Physics discussion thread

I haven't got any impression that people are somehow attacking or dismissing Baldwin here or his video, we are just discussing his take on the physics and handling of that Ginetta, which he claimed to be unrealistic in the video. He has for sure more credibility that most, but it's not like he would be expert of all the cars IRL and as far as I know, he isn't know being a IRL Ginetta G55 GT4 driver (or ever driven one), so claim that the handling is unrealistic isn't based on any IRL experience of this particular car.

As many know, he isn't just occasional IRL GT3 driver, but also know esport sim racer, who looks for fastest lap times, even if it means abusing flaws in the physics. I think it's fairly logical that he goes and tries to see if turning of TC and ABS leads to better results than having them on (not very uncommon in sims to have "flaws" like that). His claim that TC at 0 is best is probably also true for single lap pace, but is it really for even for sprint race length... I think he should had done more races with both TC on and off to see if it's really the case.

I really enjoyed the video (and Baldwin in general), I especially like that he was one of the very few in a long time who did video racing in MP in AMS2/LFM and not just against AI like most content creators these days (okay some do some ad hoc custom lobbies as well that can be hit and miss). I would just really hope he would have time and interest to do a bit more of a deep dive, actually try and race with multiple cars and classes, tracks and conditions (and not immediately with some funky TT setups), before giving his take on overal realism. Those content creators a lot followers and viewers, so their take on things like this do influence general perception of the game and I would hope they would put more effort into it before giving their opinions.

Snetterton one is with supercup though, which at least in the version we have in the game has no TC/ABS (and from the video, it's clear that the driver isn't giving tons of brake input)

But yeah, looks quite a lot like how it handles in AMS2 to me as well.

 

Ask 3 different people and you'll get 4 different opinions. I'm pretty sure opinions are divided even between real-world drivers, let alone wider community.

It's good to know which aspects is repeatedly reported as incorrect, but at the same time every title have a flaw here and there. I think the quest to find the most realistic sim is doomed from the start.

 

A little bit of delay is realistic, cars don't turn instantly the moment you turn the wheel, as if they were literally trains on rails. Unless you also have input and/or screen lag, in which case the delay can be too much, I used to have this problem in the past due to bad settings. To check this, watch the wheel shown on your screen, as you turn your real wheel, the image should match instantly.

The esses section in Bathurst is my most hated section of all tracks in all sims, so I avoid racing there, but will give it a try tonight. Which car are you using and with what TC settings (if any)? With all the cars/tracks I tested so far, I don't see any sliding, it all just feels natural now. I'm using completely default setup though, possibly silly setup values can make the cars feel silly again.

 

That describes what I feel too — it’s like the cars only turn when the rear slides, not when the front has grip.

 

Me too, i need to anticipate when in my braking point, i can release the brake and turn the wheel. Its a bit different than lmu. But gt3 really have been improve in the last patch

 

When it comes to "realism" in racing-sims, I think our expectations are . . . simply unrealistic.

 

You are going downhill, weight is already on front and then for that corner, you also need to slow down and brake putting even more load on the front, so it's not surprise if the rear loses grip and front takes over easily there if the setup isn't tuned to counter that in any way. Nords is a quite a long track with plenty of corners, with many downhill as well, so you would need to be specific, which ones are you meaning. Spielberg T3 (or T4, depending on how you count), is similar, very easy to load the fronts so much that the rear starts to flip on many cars.

As for delay... I at least don't experience such in the way you are describing. That coming down the mountain section on Bathurst is actually one of the place where the GT3 cars actually feel heavy with inertia pushing the car in the corners quite a lot.

For me, there is at least very distinct difference in feeling thru the FFB especially when the rear is slipping and car rotating vs. turning and it definitely don't do that so easily anymore on GT3 on default setup unless you have very specific kind of corner or apply braking deeper to get that behavior. I'm not sure do you mean this or generally all situations where you are turning the car (in which case, I definitely don't experience it the same way)?

This rotation vs turning, reminds me of this video

(it might use a bit wrong terms at times, but it's good nevertheless).

