac.

Other weird occurrences are when not using CSA (counter steering assist) in a rear wheel drive car. Cars tend to drive just fine until the grip is lost completely and the car becomes a missile heading anywhere but not where you think. It's impossible to predict where the car is heading, how much you need to counter-steer, how much throttle can be applied. There is very little to no feedback (when using a wheel). When using 4WD cars, they seem to drift like crazy, they go sideways even under little throttle, again, no grip just spinning like and since all wheels are driven the car seems to go sideways rather than go straight. All tyre compounds are affected, including racing slick tyres.


Another examples:



Stock car, stock tyres (racing slick tyres) at 0:28 take a look how the wheels spin in 3rd gear like crazy, the car won't stop for a few seconds. Impossible in real life - the car has nowhere near as much power to outgrip the slick tyres - and even going backwards.






a 900+hp Supra, wheels won't stop spinning even at 300kph




https://www.youtube.com/watch?v=Aseebqcd5Ro

250 hp Impreza, no grip again, look at 0:53 especially.



https://youtu.be/IrsS-a9pAmw

Here's the Scaff's video.


Please keep the discussion on topic, this is a thread about tyres, not about driving physics.

 

[TheGeologist]

I personally have not experienced what you are referring to, especially when you say predictably issues with a rear wheel drive car when CSA is off. I've never used CSA, and I've never had any issue like that.

I have played extensively with both a DS4 and a wheel. Dr B, Sr S

 

[Lord Protector]

I personally have not experienced what you are referring to, especially when you say predictably issues with a rear wheel drive car when CSA is off. I've never used CSA, and I've never had any issue like that.

I have played extensively with both a DS4 and a wheel. Dr B, Sr S

It's very unpredictable compared to real life or assetto corsa for example.






Here's the video from Scaff, very helpful to understands to problem.

 

[Highlandor]

It's not the tyres, it's whatever the equation / calculation / algorythm PD are using to give a track (and it's weather) their grip levels.

If you drive the same car repeatedly (private online lobby), but only change the track and weather (keep every other variable the same), sooner or later you'll see that sometimes you get way too much grip (and the car feels like it has super glue for tyres), other times you get understeer and wheelspin from a lack of grip.

This takes into account tyre wear and temp btw.

Try it with something like the Lambo LP400 (sports mediums - all driving aids off) on 3 different tracks like suzuka, Nurb gp and Brands gp, cycling through the different times of the day. Keep repeating this week after week and you'll soon see the difference.

 

[polizei]

I think if we all just agree this is much more a game, than a simulator, these things will matter less and less.

 

[azrilnazli]

netcode matters...simple calculation for better netcodes.

imagine the computing power that can host 60 online cars and each of them have very complex tire model

 

[MatskiMonk]

Here's the video from Scaff, very helpful to understands to problem.

Just noticing that where @Scaff says the there's no rev drop in GTS, it worth noting that the ST in his video bangs on the limiter between about 6100 and 7000 rpm. At the point the car launches the revs do drop to about 4500rpm, but then climb back up much faster and it's back on the limiter again within an instant. It's ability to do this, is very much linked to factors other than simply the tyres.

edit:
For what it's worth, my take on this is simple;

Assumption: The game has values for both static and dynamic coefficients of friction (µs and µd respectively), for each type of tyre.

Therefore, if there is an issue with the physics in a dynamic friction situation (i.e. the tyre is sliding), the most obvious explanation is that the values for µd are too low. Simple as that.

 

[Scaff]

Just noticing that where @Scaff says the there's no rev drop in GTS, it worth noting that the ST in his video bangs on the limiter between about 6100 and 7000 rpm. At the point the car launches the revs do drop to about 4500rpm, but then climb back up much faster and it's back on the limiter again within an instant. It's ability to do this, is very much linked to factors other than simply the tyres.

When I say no rev drip, I do mean that it doesn't drop and then recover in the way it should, rather it drops and bounces straight back up and off the limiter.

