How accurate is Assetto Corsa?

[Bruno Iwanczuk]

If i'm 5 seconds slower than a real life F1 lap world record. Am i really 5 seconds slower?

 

[Mitja Bonca]

you can't attack kerbs in ac like you can irl, the kerbs (inside kerbs especially) are like ice in ac

Eh, not true at all!

 

[Ryno917]

You mean Sam Collins? He is the main commentator for NismoTV. I know him, I talked to him, but he said he doesn`t know anything in specific, that Japanese are hiding all the data. Thats what he says, go figure whats true.

Yep, he's the one. Ah well I tried

 

[Denis Betty]

Eh, not true at all!

Yes true.
Kerbs on apex in AC are (for the most part) totally a no - go zone in AC. Unlike the other sims I use.

 

[Ryno917]

Yes true.
Kerbs on apex in AC are (for the most part) totally a no - go zone in AC. Unlike the other sims I use.

I regularly use (and sometimes abuse) the curbs in most cars, on most tracks without issue.

 

[Mitja Bonca]

Yes true.
Kerbs on apex in AC are (for the most part) totally a no - go zone in AC. Unlike the other sims I use.

Have you thought of the other way around, that the cars are the problem, and not the tracks?
I found out exactly the same a week ago, when I was doing physics for new DTM 2018 mod of URD. On the end, it was way better (even driving on the inside curbs), still not perfect, but ok, driveable. So that means the tracks are perfectly fine (I know that for a fact, I create them), but the cars are lacing of same setting(s) in their physics. Maybe its Kunos engine`s fault, who would know.
The Nissan GT-R GT500 car is even better in terms of physics, which we were making for the last 2 weeks, and will go public in a day or two - really recommend that you guys try it out. It`s as close as it can get sim to real life.

 

[mclarenf1papa]

It`s as close as it can get sim to real life.

If you made it without aero data (and with Kunos-style load sensitivity), I can assure you that that's not the case....

 

[Ghoults]

Yes true.
Kerbs on apex in AC are (for the most part) totally a no - go zone in AC. Unlike the other sims I use.

Can you post some examples of no go zone kerbs in ac (car/track/corner combos)? My experience is that in ac the main effect from kerbs is the traction loss because kerbs have lower friction coefficient. The shape doesn't matter that much.

I have not driven mugello for a while but I assume the issue where cars get stuck on kerbs and spun around was fixed long time ago? The issue which could be massaged away by setting damping of the tire to 500 or so?

 

[mantasisg]

About losing your rear: That's difficult to tell. Watching F1 season 2009 a lot of drivers spun!
Also look at this (didn't find anything better that quick):
The difference is that in real life you start to feel the slide at about 0.01° of drift, while with the average simracing setup, you start to feel a slide at about 2° of drift

Where do you get this nonsense about 0.01° and 2° ? These are super low angles, you wouldn't be able to tell if it is not yet UNDER optimum slip angles at these angles. Also don't you feel weird putting on Schumacher spin next to your statement ? Worth to mention that it is important how quickly the dropoff happens, you might never have the chance to catch the slide in some cases.

I am pretty sure that "seat feel" from real car during oversteer is directly tied with the reduction of g-force as you start loosing your radius. It should be that the closer to rear wheels you are, the more seat feel you should have, as well as visual cues. Another thing is that the lower the g-forces are, the more seat feel you should have.

For example: constant 40m radius turn, at 80km/h and you are in a car which wheelbase is 4 meters and seat is in exact middle, during oversteer you get up to 7 degrees of yaw angle you move just 25centimeters out, as rear wheels move 50 centimeters out. How much lateral g force drop ? It drops from 1.258g to 1.251g, surely thats a lot of seat feel with 0.5% of difference.

Now lets say you take 20m radius turn at 40 km/h and same thing happens with same car. G force drops from 0.629 to 0.621, thats 1.3% difference, perhaps that could be felt slightly.

Schumacher was doing there about ~230km/h and the radius there is pretty much 100 meters, he was pulling 4.1 g's pretty sure there were little difference in seat feel when yaw angle got too big. Moreover you also have to consider that reaction times are not that instant. Also it obviously takes huge talent, experience and probably even biological capabilities to learn how to minimise reaction times and develop precise motorics by using proprioception, vestibular system and visual cues.

