What I believe to be wrong with the RF2 tire model / grip levels / lack of control

I used to drive LC and LH Torana's, HG Monaro's, HQ Monaro's and many other Australian V8's 20 yrs ago.....I knew nothing about driving yet would take my LH Torana{original engine was 308, but it was 330cid stroker at one point before engine catastrophe, LOL}, and I'd drive like a manic in general, but also around Mt Cootha, and IMO, if cars where that hard to drive, I and others would be dead many times over.

I can appreciate laps-1-3 or further representing a problem, but we're talking about experienced sim players driving purpose built race cars, and the GTR for example strikes me as having an incomplete rear end, and I've tried a number of set up changes, but it sux hard, yet that thing would utterly wallop the cars I used to drive, granted they were all modified for better overall performance etc, but I never experienced any crazy low speed snap spins, plus, the rear of the GTR comes around and then keeps on coming, but it doesn't seem in proportion to my road positioning or the amount of throttle applied.

 

Fair play,

But as you said, "the rear of the GTR comes around and then keeps on coming", which you have to appreciate is simple inertia. In sims, people tend to think of inputs as finite and definite causes for reaction without considering basic physical principles they cannot feel as far as I can tell.

You need to cite come reasons why the GTR's end should STOP coming around when you've previously commanded it to do so. Simple friction (which is not so simple with tires) is one thing but there's a lot more at play as everyone here knows, and something as simple as drivetrain intertia gets overlooked more often then not. Just because a driver lets off the throttle after causing some notable wheelspin, doesn't mean that a lot of energy he put into the system just magically disappears, and since the tires are at that point (especially racing tires) hot, the drag down of energy through heat is not going to be so drastic and a rotating driveline; backpressure notwithstanding; has a lot more inertia than I think many people realize.

That explains the fairly common video (which I'd find if I was not such a lazy bastard) of the F1 car at Monoco continuing a spin at speeds that are so slow it looks surreal despite no throttle input.

In this debate I really think you have to go by reasoning and math since "feel" is just so subjective that it's irrelevant. I'd rather adjust my "feel" to what's empirically true instead of adjusting what's offered by the sim to match my preconception. Any sim racer who is passionate about realism needs to be open to a paradigm shift here and there in my opinion!

 

GT1 car is capable of cornering 2G, has weight of 1325kg, it has 600hp and 660Nm of torque, @2000rpm car has already 369.6Nm of torque. Camber is around -3 degrees depending from setup of course.

Because it is designed to perform really well at speed, corner at insane speeds, it has also very stiff suspension with quite small motion range of wheels.

At slow speed it just can't bite the road, reducing rear camber, softening suspension as much as one can, it can do bit better, but I think it has still big issues of tire contact with road.

I'm getting 400N at inside rear at tight corner, around 80kph speed and that is rear ARB detached, there was also bit of cobblestones for added challenge.

Racing tires are also quite different from street tires.

I really can't say that it would be tires fault that GT1 is a dog, there are numerous things at play, but when setup is adjusted and tires are warm, one can push throttle straight down and countersteer, quite fun to exit corner wheels spinning and engine hitting limiter

I need to compare GT4 and GT1 tire loads and speeds at corner exit to understand better what is happening there, GT4 is such that I have been really sideways without issues, only when tires heat well above 200C issues start. Last corner at monaco with GT4 car is 4 wheel drift every time.

When I tested F1 car, I could not notice very well steering becoming heavier with speed as downforce builds up, I wonder if such should happen?

 

Oh, how silly of me, that means all my other sims are utter crap as they have a more progressive nature....and as I said, NONE of my many muscle cars did anything like this, and I had no knowledge of racing technique at the time.

That said, the GTR also seems to lack much in the way of weight transfer, so I keep crashing at T1-2 at Palmbeach despite having constantly crashed there and being ready and trying hard to be smooth etc.....the front end and the brakes on this car are as good as any sim, in fact probably better, but the rear behaviour is a sheer nonsense.

Anyway, that's my opinion, and if ISI fail to address it, I suspect it'll be yet another car abandoned by most rf2 players.
Not much point building them if people don't want to drive them.

Do you have GSC 2012?.....you must think that's arcade as it has incredible feedback and an excellent sense of weight transfer, as does GTR Evo, another arcade game.

 

Btw, one of the prominent modders who hangs mainly at Nogrip, had the opportunity to speak with racing drivers, and he claims they always tell him the cars are on rails.....he also mentions that initially, they're as bad as anyone, but once they get dialed in, they're quite impressed by GTR evo and some of the WTCC packs etc.

 

one of the problems I think is that allot of people think real cars are hard to drive.

when in reality its hard to drive real cars fast and get that last 0.300 out of them but its not hard for even a moderate driver to keep a car on a track especially anything below GP2.

Think how many hours sim racers put in compared to real drivers and yet sim racers find it harder to keep the casrs from spinning / on the track evan at the top levels of sim racing with people that are putting 400x more time into track time than real world drivers.

