Can we objectively judge car behavior or car physics in sim racing?

Lately I kept reading posts about car behavior and car physics in AMS2. It was kind of strange that one user would say "car x is incredible bad" while another user said "car x is great but car y is really bad" and so on. That kept me thinking about how do we actually judge that car behavior? What is influencing our thinking (or feeling) process? Are we really able to be objective?
So, here is what I came up with what influences us. We have different sim rig setups. It goes from game pad up to motion enabled rig. We have different skill levels. We have different expectations. We have different favorites in different car classes. I believe that all this factors influence us and our objectiv thinking and our overall experience in sim racing. Most of us have never driven a car on a real race track and even less with a race car. I would be interested to hear some thoughts about why or why not we are able to judge car behavior and physics.

 

Yes you can judge physics by just watching what the cars does and comparing it to what your RL cars does. If it jitters instead of rocking softly whan you steer or brake, something is wrong with the way damper or springs are modeled. Same if a light cars with low center of gravity feels top heavy or when a rear end is too unstable to be usefull as a racecar, but feels like a drift setup.

 

To me, the only ones that can judge a car physics are the drivers that drove it to the limit in real life. All the others judge it based on supposions and beliefs that most of the times are wrong and not according to reality.

 

Well, sort of ... I was lucky enough to be able to drive some race cars at a limit and some times on a closed track (non track days), namely things like race-spec M3 e46, Porsche Cup (the nasty one with no ABS or TC), formula renault 2.0 and some non important others and yes its a totally different experience BUT the feeling in the simulator isnt that far off tbh. Esp. things like the AMS2 custom file or RF2 tyre feel are extremely close. Porsche Cup in both RF2 and AMS2 is actually awesome.

In terms of physics, you still need to me some what knowledgeable in the mechanics of the car in order to tell if the car does or doesnt feel right in the sim compared to the real thing. At the end of the day though, you are the person sitting in your rig and it needs to feel good to you I know there were a lot of discussions around physics here and I myslef dont agree with some of the things that were done in AMS2 (not just physics wise). But at the end of the day, its a product by Reiza and they have to be the judge of what is and isnt right for the game. We can only speak with our moneys

 

i'm no race driver but i expect a 1995 v12 f1 car with no tc to have some degree of slip when exiting a corner on full throttle, not just a sloppy auto centering lateral movement with no challenge required by the driver to keep the car on the road.

 

Good old Gerhard didn't seem to have much trouble controlling his slides at Monaco.

 

I guess its a case of startle versus surprise.
For example if you are driving a high performance car on the edge and you suddenly lose grip that could provide a startle response. A surprise response is when you get a reaction that doesn't conform to your expectations, and is separate to your startle response.
To give a car specific example, if the rear starts sliding in an oversteer manner that is expected. It can startle but shouldnt surprise. But if its followed by the front then rotating back around the rear in the same direction, so you end up with a kind of reverse front oversteer that brings the car back to alignment (if that makes any sense), that does not and so is a surprise.
Ive noticed it on the formula trainer mostly. I think its this kind of processing that results in a lot of the physics based discussions we see throughout the sim community.

 

I'd say that you can judge physics best by comparing data (numbers).
If data from the sim (e.g. tyre pressure/temperature or travel of the springs) is very close to real-world data (same car & same track, same conditions, ...) I'd say that the sim has a high simulation value (I know that this term is ... problematic ).

And if you don't have access to numbers you can relate to visual observations like McClutch has already said. You can compare the movments of the sim car with the movements of its real counterpart in certain situations.

You have to be sure, though, that sim car and real car are set up equally (not only the very same car but also [nearly] the same setup).

Otherwise it's very difficult to draw valid conclusions, I guess.
I wouldn't rely on "feeling" because a wrongly set up FFB could easily lead to misjudgement of the "physics".

Edit:
Another approach could be to diagnose the influence of setup changes in the sim like a few of you were doing when you were playing with differential settings.
In the end it led to changes in the sim which have been embraced by many people as a step forward regarding simulating real word behaviour. (Sorry, I'm struggling with words because English is not my first language).

 

And that's where the issues come in. People expect something to drive a certain way with no real knowledge or background, and judge a car in a sim on those expectations instead of asking/investigating. I've been guilty of that myself in the past.