Unless FFB does the reaction for you, and the drop off never happens too fast... and simulation ends there.

Yes you probably shouldn't be able to beat laptimes of F1 drivers even in a sim where you are not afraid, don't feel psychological pressure and your neck and feet is not being torn off by 4g's.

The only thing which IMO could dictate a little bit of "help" is compensating input lag, but I would even leave it to compensate psychological pressure. A sim where you can casually get anywhere near to world reccords is not a sim IMO, even if you are champion simracer, it should still take lots of effort.

 

[CobraCat]

How many fatal crashes have we all had in AC before beating an official laptime ?

 

[RasmusP]

Where do you get this nonsense about 0.01° and 2° ? These are super low angles, you wouldn't be able to tell if it is not yet UNDER optimum slip angles at these angles. Also don't you feel weird putting on Schumacher spin next to your statement ? Worth to mention that it is important how quickly the dropoff happens, you might never have the chance to catch the slide in some cases.

I am pretty sure that "seat feel" from real car during oversteer is directly tied with the reduction of g-force as you start loosing your radius. It should be that the closer to rear wheels you are, the more seat feel you should have, as well as visual cues. Another thing is that the lower the g-forces are, the more seat feel you should have.

For example: constant 40m radius turn, at 80km/h and you are in a car which wheelbase is 4 meters and seat is in exact middle, during oversteer you get up to 7 degrees of yaw angle you move just 25centimeters out, as rear wheels move 50 centimeters out. How much lateral g force drop ? It drops from 1.258g to 1.251g, surely thats a lot of seat feel with 0.5% of difference.

Now lets say you take 20m radius turn at 40 km/h and same thing happens with same car. G force drops from 0.629 to 0.621, thats 1.3% difference, perhaps that could be felt slightly.

Schumacher was doing there about ~230km/h and the radius there is pretty much 100 meters, he was pulling 4.1 g's pretty sure there were little difference in seat feel when yaw angle got too big. Moreover you also have to consider that reaction times are not that instant. Also it obviously takes huge talent, experience and probably even biological capabilities to learn how to minimise reaction times and develop precise motorics by using proprioception, vestibular system and visual cues.

Unless FFB does the reaction for you, and the drop off never happens too fast... and simulation ends there.

Yes you probably shouldn't be able to beat laptimes of F1 drivers even in a sim where you are not afraid, don't feel psychological pressure and your neck and feet is not being torn off by 4g's.

The only thing which IMO could dictate a little bit of "help" is compensating input lag, but I would even leave it to compensate psychological pressure. A sim where you can casually get anywhere near to world reccords is not a sim IMO, even if you are champion simracer, it should still take lots of effort.

My post was not meant to be scientifically accurate at all. All numbers were fantasy numbers and the video was just meant to show that real F1 cars can just spin without anything breaking or an oil spill etc. It was about oversteery setups. Or oversteery in some occasions, like Schumachers F1 car seems to be. I've watched the highlights of the whole 2009 season too and lots of spins throughout the season and its sessions.
It purely was meant to show that with "real" setups, it's possible to spin the car around if you turn too much or do a slight mistake without the need to touch the grass or anything.

But feel free to tell me my post is nonsense (which it is, since the numbers weren't meant to be accurate in the first place).
Nice attitude there

I'll make sure to put some "fantasy number for giving an example" next to all of them.

 

[Ghoults]

I think in sims we can get below the real world records in same car/track combo. Even if you had perfect physics model of everything and perfect data you'll still be faster in sim. You can do more laps, you can do laps at any time you want and you can adjust and ignore things that slow you down in real life. Physical gs on your body in real life allow you to feel slides earlier as you feel the change in acceleration before when the car begins to slide. Not see the slide when it happens. You can brake harder and drive smoother.