For a while now, I have been thinking that on AMS2, when you slow down and the cars is diving/pitching weight transfer going to front, it's like it reduces the rear grip too much (even if it's just little braking, maybe just from engine braking) and too easily, but not only that, when you ease off the brakes, the rears don't regain grip as fast, which leads to those slow spins.

If you have corner that is also going downhill or at end of a descent, like T1 on Nurburgring, Spielberg T3 or that Forrest Elbow on Bathurst, it's even more excessive.

I think that doesn't happen that easily with the GT3's anymore outside on some specific situations (and makes them a lot better to drive). It was negated already in the September patch with the driveline/diff fixes on those cars (before that it was really hard to trail brake). This is on default setups... you can still get that if you change the setup more dramatically though. Before those changes, you could reduce it by weakening the front, adding a lot of bump stops or stiff strings to front to mitigate the behavior, reducing the front rebound dampers to make the front "bounce" back the weight more to rear faster and things like that (which you can still do if some track causes problems).

GT4's on the other hand, has still more of that going on (especially the Aston the last time I tried it).

ps. maybe mods could move these conversions again to the physics thread...

 

Having just looked at a few of his videos for the first time a couple of days ago, I don't think anyone could credibly dismiss him as talking nonsense. He is one of the rare people who is able to perform fast laps in all of the products we call "sims" and who also has real-world racing car experience, including coincidentally in one of the cars that is in all of the sims. From what I saw, he is much more cautious and measured about saying anything definitive about any sim, which is wise IMO.

His GT4 video mentioned that he was not using Default or Default+ FFB, which is a problem for me if comparisons are a big part of the presentation. Even if using one of those, we all know the quirks of FFB in AMS 2 and how many people mistakenly misadjust their wheels and damper settings.

He chose to not use TC, which is something I usually do when learning a car, or when using the similar Cup car because it doesn't have TC! In a race scenario, even in the dry, I would likely use a low-level of TC for safety purposes and tire management, but I don't have his level of gaming skill, so I wouldn't pick on that. It is possible that he was overheating his tires at points, but how would he know that? The HUD does not clearly show flashing temps on the tires, so someone playing for the first time or first few times may not figure that out.

He borrowed a fast TT set-up as a short-cut, and then used it in a race probably not realizing the depth of simulation in AMS 2 can punish you handily for that choice.

He discusses the unrealistic ability to lean on the car in corners and abuse the slip angle. We all know this is the key "debating" point about AMS 2 versus other sims. We know that the cockpit cam settings can exaggerate the perception of this for some. We know that the default vivid audio feedback (announced as a positive and will also be used in the upcoming PMR sim) can exaggerate the perception of this for some. We know that the advanced tire model in AMS 2 has many layers and adjustments and it can be more or less accurate to the real world depending on the skills of those programming it. But we also know that when you take into account variable and evolving track and tire temps, and track grip, the same car with the same tires can feel quite different from one session or minute to the next...something that is again more "vivid" in AMS 2 than in most of its competitors.

With all that said, it is possible that the GT4 Ginetta is still a bit too "slidey" in corners? Of course it is, because we already know it has been adjusted multiple times by the devs. Many other people have said the same thing, based on experience with other sims. Now we have someone with that experience, plus real-world experience with the car saying it again.

If I factor in the likely sub-optimal FFB being used, watch his skill level with the car, take into account his relatively top-level experience, I think I can safely say I respect his assertion. I think the level of adjustment needed to "correct" the situation is quite small, and it may not all be in the tires. The physics are not broken in any way, shape or form, but a couple of per cent here or there could make it a bit more realistic.

Given that he isn't a clown, clowning around, like some other YTers, and attempted to do a race with no TC after having only run time trial, his presentation made both himself and AMS 2 look pretty good overall. I suspect when he gets a chance to spend more time preparing something, he will have a thoughtful review of some combo in AMS 2. He may realize that the game is realistic enough to humble even someone of his talent level who cannot skip the use of TC in race conditions or rely on "meta" set-up tricks to gain "free" lap time.

