True. Harder isn't more realistic. Back in the days drivers were able to keep a car on track for 24 hours which wouldn't be possible if the car would slide all over the place. Or as Jackie Stewart put it " I always wanted the car that the easiest to drive".

Also depends on your driving style/aggression. Its common to push on the limit 100% of the time in a sim, where in reality you would race much more cautiously/conservatively. The cars are a lot easier to drive even 5% under the limit.

@@GPLaps Not really true in the way people think about it ie: "they just drive slower in real life". It is very much so true for amateurs; not for good professionals. The only area real drivers typically struggle in compared to sim is braking at the very-very limit in extremely fast cars like prototypes. You can pretty easily compare limit behavior in cornering situations via telemetry, video feed and opinions in most situations.

@@GPLaps Indeed we are always looking for the higher difficulty levels when in reality things were often different and not so hard nor the drivers used to push so much. A mistake could mean death ... not reload.

@@jamesbehra2690 correct

He did a video on it; it’s an amazing car!

Different car! Its on my list

@@TheTotallyRealXiJinping this is a new Delage from 1927, the Delage already reviewed was from nicecupatea's 1923 season pack.

ooh that sounds exciting. Can’t wait!

I was going to mention Power and Glory mod, would be good to see him try it out. For me the best mix would be Solitude with the Ferrari 275, one of my favourite combinations in sim racing.

@@carlosfandango2419 , He's uses Power & Glory mod for some of his videos. The 1000km Monza 1966 video he did recently is one.

@@Dethmeister OK, cheers for that.

I generally agree, but AMS2 and pCars 2 also put modern tyres on their cars while AC simulates old tyre-compound. That's a huge difference. Former DTM-driver Volker Strycek was driving his Opel Omega DTM at the Norrisring after the car didn't move for 20 years. He just put new tyres on, new brake fluids and was 4 sec. quicker than 20 years before (1:04 vs 1:00) on a track with basically four corners, a less competitive environment and two decades in retirement. And than you get always fresh tyres on the rim in sim-racing even if just changing setup. In AMS2 the handling can change dramatically in some cars once the sticky layer is off and I already suggested recently to give us a used-tyre-option since basically every racing-team I know is practising with old tyres first.

@@Leynad778 AMS2 should implement vintage compounds which are less grippy

Weirdly, gas karts felt like modded forcefeedback from project cars 2, similar to rfactor 2 but I can't get anyone else to agree

no. the mic compression is hurting my ears.

@@lxvideostuff7200 I was going to say the same, need to cut a bit of high frequency to make it comfy on headphones. Other than that I do think they are good.

a mk 17 Legrand??

wow! thank you, glad i could help :)

It is true that vanilla AC has a really bad SAT curve in the tires. It does indeed go light too soon. It doesn't *really* affect the driving physics, but it doesn't feel correct. Some cars with low caster trail suffer pretty badly from that. It's still better than canned FFB that gMotor sims without RealFeel have, but it could have been better. I'm unsure why KS did it how they did, maybe to give "more feel" as a gamey thing. I don't really agree with that approach; wheel torque would ideally be 1:1 to IRL. Cphys does add much better SAT by default and you can tweak it if you really want to but the majority of content doesn't use Cphys at all. Also, most cars that people will have driven have powersteering, which feels entirely different from manual steering. It has that "always stiff no matter what" feel and you can't feel the SAT dropoff *at all*, so to people who have never driven a manual rack, the very presence of SAT to begin with will feel wrong. Extra wrong if it's biased towards being too light. It might be one reason why people think cars are always understeering in AC, when the tires aren't even on the limit yet! If you want an idea of what a manual rack on reasonable tires can feel like in AC, download the NSX from RD, and make sure to get the hotfix. I implemented most of the relevant suspension and tire Cphys features.