Maybe the 1995 V12 hat a lazy engine, maybe the engine weight gave the car better rear grip. Maybe you're right and the car in sim has too much grip. But your expectations are no basis to reach a conclusion on that matter.

 

Agreed. There are too many people in sim racing that still expect sim race cars to drive like this (like they did in 2003):

 

I totally agree.
Maybe we can drive these car with less tyre pressure than they could/can in real life ... maybe a rubbered track in AMS 2 has far too much grip (no accusation but only an example) compared to the original track conditions of that time.

 

And that is the most tricky part. The teams will never tell us what setup they are using even though we have the option in sim racing to do the actual setup. That brings up another interesting question. Would a real world setup work in sim racing? With that I mean would the real world car setup translate correctly to the car behavior in sim racing?

 

I have collected real world setup data over the years from diffrfent sources, also OEM data available for car shops and racing parts manufaturer data. This is what I use for setups and I then judge the cars by their behaviour and how they compare to production models from Porsche and BMW that I have driven in the last 30 Years.

My current reffernce cars in AMS3 are the Cayman and the MIni Cooper JCW, simply because I had two laps in a rented Cayman on the Nordschleife in 2017, and I owned a 2009 Mini Cooper S.

 

That is really interesting. It means that the real world setup does work in sim racing.

 

OMG, you already have access to AMS3 - lucky you. (Just kidding.)

It's great that you can relate to real life experience. I wished I could.
The fastest car I've ever driven was a VW Touran with 140 hp.
I mean ... as a teenager I spent a couple of laps karting but there are no memories left in my brain to compare with the behaviour of the karts in AMS2.

And as Micropitt asked:
Could you please tell us a little bit about real data and your experiences with it in AMS2? I'd be very grateful.

 

As a general rule, wipe all default setups, never use them in anything that resembles a real race of any length, they are plain bad, only good for hotlapping cold tyres in TT.

Toe and camber
Get rid of negative Toe in the front, only very few cars benefit from negative toe, and those that do are FWD or Race cars meant for very tight corners, Autocross or Car Slalom. better increase steering angle instead and go for neutral or mild toe in. Don't use excessive camber values, -3.0 is already a lot. Always use toe in at the rear, always, edxcept for FWD cars, there you can use more camber. Driven wheels shoudl always sport less camber then not driven wheels.

Damper roll bars and Springs
Springs and Dampers, I prefer softer springs but stiff dampers, particulary the rebound should be stiffer. I always reduce roll bar a bit, more in the rear.

Also aero. its often to much downforce in the rear....I use the entry to "flugplatz" as a test area, the front should not lift there at half throttle.

Experiment a lot, the McLaren F1 Street car is a interesting testbed since it allows for quite some drastic changes.

But for our luck, the general known rules of car setups work flawlessly in ASM2 as well as in RL.

Im a bit out of luck on LSD, but I try the most common LSD setting first: 40/60 for BMW or 60/40 for Porsches, 60-80N preload, 100 for high torque models. for FWD I prefer open Diff or 75-80, equal ramp angles.

To start your own investigations, look here: Racing cars and their history - Racing Cars (wikidot.com) from there browse the online catalogues or race parts suppliers etc....or have a chat with the lad from local motor sport clubs...and repair shops. A lot of repair shop owner have had their own Motorsport or tuning experiences, very often in rally sports since it has a low entry levels.

For myself that madness began at school, aged 17 with planting a 100HP FIAT Engine into the rear of a naked VW beetle chassis and trying to make it somewhat drivable.

Did I mention that I like the Fusca?

Motorgilde Bieberstein - YouTube

 

Thank you very much for your write-up.
I can relate to most of it (concerning setup changes in the sims I play) but I've also learned somerhing new.
Again, thank you very much.

 

Some interesting points here from someone who knows a lot about simracing but has hardly ever driven a real car

I think he's expressed it quite well, and he nails many of the differences between real life and computer simulation, probably because his impressions are so fresh.

 

Back in my GT Sport discord league days there was this guy who’d say “the blabla car drives exactly like it drives irl, I know this because Ive driven one”.

He was using a PS4 controller.