Which is why I describe it as a bounce rather than a drop.

 

[Amaru]

The tyres in GTS aren't as we knew them in previous titles.

Comfort Hard for example, in GT6 felt almost identical to GT5's CH tyre (& vice versa for pretty much every tyre in GT5 & 6).

GTS is quite the opposite, none of the tyres available in GTS feel like they felt in previous titles. I've been convinced of this since the closed beta after about a week of gameplay.

I'm glad to see @Lord Protector 's findings verify what I have been experiencing.

 

[Knukel]

I remember noticing huge difference to GT5 when first time doing the Clio tutorial run around Brands Hatch. And tyres felt so lifeless when went back to GT5 to test shuffle. (Some gave up with GT6 because of too 'snappy' to control on pad or drifting was too hard & and i think this is related to tyre model). I compare GT Sport more to GT5 than GT6.

IIRC, in Gran turismo 2 TVR Speed 12 on economy tyres couldn't accelerate past 300kph because of endless wheelspin

 

[Lofty L]

I think if we all just agree this is much more a game, than a simulator, these things will matter less and less.

Yeah but they spend all this time in making this game and it is nowhere near to what they claim it to be my brother has a Mclaren 650s and it’s far easier to drive in real life than in this game

 

[TS050]

my brother has a Mclaren 650s and it’s far easier to drive in real life than in this game

we gon need some proof

 

[Lofty L]

we gon need some proof

Of what

Proof of what

 

[Lord Protector]

Any high power rear wheel drive car is easier to drive irl than in GTS.

 

[Lofty L]

Any high power rear wheel drive car is easier to drive irl than in GTS.

What you need to remember is a Ford Focus does not run the same tyres as a Mclaren 650s. In real world but we have sport soft tyres that apparently does for everything in GT Sport

 

[TS050]

Of what

You can’t just claim that your brother owns a 650S and expect everyone to believe you without proof

 

[Lofty L]

You can’t just claim that your brother owns a 650S and expect everyone to believe you without proof

I don’t care what people believe we have over a million pounds worth of Ducati motorcycles it’s got nothing to do with anybody else you wanted an opinion you got one end of what we have is irrelevant

 

[Firehuntah]

I think overall it's not too bad but there are some cases where I do think there's not enough grip. I especially notice this with certain Gr3 cars in low speed corners. Even with low throttle input you often get oversteer on slicks. This doesn't really seem realistic. Now I've not driven in a real racecar before so it's hard to compare but it doesn't seem right.

FWD cars did get better after one of the updates though, feels they have a more realistic amount of grip in corners. Same for road cars, I haven't really noticed anything wrong. But looking at your videos and especially the 0-100 km/h times there's definately something wrong. Not as bad as the Tesla Roadster in GT5/6 (that was REALLY off ) but the difference in time is still too much. Don't know if you tested this but do the times at least compare when you use racing softs? Then it's definately a tyre/grip issue.

 

[SlipZtrEm]

If the 911 GT3 RS was as unruly in the real world as it is in GT Sport, I probably would've been a dead man back in October.

GT's had an issue with acceleration times — especially those of front-drive cars — since around GT3. Low-speed physics seem to be tough to nail for most of the sims out there. I'd guess it's less of a concern to devs since most players won't be exploring that part of the speedometer as much as the others, but it's important nonetheless.

 

[Voodoovaj]

Any high power rear wheel drive car is easier to drive irl than in GTS.

To be fair, we all asked for this.

During the reign of both GT5 and GT6, the cry was all about how unrealistic the tire grip was. Going back to GT6 recently, Sport Hard in GT6 is surprisingly sticky and (almost) akin to the GTS racing hard.

In GT6, the racing cars were all much better, and the racing in general was better, on the sports tires than on their racing tires. It would seem to me that PD did exactly what we asked for and gave us less grip.

 

[PirateAlan8]

Tyre model what? But I sure am enjoying playing this game!