 

In my limited experience with the current Ginetta GT4 in AMS2, it is indeed faster (and safe) to run it with TC off. And it's not the only one, many other cars are like that, including the Alpine GT4 (very stable) and some of the GT3s, for example the McLaren can be raced with TC off just fine. I think (and I'm not saying I'm sure of course) a problem in AMS2 is that as mentioned earlier, TC even at 1 is very intrusive and cuts too much power for too long (it was the same in PCARS2). I cannot know if this realistic, but I can only suspect that it isn't. So bottom line, in many cases it's faster to have it turned completely off (unlike ABS). Maybe that's my only still current complaint for GT3s, that their TC levels must be adjusted.

And it wasn't tire overheating that caused his spin at the end, he just touched the grass on the left at full throttle and that sent him. I agree he is not a clown like other youtubers, but still he's doing videos for the views, so he has to make them exciting one way or the other, so we need to take this into account when judging his comments. But it was very good to see that he didn't blame the game/physics whatever for that spin, I think he accepted it was a rookie mistake from him.

Back to the Ginetta, the biggest thing I dislike about it is how it often slides (not spins) massively when it touches kerbs, for example the inner kerb in the last turn in Barcelona, while as shown in the video above, the real car can handle it just fine. There are other cars in AMS2 with that behavior as well (another place is the kerbs in the right handler after the 2nd fast left hander in Spielberg), GT3s used to have this, too, but with the latest updates they can handle such kerbs just fine now. So I'm hoping the latest great improvements on GT3s will propagate to other classes as well.

Final comment, about TT setups, yeah this is a problem now, there are completely outdated now setups there, and when people try to use the top times setups because they think they will give them an advantage, they will probably have the opposite effect. So I'm hoping Reiza will find a way to clear leaderboards of classes that have had massive updates recently.

 

What I mean is that the cars (as a general observation) don’t respond to the fronts to turn, they 'only turn' when the rear steps out. This could very well be related to FFB or POV, but I just can’t figure out what it really is, so I’m putting forward these propositions as food for thought.

Exactly that: too easy to get into a glide, too slow to settle down (that’s what I mean by “slidey”).
Some people don’t like comparisons, but pragmatically, that’s the best way we can describe our perceived reality. I don’t really think R3E has a better tyre model (as in a more complex, detailed one), but it feels much better in many respects. The cars dive into corners with confident grip, and the rear steps out naturally after being abused. When that happens, we get an FFB warning and a quick spin soon after — but once we master the car and react fast enough, we can risk using that small window. If not, our lap times pay the price!

 

To add some element to what is getting discussed in the patch topic and surely might be moved over here. I just recorded just a quick Z06 GT3.R race on Mosport

The Corvette is quite agile which allows it to turn in sharp. Yet I have to really work very clean with the brake to get the car to rotate. If I just throw the steering I will just understeer off track. I can easily feel the weight shifting forward under braking and the front wheels making the turn in, compared to other comments they always have to step the rear out to get rotation. I never got that feeling at all, at all times the rear felt very planted where it should, not once did I get the old feeling of risking to throw the car away. (of course breaking downhill towards the tight u turn always needs good attention to what you're doing since you can totally feel how light your rear is there). Even stepping on the throttle to easy in some corners never resulted in overrotation but plausible understeer, you can see it in T2 on multiple occasions that I have to step back on throttle to get the cornerexit

Since I ran the base setup with no tuning just TCS one click you can see that tires started to overheat most likely to the not optimal camber for this track, so the car punished me a bit for still pushing that the last lap was a bit wild but all in a very reasonable amount.

I wasn't in my best pace in the end since AI was a bit too low at 107, but with a 1:25.7 at best lap it was alright. With proper tire heat and better matched AI pace a low 1:15 was in it for sure.

So now my question is just how would the same track look for people who say they have to step the back out to rotate or change clutch/engine brake to not fly off under braking. And this is really just meant as informative as it could be to maybe get more clues why people might still struggle with GT3s. So maybe more people could share their Mosport laps, preferably with Aston or M4 since they are even more planted.