@@ArchOfficial sounds like you know what you are talking about. tell me how to make ac better, as i dont really get it from the text above

I use the camera in free cam mod (F7) and record those shots

@@GPLaps thanks. As simple as that :)

It depends mostly on how much it has been driven in, since (I think) assetto doesn't change the surface itself depending on the setting

To me ac feels extremely progressive and intuitive to drive. A bit too forgiving but there are mods that are very good to drive and have more limit. What I mean by that is that the cars are not as slidey at the limit but still not on/off like I feel gmotor sims feel to me. Overall I feel ac is too slidey and if you are not comfortable with the car sliding and want more on rails experience (like raceroom) then ac will feel odd. To me gmotor games (ams1, rf2 and raceroom) have much much more pronounced and clear limit and generally once you go over it is a lot more about saving it than controlling it like in ac. In ac if I slide I can generally feel whether I am above or below the limit. Just like in real life the car can slide will still being under the limit. After all the peak grip is generated at certain slip angle. In gmotor sims when I slide the car it feels like it is already beyond the peak slip angle instantly. I have always struggeled to really feel the cars in any and all gmotor sims. The limit is just so much narrower and once you go over the limit the car behaviour becomes super nervous and the throttle just doesn't do much. It is all about just steering the exact right amount to stop it spinning and stop it from doing a tank slapper. In ac I feel the throttle is more important than the steering.

@@erwinlommer197 Thank you for your feedback!

Well, the limit of dynamic friction is significantly lower than the limit of static friction in almost all cases, certainly with tires. Once a wheel spins, it wont stop spinning, until the force the road surface exerts on it becomes significantly lower (ie you step off the accelerator for a while). And Im absolutely certain this is modelled in AC, its such a basic rule of physics and car handling. If you feel like the specific values used are too forgiving, well, just mod them.

Definitely iRacing, RFactor maybe 🤷. I don't know. I love AC

No, AC is still regareded as one of the best sims. Of course there are people who prefer another sim over AC, but it's the very same the other way round. I have no idea what Jake was up to when he tried to talk AC's physics down.

Its elitism. And the people talking about it usually have no idea about the underlying systems. But the level above AC is simulating tires as 3D softbody objects, with deformation due to load, and unequal load and grip across the surface cculated in real time, as well as location based wear and temperarure modeling. I know BeamNG does some of that (the former sertainly, the latter I believe in part), but I dont think any of the "hardcore racinf sims" do it. All they have is a handling model tuned to be less forgiving, which makes people automatically assume they must be more realistic.

Scaling feels ok to me, but, you can change the world scale on a 'per application' basis in SteamVR

I think i get what you’re. Some of these tracks aren’t realistic to the time period because theyre far too smooth, but the settings still help complement the experience.

@@GenuineRage Yeah, and also, in case someone tries more modern cars to period tracks. That situation is more or less only a one-edged sword; older cars work well in newer, smoother tracks, less bumps to worry about, but the newer cars would struggle on period, bumpy, and possibly dirty road surfaces of period tracks, since their stiffer suspension would make them buck all over the place I think...

@@jcgabriel1569 yeah that makes sense as well. Im still new to simulators and i run AC and ACC because theyre the easiest for me to get into and mostly because of the modding capabilities, not necessarily the extreme realism aspects, but thats still very important to people who really take sims seriously. Cheers.

Everything you just said was completely incorrect

yes!

@@GPLaps AWESOME! thank you, enjoy the videos, enjoy your style. no one else is doing what you do and it is inspiring.

Logitech G29 should be on sale right now and is definitely the best entry wheel!

All of the modern sims do as far as I know. Even Gran Turismo Sport simulates that and it's not a "hardcore" sim.

AC does.

@@ImpliedVolatility704 At the time I was running iRacing, which was years ago, it didn't model dirty tires. If you cut the course, you would get an on-track slow down penalty. Later, they would issue warnings and after so many warnings you could be be disqualified. Maybe its changed by now. I always found it unrealistic that the fastest drivers were always testing the limits of the track boundaries beyond all reason and find spots where they could cut the course and in doing so run over dirt or grass and suffer no loss of grip in the process.

