“ are rfactor2 physics broken” video

That's kinda how I read that video. Here's some ways we/I can tweak rF2 to have some interesting results. Now, maybe if I was sim racing for big$$, I'd be exploiting other tweaking avenues...but I don't have the time or inclination...even though we're still in a lock down. Keep safe y'all and see you on the other side of 2020. Interesting discussion.
"Setups don't really make you go fast. Setups make the car behave in a way that you can drive fast." Now who wrote that?
 
Slip_Angel said:
This, people need to ask more about stuff that actually matters in a simulator. RF2 can still use OLD U.I but if it has most realistic physics people will buy it no matter. ACC is ahead in all other aspect except physics and hence I am not playing it.
Click to expand...

The walls of texts asking for physics improvements come from time to time, it´s just that most users are aware of where the improvements are needed, we´re aware that the devs are aware and we´re aware of the roadmap so we know when we can expect this stuff to actually be worked on.
 
BazzaLB said:
So what you are saying is the example he shows could be issues with rf2 aero implementation and more examples of cars that are less aero sensitive need to be shown first before the developers should even contemplate investigating?
Click to expand...

Without yet seeing the video, but from what I've seen from some discussions, 'trick' setups in these cars might be at least partially down to the aero implementation in these cars. There is at least one parameter that can lead to unrealistically large aero forces if not controlled (and can also lead to flying cars). You could tweak the car setup in an unrealistic fashion to make use of that aero potential. There are also several parameters that have an impact on what happens with yaw, and it's easy to think something might be overlooked which allows some extra slide angle.

I'm not suggesting the devs not look at anything until other cars show issues, and I won't fall for your leading question :D

For me, first step: identify the issue. Quantify it. Isolate it. Then fix it. I doubt rF2 itself needs fixing (fixing a problem I mean - not touching potential improvements), almost certainly any actual confirmed issues are down to the content.
 
I recall an interview Juan Pablo Montoya (a real race driver from F1, nascar, indy, weather tech) gives after the virtual 24h le mans; he stated that to be fast in the simulator the car must be drive different than the real car; if the car in rFactor2 is drove the same as the real car it would be out of pace against the virtual racers.

As a side note he had drove the same car in the 24h Le manse 2019 and 2020 (oreca 07 lmp2), so he knows what he's talking about.

If I find the interview i'll port it here
 
Last edited:
Emery said:
Running faster by detaching swaybars and lowering the rear wing are issues? Not in my mind since the default setup is understeery. You can see him struggling with the default setup's understeer and rather than tune himself to drive the setup, he just keeps plowing into corners. 1.5 second improvement with a setup that suits his driving style? Sure, no problem.
Click to expand...
No problem ? lol do you even know how setup and physics work ?
He got faster by running lower camber ....LOL
Sebring is for fact high-mid downforce circuit and he got faster by using minimum rear wing ...LOL
Detaching sway bars in car that has underbody aero plus it with affect the roll hence tyre contact patch hence car will need more alignment tuning but still no aero loss ? no mechanical grip loss ?
Excessive slip angles,excessive steering wheel is another problem that is not discussed in the video but it exists FOR FACT still no problem ?
the only problem is guys like you getting in way of simulation.
If you don't know how things should work then please don't comment on such complex topics you are actually misguiding people.

RF2 needs big physics updates for fact , even more desperately than some ui or comp system.
If this "sim" is to be sold as a simulator then it NEEDS has to obey laws of physics and vehicle dynamics PERIOD.
 
What does that mean in conclusion? My personal expertise in everything to do with cars starts with changing tires and oil and ends there again.
As I understand it now, is there actually no simulation that comes close to the RL?
So at the end of the day it's all just "games"?
According to the Montoya statement, the video and the current discussion, yes ...
That is very bitter .. to be honest ..
 
Hooray.

Maybe this thread will FINALLY after 4 years react S397 internally.

Just as we hope that they will have finally understood the importance (relatively urgent) of improving the offline game (AI, rules, equivalent of the Log Analyzer software, etc.).

But physics is not even debatable, it is the most important thing in a simulation. POINT.
 
memoNo1 said:
What does that mean in conclusion? My personal expertise in everything to do with cars starts with changing tires and oil and ends there again.
As I understand it now, is there actually no simulation that comes close to the RL?
So at the end of the day it's all just "games"?
According to the Montoya statement, the video and the current discussion, yes ...
That is very bitter .. to be honest ..
Click to expand...
I think most of these sims comes close enough to real life in that most of the driving skill gained is transferable into real life driving. It's just that the virtual car's behavior is not going to be exactly the same as with the real life version of the same car. It's like going from driving a Porsche 911 to driving a Corvette. It takes some time to adapt to the differences, but after that you get to use most of the driving skills you learned in the former car.
So in my mind virtual car and real life car are both cars, and it's driving you do in both of them, but they are not the same car.
 
memoNo1 said:
What does that mean in conclusion? My personal expertise in everything to do with cars starts with changing tires and oil and ends there again.
As I understand it now, is there actually no simulation that comes close to the RL?
So at the end of the day it's all just "games"?
According to the Montoya statement, the video and the current discussion, yes ...
That is very bitter .. to be honest ..
Click to expand...
I would not paint it in a such a pessimistic way. Like the laws of physics "work" in a certain interval of parameters because they are not "perfect" (Newton works in our macroworld, but no longer in a quantum scale). Similarly most simulations work well within a certain interval of parameters based on how they were coded. When you go out of bounds it starts acting weird. rFactor 2, it seems, allows too much or too drastic setup changes and it goes out of bounds producing unrealistic results. If you don't race online competitively and don't have to pursue these hacks, you should be alright, believing the results this simulation produces as much as it is possible with a commercial public product.
 