Additional note: I moved from Custom FFB to standard when the updated ones came up since the custom files totally washed away any feeling of grip to me. I could turn and the steering just kept having reistance even when grip was already gone. With Standard FFB I can feel the exact grip dropoff moment where I steer to much an the steering goes lighter

 

The problem with AMS2 is the incorrect and very peculiar wheel load. This is caused by the vertical forces (Z-axis). This means that the static and dynamic forces that occur during braking, acceleration, or when driving over bumps are simulated incorrectly. All these forces ultimately affect the wheel load. But I've come to terms with it and use other simulations instead. Today I'm testing AC Rally, and I think the physics will be realistic; I'm curious.

 

Since this is the Physics thread, could you please explain on what basis you are assessing the correctness of the physics engine of the game, which is a completely different animal than custom FFB files. If you don't have any direct access to it or knowledge of it, please stop repeating that the physics are broken/wrong, especially when it is highly likely that virtually identical calculations for wheel and tire load are being used in every modern sim. This is not one of the complicated or mysterious parts of the simulations.

 

I really respect the work and time you put into the custom ffb files you modified but clearly you have a negative bias towards AMS2 now for whatever reason. Personally I disagree with your criticism of the physics, what you have been saying recently made sense to me prior to the last couple of updates but honestly since those updates the change feels so much better to me using the default ffb.
Anyway each to there own of course, it's purely subjective when it comes to feel.

 

That's precisely the problem. I know my custom builds inside and out and know exactly when the physics have changed because it would also change the custom files. What you're noticing are the new default force feedback (FFB). This improves the feel but doesn't change the physics. The reduced sliding and less floaty feeling comes from the new FFB, although it's still present (just less so) and unfortunately has other anomalies. But it's just a masking of the physics, just like I've been doing with custom builds for a long time. But at some point, there's a limit because the physics simply don't allow it anymore. I don't dislike it; I'm just disappointed with the physics because, for me personally, it hasn't improved but remains the same. Take a look at AC Rally; I tested it today. For me, that's good physics and realism, even though it's still in Early Access. Yes, everyone has their opinion, and I have mine, and other people have theirs. I don't have to justify myself, and as I said, there are enough options for everyone to have fun. But I'll let you know if AMS2 physics really does improve, at least for me

 

You are aware that there have been innumerable updates just to the tires over literally years now...and that many other aspects of the physics modelling have been updated? There have been so many physics-related updates to this title that it has actually annoyed some people who would prefer just to stick with what was there for their TT or set-up development purposes. I also wouldn't want to be the one to inform the Reiza team members who have worked tirelessly for years that all their work was actually just in their imaginations because it had no effect on physics. You have lost whatever credibility you had left with these ridiculous statements.

Back to reality, it is true that the recent FFB changes may be more noticeable or more immediately noticeable on some cars than the physics changes--probably because the physics changes already occurred over the past months and years. GT3 cars, as an example, have had both noticeable tire/physics and FFB changes in the most recent cycle. This is readily apparent from the user feedback here on the forum commenting on both.

 

No one needs a custom FFB file as the FFB is perfect.Just adjust your wheel software and ingame settings for your likeing.Simple as it is.And as i see mostly Moza users are moanin'...

 

I've actually been using the "Basic" built-in default custom FFB and actually quite like it. I should switch back to the others to see how it compares. It seems to work very nicely with the Ginetta G40 GT5 which is what I've been driving lately. I agree nobody actually needs a custom FFB, however I understand after driving in R3E for a bit recently the feeling that DanielKart was going for with his files - it's quite different to the way AMS2 feels. Better or worse is completely subjective.

@Dady Cairo I use a Moza R9 V1 and I have zero issues with the FFB only problems I've had lately were that Moza had a major update to Pithouse that reset everything at the same time that AMS2 released their major FFB changes so it took a bit of work to get things back to a way that I liked (in Pithouse).