iRacing is really not very accurate at all. Their data is probably good especially for the suspensions, but the tire model appears to be impossible to correlate at all, even if they're able to tweak the curves after the fact. Thermals are still bad, which is a step-up from terrible that it used to be. If this has to do with the model or the parameters they put in, I don't really know. Certainly they do a lot of tire testing so it should not be much of a data issue so I am inclined to believe the model has issues. rF2 and it appears Pcars/AMS2 has much of the same problem. If you care about making a sim car for driver training or setup testing, you should use AC with CSP if you must use a consumer sim. Your lack of data will limit you much more than the sim ever will, at least for typical road-race applications. I'm sure gMotor 2.55 and the Reiza version (rFactor1, AMS1) are good; I am not too sure about gMotor 2.10 (SimBin titles). I have bad memories of GTR2 and I've talked smack about it in the past but it can be 100% just bad parameters (It was years ago!) so I will refrain from giving any real opinion. I think the main difference in the new gMotor versions is largely feature additions, not changes to the actual modeling, but I am unsure if Reiza changed some of the more esoteric parts of the modeling. The gMotor used in rFactor 2 is significantly different tire-wise and should be avoided for any sim stuff.

Yeah, I was wondering what sims he was talking about hardcore sim racers moving on to. There aren't many newer ones. I've heard good and bad opinions about all of them. Assetto Corsa Competizione, Automobolista 2? Certainly not PCars 2? RaceRoom, rFactor 2, and iRacing are as old as AC too.

​@@Dethmeister People generally get swayed by the sensationalist marketing talk about physical tire models in rF2/iR/AMS2/PCars so I'm gonna assume it's one of those. They say it's clearly better than the empirical brush style models but can't elaborate how nor am I too sure they have ever even thought about it. The other side of the coin is the people who don't like physical tire models, but don't really understand empirical brush models either and just go by how many load contact points the model has. I assume they will prefer a model with 10 points that just averages the load compared to an actual good model with 5 just because 10 is twice as big as 5. The people who actually know what they're doing and want to build a sim car just use rF1/AMS1 or AC.

Do any models with real time softbody tire simulation exist yet, or is that still beyond the capability of consumer processors? One ought to think unloading it to GPU should be enough.

@@egoalter1276 BeamNG, but it doesn't feel the most accurate compared to things like AC or rFactor2

Yup, ‘ice rink’ is not fun.

All of it is an illusion. No mod can fix the empirical one-point tyre model that cannot ever capture in a realistic way the old bias-ply tyres. So instead people just claim that "whatever feels good to me" = "realism".

​@@azarisLP If you'd ever read a paper on brush-type tire models you'd understand in the early chapters why that's not relevant. There is negligible to no difference in results between a single point vs multi-point brush model when the track surface is relatively flat ie: when driving over anything that's not a significant kerb or pothole or something. The real advantage you get from multiple load contact points on a brush-type model is being able to more accurately represent the loads the tire is under when traveling over an un-even surface. There's a reason many non-realtime brush models are a single disc. You can correlate bias-ply tires just as closely as you want irregardless to load contact points. There will however be some error in all single load contact point brush model tires when ΔFZ is higher than it would be in real life ie: when driving on and off some kerbs. There's also a bit more going on there than just averaging out the loads between the points that you would want to appreciate to get accurate behavior. By the way, a bad multi-point model implementation will offer little to no advantage over a single-point model and might even be worse in some situations, just in a different way. You can guess what I am referring to.

@@azarisLP Oh, I noticed you also have "empirical" in there. They're far better and much more usable than physical based models right now so I don't know why you added that in. It's not really possible to correlate tires on a physical model due to the great interactivity and esoteric nature of some of the inputs (Similar to some arbitrary values used in good old Magic Formula) hence why to my knowledge no actual sim uses them. The only exception would be if you generate the tire physically, then just manually tweak the curves one by one, but that's not much of a physical tire model now is it. I suspect for the time being they're limited to theoretical development for tire companies only.