@Rastas gave his statemant in relation to physics in the YT video.
His Mercedes Mod comes very close to his car in the RL.
Ok, S397 have to fix these shitty setup "hacks". No question at all.
When I see Marcos Mods, I am convinced that this is very close to the RL.
Whether Formula 3 or the ETCC.
Also the impressive STC2000 Mod from VRC.
But why do the LMP2 have to be driven differently than in the RL?
Because of these possible hacks?
Let's assume this doesn't exist.
Would the LMP2 behave and drive or feel more like their counterpart in the RL?
But why does it work with Rastas with his Benz mod and with the LMP2 as an example not?
Maybe because S397 does not get all the exact physics data from the teams?
Because of team secret data?
Would be understandable.
Ok is offtopic.
Nevertheless I want to know.
And it goes without saying that these "hacks" should be fixed.
 
memoNo1 said:
@Rastas gave his statemant in relation to physics in the YT video.
His Mercedes Mod comes very close to his car in the RL.
Ok, S397 have to fix these shitty setup "hacks". No question at all.
When I see Marcos Mods, I am convinced that this is very close to the RL.
Whether Formula 3 or the ETCC.
Also the impressive STC2000 Mod from VRC.
But why do the LMP2 have to be driven differently than in the RL?
Because of these possible hacks?
Let's assume this doesn't exist.
Would the LMP2 behave and drive or feel more like their counterpart in the RL?
But why does it work with Rastas with his Benz mod and with the LMP2 as an example not?
Maybe because S397 does not get all the exact physics data from the teams?
Because of team secret data?
Would be understandable.
Ok is offtopic.
Nevertheless I want to know.
And it goes without saying that these "hacks" should be fixed.
Click to expand...
exactly these exploits needs to be patched out if this is to be called a sim.
Hopefully the problems are in the implementation rather than physics engine itself otherwise S397 would require complete upgrade.
 
If I have to bet, the main source of discrepancy between simulated car and reality is aerodynamics. Aerodynamics is like voodoo to science. So much voodoo that we still need wind tunnels because we need to check what happen in the real world, because simulations can't come close enough. And when a fast moving vehicle relies heavily on aero to stick to the ground, things get messy pretty fast.... both in RL and in Sims.
 
lagg said:
From that point of view, sims don't exist (in the private market)
Click to expand...
If they all have such unrealistic setup behaviour and overdriving characteristics like excessive steering wheel , easily controlled high slip oversteer
anything that doesn't get as CLOSE as it can get to reality then yes. there are none and software industry must do better.
You can't just stop here and say yes this is as good as it gets ...NO !
If there is a room for improvement then it must be worked on and if there are broken physics then it needs fixing ASAP.
People are paying here real money to get what is advertised i.e simulation of cars. And if a "sim" is failing so hard in that aspect then it does not deserve a "simulator" tag.
It is misleading. by definition any simulation is based on real laws of physics and hence it should act accordingly. RF2 is exploitable in MAJOR way and it cannot be ignored. any real sim enthusiast CANNOT drive without thinking about it.
I'm pretty sure this issue are fixable and developers are not taking any action on it, hopefully they are not planning to add another electric car that there are not even goggle images of.
 
I just want to point to a little detail: The guy in the video did not remove the front packer
And with 20mm front packer you hit the bumpstop at 42mm ride height allready approx. in the CV8 (could be a bad suspension modelling). So he went with a pretty stiff front around some corner sections probably, even if he removed the front ARB
 
Stefan_L_01 said:
I just want to point to a little detail: The guy in the video did not remove the front packer
And with 20mm front packer you hit the bumpstop at 42mm ride height allready approx. in the CV8 (could be a bad suspension modelling). So he went with a pretty stiff front around some corner sections probably, even if he removed the front ARB
Click to expand...
Good observation mate, even in my testing i found out that C8R requires much less packers than other cars i own.
As you said it could be a suspension issue or my theory is car has softer wheel rates.
If i'm correct wheel rate could be totally different from car to car even if springs rate are same (pretty basic stuff sorry about that but I had to post it to avoid misunderstandings).

You are correct that despite detaching arb he is basically running on bumpstop so car is still very stiff
BUT
here is my theory now since all the rolling stiffness is provided by front bumpstop only the car should be very compromised mechanically, it should understeer HARD. as not only the front is provided majority of stiffness BUT also bumpstop are IMO much stiffer than springs and ARB, bumpstop are really there to stop the car from bottoming out and prevent extreme pitch and squatting so you can imagine how much stiffer a bumpstop would be.
The fact that such compromised setup is getting quite fast lap times on bumpy track is ridiculous
 
It is bumpy, but not bumpy as RL, in simulation there are no resonating effects that can literally destroy the car, so he can get away with such setup, but .. can he consistently lap like that with that setup? Because is in the long run that you see if a behaviour is broken, or if he is just good for 1 or 2 laps. In RL a setup like that would not be used, but this is different from saying it would not work, again, for 1 lap. But who has a car to wreck for the sake of verifying?
 
You must log in or register to reply here.
Back
Top