But there is a drawback as well. When you are being shaken around your inputs become less precise. In sims you can also drive every car with paddle shifters at your chosen ffb level and steering lock and brake pedal stiffness. Your car is always in perfect condition with no wear and tear and your setup changes are not just instantaneous but everything in your car is changeable instantly. Unlimited new tires and no care in the world if you break something. And in case of f1 car they bring different parts to almost every race. How many parts for different races do you have in sim? Are you even comparing the same car? Always same weather conditions, always same track condition. Only weakness sim racers have is that they don't have teams of race engineers and strategy engineers adjusting the car with data. A sim racer can do it but in the end it becomes a question of time and real teams spend thousands of hours at the top level to get it right.

That being said you still need to compare the best to the best. You can not just try to compare your own hotlap records to the schumachers, vettels and hamiltons. You need to compare the greger huttus and bono huises to schumachers, vettels and hamiltons. And with equal physics I'm sure huttu and huis will go faster than the real life drivers do in real life. Fear of death or injury has nothing to do with it. None of the greats lift through eau rouge if it can be taken flatout. Some years ago they did it one handed. But the real life drivers have lots of limitations that slow them down. Sim racers only have one and that is lack of g-forces. But with regards to lap time I'd say not having it makes you ulimately faster.

It might be interesting to see how fast a sim racer could go in some kind of radio controlled real car against equal skilled driver in the same car. With equal weight and all other setups.

 

[mantasisg]

Nice attitude there

I just seek for accuracy.

...You can do more laps, you can do laps at any time you want and you can adjust and ignore things that slow you down in real life. Physical gs on your body in real life allow you to feel slides earlier as you feel the change in acceleration before when the car begins to slide. Not see the slide when it happens. You can brake harder and drive smoother.

....Only weakness sim racers have is that they don't have teams of race engineers and strategy engineers adjusting the car with data. A sim racer can do it but in the end it becomes a question of time and real teams spend thousands of hours at the top level to get it right.

It might be interesting to see how fast a sim racer could go in some kind of radio controlled real car against equal skilled driver in the same car. With equal weight and all other setups.

^^^
It is wrong. First of all, if you don't have consistency and crash frequently it is first sign of lacking skill, and it is very very unlikely that the perfect lap will happen, maybe bigger chance to win in the lottery than to somehow accidently do it right in all curves, all phases and in perfect trajectories and timing.

Also STOP the false about the "body feel". It is so widely spread now. There is absolutely no way to feel the slide BEFORE it happens, unless you are Senna and have paranormal understanding so you can guess it before it happens, and you can do it in the sim as well. It is just impossible to feel acceleration before it occurs, it makes no sense.

And no engineers team is quite a difference, but simulations are not that deep anyway.

Also interesting stuff. Hamilton kinda discovered his talent with RC car IIRIC. He got a chance to participate in RC race and immediately was very good at it.

 

[Ghoults]

Also STOP the false about the "body feel".

Your body does feel acceleration. When you brake your body wants to move forward which means you feel the belts against your body and your foot against the pedal more intensely. When you drive over a kerb you feel the ffb even if it might be small enough to not actually move the wheel. When you stand in elevator and the elevator starts moving up/down you feel the initial change in acceleration. And just like acceleration is the first derivative of speed, jerk is the first derivative of acceleration. The rate at which acceleration changes. Human body can feel that. So when the acceleration changes either during braking it is sign of tire gaining or losing grip or in lateral direction it is the same thing. Or vertically when going over hill or being at the bottom of a dip.

So when you are in the corner and you keep turning the wheel more and more until you start to oversteer you do feel at that moment the change in acceleration. As the tires lose or gain grip the lateral acceleration changes. Before the car starts to slide the acceleration changes. You feel the change in acceleration before you see it. Senna did not have any paranormal abilities. He just had good feel for the car which also means good feel for acceleration.

Best example of feeling acceleration is the operation of clutch. In sims that is super difficult because there is no g-forces. You don't feel the bite. You have literally almost no feel. You just need to look at the screen and react to it when you see motion. In real car you can operate the clutch much easier because when the clutch starts to bite you feel it. It is super easy to learn it and get perfect getaway in uphill, snow, ice and a lot harder in sim. That is because the clutch operation when starting a car from standstill is 99% just feel of acceleration and jerk.