Load transfer can't be simulated accurately? That's one of the easy things that you can get 100% absolutely correct. It is not that difficult to code a kinematic solver that works and is correct, compared to coding one that works but is incorrect. This must be the one thing that all decent sims do exactly right. So are G's; the math is not terribly complicated. Nothing is preventing you from getting exact~ G's and cornering speeds around any given track. In an empiric tire model, you just put in the curves you want and the correct result will come out. Now I will be honest and say that no consumer sim has correct elastokinematics and flex for suspension/chassis components, and those will affect your wheel angles and roll stiffness distribution. Those are more like missing features than the sim being inaccurate. They typically don't matter until your car is already at a high professional level and I think most pro sims don't even support them, but it's a reality. You can't really make a kart into a typical hard-body sim because they rely on the component flex for springing. Tire behavior is more or less accurate in AC + CSP, you can correlate a tire set basically exactly even a good bit over the limit and that's with heat included to a reasonable range. Some minor interactions are missing, but it doesn't really show up on telemetry nor is it humanly possible to feel. Older gMotor has quite good tire behavior too although heat is bad like in every other sim apart from AC with CSP. Thermals wise, none of the consumer sims really do a good job, not even vanilla AC, but you can get a really good simulation if you use AC + CSP. You can get temperatures to very closely match even a bit outside of typical driving parameters, like when sliding a lot and over-driving, or just rolling down the highway. They cool down correctly too. That goes for brakes too. Of course, the tire damage and whatnot is more complex IRL than in any of the sims, but the heating/cooling/grip relationship can be very precisely replicated in AC + CSP. No sim that I know of can yet truly do temps while drifting right, where the surface is 300 - 500c but the carcass isn't heating up too much, but you can emulate it with some temp/slip interactivity in AC + CSP. I don't exactly foresee a technical limitation for simulating this correctly, but getting usable data from a tire at 500c can be quite challenging. I don't know where this idea came from that some things "can't" be simulated accurately. It's really more an issue that developers don't have good enough data to use as a reference; but some developers do and they can make an accurate simulation. The main reason most consumer sims are not accurate is because the data used to build the cars with is not accurate. A large part of it is guessed. Sometimes devs just make subpar implementations too. It doesn't necessarily mean the sims *can't* be accurate if good data and a skilled engineer is used.

Absolutely. Empty Box made a great video about that, it's called "Perception and Feel - "Sim" is a Lie."

​@@rideroundandstuff Ah yes, that video. The one that people keep using as a reference even though it's wrong. I'd guess and say that Emptybox hasn't developed a simulation model for a car before either so it's not like he is speaking from experience. It might be relevant for 10+ years ago when all of the consumer sim tire models were junk and all the peripherals were junk so you could just give up trying to make accurate models. You can get very, very realistic car models done nowadays if you use the right software and there is very little question about that. It's not a subjective thing at all. Even the consumer hardware peripherals are becoming somewhat usable for simulation purposes. Of course the very nitty gritty stuff is difficult to match to IRL and most models don't actually do the tire dynamics exactly like IRL (You don't have to, though. Makes almost no difference to the point where it's okay to just use slipvelocity and a simple flex model for example) so there will always be some subjective aspect to it when you're correlating the car with your drivers. Indeed the seat, steering wheel, graphics, sounds etc. will affect how accurately the information from the sim is communicated compared to the real car. However just because correlating a car accurately is difficult doesn't mean it's impossible, or that it comes down to subjective beliefs as opposed to quantifiable outputs.

@@ArchOfficial Tires and road surface will stay guesswork for a lot of years and you're delusional if you think otherwise. Oh you have super fancy data sheets that are 100% accurate? Take a look at them again in 10 years and have a good laugh about how "100% accurate" there were.

​@@rideroundandstuff Have you actually built a car model or are you just parroting some stuff you've heard? Maybe for consumer sim game developers it's guesswork now and it'll remain guesswork in the future, but that's just a lack of data and in some cases knowledge and ability. Consumer products are not simulations, they're 'realistic' games. Nobody will use the time to actually correlate a car or a set of tires. A car correlated to high accuracy in 2021 will be just as accurate in 2031. Are you saying that models depreciate with time?

Even the GT3-cars in AC are more fun to drive compare to ACC, but both have in common, that the (non-dynamic) rubber-line is by far quicker. That's what makes vintage cars more difficult to drive because they go more wide and more often land with one or two tyres on the slippy part of the track.

some people’s trash are other’s treasures

50 years from now, we'll say the same thing about today's race cars.