Sure real life gets hard from the fact that if you crash its going to hurt scare the **** out of you and cost you allot of money so that affects how people drive ( but not so much at the higher levels)

Some people will also find the G's and general violence of real world driving off putting and scary but for people that dont and spend 6+ hours a week from the age of 15 driving are going to find cars stupidly easy to drive when it comes to simply getting them around the track within 1 second of the pace.

So when people say NKP is to easy maybe its the case the RF2 is needlesly hard , personaly I would not use the words Easy and hard as I can still drive RF2 with most cars in the game to within 1s of the pace after a good number laps I would simply say RF2 takes control away from the driver far sooner than NKP and it is in fact harder to drive a fast lap time in NKP than it is RF2 because there is more depth and degrees of angle and grip at that point where tires begin to go off.

 

people still ignoring my last post lol.
i think im also going to relink something that Domi linked in another thread, to add to my last post:

http://www.radicalsonline.com/conte...peaks-about-his-life-simracing-and-netKar-PRO

and im gona relink my video of me driving, and michael schumacher driving.
im going to keep doing this until one of you rfactor fans tell me why you can do this in real life (and netkar pro, and to a lesser extent, gtr evo), but Can't do it in rfactor. okay?

 

Ummm, maybe because they are still working on the tire model as has been said so many times in this thread.

Do you really think that if you post the same things over and over it is going to change anything? All you are going to do is basically get ignored because you are the one not listening. You are comparing things that are full released products and real life to a beta that is still being worked on to get things right.

 

oh okay, so now you are admitting that rfactor is majorly flawed in terms of grip levels when at and over the limit then, and you're putting it down to the fact that its still in beta?

fine, im happy with that.

look im just trying to get a discussion going based around some evidence. The best evidence we have is in the form of video. i could harp on all day about my personal driving experiences in real life and in simracing at a high level, but none of it matters and we'd get no where.

might as well focus on video.

i put the Q&A from Kunos in because it basically 100% says and agrees with what i think and believe to, so all my thoughts are laid out, in that Q&A with him.

Also.....yes .... of course we are comparing real life and other sims to this beta.....Thats the point of this entire thread. There are people argueing that rfactor currently is the most realistic, and the cars handle in a more realistic way than other sims, and is closer to real life.

There are others, like me, that are argueing against this. This is what the thread is about.
Your response to my question (and the videos i backed it up with) is
'its still in beta'.

Fine, i'm happy with that.

 

We have been told by Tim, and we have discussed here that the tire patch is the area currently being worked on by ISI. We have also discussed that doing what you see in real life is a lot tougher here because of the lack of the seat of the pants feel you have in real life. No matter what you do in a sim, you won't have that seat of the pants feel unless you have spent huge amounts of money on a motion sim to get the G-Force feel at all 4 corners as well as the feedback in the wheel.

That being said, we have discussed that the reason that NKP feels and handles the way it does is because they have fudged with the FFB in order to put information to the wheel that would not normally be there in real life. We have also discussed that slip angle and grip are key factors in being able to control the car. If the tires are already at maximum grip level and a slide occurs, physics is going to do what it will with the car until such time as grip levels come back.

What you did in real life, and what Shumacher did in that video were well under maximum grip of the tires and/or was done with the feedback from the seat of the pants letting you know that you were starting to slide and could correct.

Does this mean that rF2 is majorly flawed, no. This only means that something as important as tire grip and slip angle is still being worked on. Up to the point of maximum grip, the tire physics works and there has been no complaint about that anywhere from what I have read. past the point of maximum grip is where things start to get fuzzy for everyone's opinion.

Yes it would be nice to be able to prevent being looped every time the back end gets loose, but that doesn't happen even in real life, and we don't and can't have all the feedback needed to do that in this sim. I still stand by my statement that I would rather have real and accurate over faked even if it makes my driving a little harder so that I can learn to live within the bounds of the real. If I can ever afford to get a motion sim, then I would expect more information and would expect to be able to handle the cars a lot better.

In real life you drive by what you feel more than any other sensory input. In a sim you drive more by what you see and hear. Big difference and training your eyes and ears to be as fast and reliable as your feeling makes it tough.

 

Lol netkar pro ffb = 20 lines of code! Others might be over thinking???

 

I don't think that's true, in an interview Stefano Casillo said this.

"Comment: There is an ongoing discussion about using fixed tables for the data or real time calculations (not a maths guy, sorry!), like ISI they say.