 

[mantasisg]

True about clutch, but it does not come without visual cues it is a mix, today as I went to my car I depressed clutch and the car started rolling backwards because it was in gear and no handbrake applied, I was looking in my backpack and car rolled for about a second or more till I reacted. Also it is not good example, because in low accelearation it is indeed very well felt, let alone accelerating from standing still. But take a lot of acceleration, and same changes becomes less obvious.

And you are wrong that acceleration will change before car will start sliding, it can not be so. And without physical cues you wouldn't be able to tell if the car just didn't slow down by tiny bit, as it is a mix of feel. And vision is definitely dominant in high speed.

If you'd check this calculator, you'd see that radius change due to 7degs of yaw angle in a car where you sit 2m away from front wheels axle gives as much g force change as change of 0.2 km/h while maintaing the radius, at least in 40m radius curve at 80km\h.

https://rechneronline.de/g-acceleration/curve.php

Vision is most important.

 

[Ghoults]

Vision is the primary source of feedback as you need to see where you are going but physical feedback is very important as well because acceleration can change massively even if there is no sliding at all. Audio is important as well and is the main source of feedback during braking because the visual motion cues when braking are inadequate as you can not really sense front wheels locking very well without it.

And there is change of acceleration before the slide. The tire deforms, the suspension heaves and rolls, the bushings in suspension pickup points deform, the parts flex. Even the tires start sliding at different time. And at which moment are the tires sliding? Is it sliding if part of the tire is sliding or does the whole footprint of the tire need to slide before we accept it as sliding? All of that changes acceleration which is the forces you feel on your body. And in right conditions all of that happens before the rear steps out. In sim you can only see tire flex if it is excessive. In some cars you can see it easily like ferrari 312t in ac but in most cars you can not see it. But if you drive the real car you'll feel it.

 

[John-Eric Saxén]

Vision is most important.

I don't know how you can categorically state this. The "seat in the pants" feeling you get when driving a real car is such an important clue to what the car is actually doing. Real life drivers have many times voiced their view on how sim racing feels different to real racing and this is probably the single biggest reason why. For example, how would you judge a lightening of the rear end and a snap oversteer by vision? Surely it's the other senses that come in first, both from the steering wheel feedback and the feeling you get in your back.

 

[Kek700]

Also the wheels outer edges can bend and will flex as much as 3cm and more.
if the tyre lets go unexpectedly that wheels rim will immediately flex back to it natural
position. This is another message that is sent to the drivers bum.

And i still want to know how they compensate for front wheels angular change
effecting the aero in f1.

 

[mantasisg]

Peripheral vision is more important than you think. It is in same mix with balance and proprioception, and I think it is most important.

If it wasn't so important then guys with no FFB would.. well you'd never hear about anybody not using FFB.

I still fail to understand how in case of loosing a car in a curve acceleration can change before it is lost, and more over change massively, it is imposible.

The tire deforms, and everything flex, compress, extend and so on before the car is sliding, thats for sure, but it does not tell that that oversteer is about to happen, if you are experienced you can only make a good guess. Also AC does not give good impression on tires felx and chassis pitch roll. And tires respond too well.

I suppose sliding is where linear range of the curve ends. But you don't feel it because you are still holding curve and increasing lateral force in tire. And if tire still can maintain and moreover increase it's lateral force, is not sliding, even if part of it is. Vehicle will continue accelerating towards the center of the curve, and you'll have no unusual sensation through g forces, before lateral force will drop. The rest depends on how fast is the drop-off plus whats the sliding friction, and if you'll have a chance to slide enough to feel that you are "leaving the orbit".

I'd agree with that if tire has large slip angles then you could relly on seat feel.

 

[RasmusP]

I just seek for accuracy.

And I edited my post
Asking me where "my nonsense comes from" and if I "wouldn't feel weird" to post what I did post might give you the accuracy you like but you sure lost my respect. Maybe think about how to achieve what you seek with a nicer tone. Or the crap that happened earlier this year will happen again at some point.
And nobody wants that!

 

[mantasisg]

Also the wheels outer edges can bend and will flex as much as 3cm and more.
if the tyre lets go unexpectedly that wheels rim will immediately flex back to it natural
position.

It is so ? I think it could only possibly happen with very light car and on very low friction surface. Otherwise, only after stopping to slide lateraly.