 

[Lofty L]

If the 911 GT3 RS was as unruly in the real world as it is in GT Sport, I probably would've been a dead man back in October.

GT's had an issue with acceleration times — especially those of front-drive cars — since around GT3. Low-speed physics seem to be tough to nail for most of the sims out there. I'd guess it's less of a concern to devs since most players won't be exploring that part of the speedometer as much as the others, but it's important nonetheless.

Thank you all these people moaning and when was the last time they drove a Veyron or am F40 or Porsche GT3 so what are they comparing it to

 

[LennyB]

Low-speed physics seem to be tough to nail for most of the sims out there.

This is the exact problem, years ago I read an article from a game programmer type guy who explained it all in great detail. It was near the end of gt3's life. It's hard to believe after all this time a solution has not been found.

I wish I could find that information again because I'm not going to waste anyone's time trying to fumble around with my understanding of it.

 

[Shottah072]

Yeah but they spend all this time in making this game and it is nowhere near to what they claim it to be my brother has a Mclaren 650s and it’s far easier to drive in real life than in this game

Not saying GTS physics are great but i doubt your brother has ever driven his 650S at the limit like we do in GTS. Of course its not hard to cruise your supercar at normal speed, but push it to the limit and it wont be that easy to handle anymore.

Wish i could affort a 650S though, nice car !

 

[polizei]

netcode matters...simple calculation for better netcodes.

imagine the computing power that can host 60 online cars and each of them have very complex tire model

I don't develop games, so if im wrong, sorry not sorry, but wouldn't that calculation be made on the client side not the server? I agree, netcode is important, but i would think the clients are all just telling the server where they are and what they are doing and the server is simply telling all the clients where the other clients are?

 

[sabamber]

I owned a TVR T350c and M3 CSL recently and can confirm its definitely harder to drive a quick car properly in real life than GT.

 

[Voodoovaj]

What the tyre model affects: acceleration, deceleration, understeer, oversteer.
What it does not affect: basically any kind of driving until tyres start to lose grip.

Not to be a fly in the ointment, but the grip modifier within a tire model affects the items listed, not the tire model itself.

For a tire model you can think of this like the carcass of the tire without rubber on it. The tire model would be how the tire behaves under load: Lateral load, gravitational load, weigh transfer, does the tire diameter grow with speed, how does the tire react to compression and rebound from bumps in the road, how does the tire flex differently when these load occur in a not rotating versus rotational condition. Stuff like that. That's a tire model. Imagine a tire in the general sense and think of all the things it does and that is what you are modelling.

What you listed are elements of grip, or IRL terms, the rubber compound.

 

[Nando deBem]

@Lofty L, I think people just wanted to see your brother’s beautiful car. Nobody really cares if you’re rich or not.

 

[tofuhero]

I don't develop games, so if im wrong, sorry not sorry, but wouldn't that calculation be made on the client side not the server? I agree, netcode is important, but i would think the clients are all just telling the server where they are and what they are doing and the server is simply telling all the clients where the other clients are?

The physics calculations are definitely done on the PS4; otherwise offline mode wouldn't work.

 

[GordonS]

I think overall it's not too bad but there are some cases where I do think there's not enough grip. I especially notice this with certain Gr3 cars in low speed corners. Even with low throttle input you often get oversteer on slicks. This doesn't really seem realistic. Now I've not driven in a real racecar before so it's hard to compare but it doesn't seem right.

FWD cars did get better after one of the updates though, feels they have a more realistic amount of grip in corners. Same for road cars, I haven't really noticed anything wrong. But looking at your videos and especially the 0-100 km/h times there's definately something wrong. Not as bad as the Tesla Roadster in GT5/6 (that was REALLY off ) but the difference in time is still too much. Don't know if you tested this but do the times at least compare when you use racing softs? Then it's definately a tyre/grip issue.

I've tested launching LaFerrari vs. Veyron TCS on with RS tyres and the drag times are quite life like that way.
The footage is here:

But yes, more normal tyres should be more grippy in launch situations.