Personally I don't do that, but I see the advantage of doing it that way. My fear is that this is exactly why I feel disconnected in their sims. Sometimes people ask me what I did to make FF good in netKar PRO. I am clueless.. I think there is a fundamental difference in how FF is calculated in the group netKar PRO, iRacing, Live For Speed, VirtualGP3, even Gran Turismo Academy felt "netKar-ish" to me. I think my FF is, I don't know, 20 lines of code?"

http://www.radicalsonline.com/conte...peaks-about-his-life-simracing-and-netKar-PRO

As an experiment turn ffb off and then do what they did in that video but ignore everything and just turn the wheel left and right as reers spin up you will see that regardless of what the player does in rf2 it has no effect , where as with NKP evan with FFB off you can still gude the car to some exstent or though still not as well as you can in real life

 

BTW, I watched that video with schumacher again and did a bit of my own analyzing and here are a couple of points you should look at. In the beginning, he is warming up his rear tires before the start, and before the car gets too crazy on him, he backs off the throttle to make sure that it doesn't come around. In the part where he is taking the corner, he is catching the slide very early. If you look at his tire angle, you can see that they don't move more than about 10 degrees, maybe less. This tells me that the car never gets very far off center at the rear. Likely it never exceeds about 15 degrees off center at the worst part of the slide.

In most cases, we don't even know we are into a slide and can react until we have passed that 15 degrees and this usually ends with us coming around and running off track.

As to the argument that Stefano Casillo writes his FFB with 20 lines of code, he never says his calculations are real, he just says that he does it fundamentally different.

 

The piont is he takes the FFB directly from the physics of the game and they are no more fudged than ISI's FFB though granted ISI with RF2 are taking the data from the front coloum alone at the moment which works well for communicating track bumps and some aspects of the car but seems to totally ignore others ( I belive ISI have said they are looking at bleeding in other info into the FFB to fix this)


We are talking about simulators non of it is Real even if you use 100% captured data its all about compromise and clever work around's to get it to run in real time.

the issue is who ends up with the most reolistic moving cars at the end of the day when it comes to actualy movment of the car in game , and actual required user input to change that movement.

and then FFB is on top of that but we are trying to keep this on the topic of tires and car behavoir.

 

Actually James, I don't use FFB at all anyways, and I can launch with some tail wagging and still keep it under control.

 

Yes you can play both RF2 and NKP with FFB OFF that's why this discussion was not about FFB , the fact is even with FFB off NKP is more realistic than RF2 in many ways.

Someone kept making the argument that the stability in NKP comes from the FFB , when I would say thats not the case, its just that NKP has better car stability and car stablity thats closer to that of a real car.

 

I would think of ISI's method like a BMP graphic where as Kunos use vectors

IF isi have the data to make that BMP of a high reolution and its an image of the whole spectrum of a car then ISI's model would win out over the method Kunos use.

At this piont in time Kunos's method provides better resolution in possable car movment and car response to that of ISI's method but the issue Kunos's method faces is that it wont directly line up to a Real world specifc car as well as ISI's method once ISI have the resolution required to fool a human.


This is also why ISI's method is better for real world race teams becuse they can 1) use there own tire data and 2) can have it at a very high resolution and run that from a specially designed PC

Though nothing is set in stone I'm sure someone could fine a way using the kunos method to make something that's near 100% realistic and I'm sure someone using ISI's method can make something thats near 100% realistic.

 

ONLY when much over the limit, when riding that edge of limit I believe rF2 is the best (rF1 was very good in this area I believe aswell), but yes when going way past limits as I said in my very first post on the 1st or 2nd page, rF2 can definetely still improve in that area. If you look at my video that I posted of a guy racing I believe build 90, you see very good and natural corrections, even though it may not be as good as NK Pro in the "way past the limits of grip area" its still not as horrible as you guys are making it sound and actually gettin pretty good.

Also NKP is the best in this very specific area, but even that isnt perfect, as its a little overdone and too easy and definetely flatters the driver in thinkin hes great and can handle any car,in any grip situation, in any slide, regardless of conditions, how off your setup is, how un forgiving a certain part of that track can be, how your treating the car etc. etc etc etc etc. It seems to give every slide just an extra tiny bit of leeway to catch it. MAYBE its some sort of SLIGHT compensation in order to make up for lack of feel and much later feel that us gamers get, and due to this compenstation it makes it feel more natural, but if you actually analysed how the car was acting/reacting with telemetry data maybe you would find that its being almost a little too forgiving in certain slides/grip loss situations???????? Hmmmm...........

But yes, regardless if that possible theory is true or untrue, regardless, yes I do believe in this area NKP is better than rF2, and more natural and "normal" feeling (part of this is due to how the FFB works in that game, but yes also with the pure physics/tyre model of the sim as well).

No way, physical car movement and response, action/reaction of all the variables which affect what a car is doing millions of times a second etc etc. is in my opinion way better in rF2. Its only the "once youve gone way over the limits of rear gripr" that it can sometimes seem a little off, but even that is getting better and better every build.

Other than the way past the limits of rear oversteer thing, I believe as a whole, nothing is close to rF2 in terms of pure physics modelling (not the numbers behind the physcs modelling, but HOW those numbers are affecting how the car is acting and reacting to well, everything)

 

Read my post above and you will see why its likely that at this point in time NKP provides more resolution in terms of posable dirver input and interaction with the physics system than RF2.

Obviously RF2 potentaily maps better to real car when its working but I would argue that is pointless if you then have massive holes in the simulation or the data has not apreached a resolution that lets you clearly experience what the real car is.