James Baldwin comments on the McLaren 720S GT3 Evo in AMS2 1.6

We don’t have to wait, of course not. But to have a race driver that says good and bad things about a Game it’s a good chance to improve the Work…. He has the experience how to drive this GT3 on the limit and how the cars behave….so as developer you can take this as chance to do some right improvements and not on the Feedback of people that don’t have this type of experience.

 

I said some not the ones mentioned here, but those who (want to) know everything better and better.
Edit: Real race drivers imo are H.J.Stuck,G.Hill etc.Every talented monkey can drive a GT3 car with a little practise.

 

How? I mean physics is maths and maths is numbers. Where are the numbers in the video?

I don't really think you can really translate hand-waving into physics improvements - certainly not if someone is saying it has to be as realistic as possible.

The way to get something as realistic as possible would be to capture data in real life, understanding the maths of physics, perhaps come up with some empirical formulas that will work in real time on the hardware you have and then see if your output data matches closely the real life data.

If doubt anyone could make a game more realistic by watching a video of someone saying "I think the cars are doing this too much" whilst waving their hands - Although Ian Bell's employees have probably tried. "Subtract 5 from these parameters...no wait, I don't think he moved his hands that much, only subtract 3"

The best comment I saw on the youtube was something like "A driver can do laps all morning, go for lunch, come back and tell you the car has changed completely even though no one has touched a thing"

When a sim developer hires a driver and uses this in marketing it's usually to dupe the customer base not because they have any great insight. It's cheaper to pay a driver to say "Yeah, it's pretty much like my car - the game is great kids. You'll love it" then it is to get a team of people spending months working on physics and handling.

Remember too it's a common scam along with "Yes it'll have a realistic-ish handling model" that arcade games have used to release an arcade game but to get people who want a sim game to still buy it - and, sorry to say, the sim community falls for it every time. One guy has done it twice with need for speed games, 3 times with his next company and will do it again soon. And codemasters have done it more than once, but especially that first Steve Hood designed codemasters F1 game they went all in on using social media and hiring an F1 guy and duping the section of the audience who wanted the new game to be a logical successor to geoff crammond or whatever.

They paid Anthony Davidson some guy who briefly got an F1 seat but couldn't keep it to add the 'an F1 driver said the game is great' BS and supposedly work on the audio. They even faked lap times because half of the sim audience seems to believe that whether the physics and handling are good hinges on whether the lap times are the same as IRL - well faking the lap time is easy.

I know racing drivers often think the team listens carefully to their hand-waving and input. But they don't. Of course they don't. They've got banks of data, phds and computers. A team of experts in aerodynamics and engineering. You know if you trained a monkey to press buttons to go into space you don't start asking the monkey stuff about the rocket do you? You just give it a banana and get it to sit in the corner until the next take off. Of course everyone cheers the monkey, we all love the monkey.

If Alonso says "The car is really bad out there today" and it isn't they say "No worries, go and talk to the press and we'll fix the setup" and then they do nothing because they have the data showing them it is perfectly fine. Similarly if the data says the car could be improved and the driver says "The car is great! Don't change a thing!" - they'll say "Sure, we'll leave it" and then they change the setup so its better. Placebo and ego - you tell the driver what he wants to hear and do the right thing with the car.

As for someone hand-waving in a video about the handling - how do you translate that to maths? You'd need all the inputs and all the output then crunch it through the physics engine, tyre model etc and then you can say "Yeah this is why the car behaved like that when you did that input" - and it might be right, it might not. That isn't this, is it?

If F1 wants feedback from the driver they put him in the car and send him out on the track and collect the data.

 

We have this amazing opportunity that members of our simracing community have got the opportunity to drive in real life, we get direct feedback, from simracer to simracer, but our ego doesn't let us accept that the sim of our choice may not be as perfect as we think. Please, do some self reflection. That guy races in that car, but has also very close ties to simracing. Who are we to say, nah, his opinion and findings are worthless. Come on.

 

Men Reiza had a lot of data an numbers for the physics they don’t do this since yesterday. The point here is that some cars in AMS2 are not credible in their behavior because they have a lot of downforce and when a guy say this in a very friendly way so I don’t know what can be so wrong. but yeah take your banana and seat on the corner and do some maths.

 

I think some of the problem here is that AMS2 is not these particular YouTubers main sims, so they don't dedicate time to crafting a setup that suits them. I'm pretty sure Mr. Baldwin puts much more time into his LMU setups than he did for his AMS2 setup. They aren't going to devote their time on something they are just trying out for fun/comparison to their sim of choice. Of all the YouTubers out there I respect James Baldwin, Jardier, and Nils opinions the most. They are honest, point out issues they see, but aren't condescending like so many others.

Edit: And they are just opinions. Doesn't mean I have to agree with them.

Edit 2: Changed "JB" to James Baldwin so not to confuse him with the other JB.

 

All sims, set up or not, have certain characteristics that may be more or less realistic, AMS 2 included. Setups are for tweaking, not changing fundamental characteristics. I haven't felt any condescending attitude from J. Baldwin, I have to give kudos to him for pointing out the negatives and positives in every sim he tested this car in (ACC, iR, LMU, AMS2). When Jimmy Broadbent pointed out how he doesn't really consider AMS2 very realistic a good while back this whole community was up in arms. Many physics revisions later... was the community right back then?

These are opinions, yes. Not all opinions have equal weight though. I am more likely to accept the opinion of an experienced person compared to a random forum opinion. Just like in other fields and topics the opinion of an expert bears much more weight than the opinion of an influencer or a populist politician. Or at least it should be the case.

 

This has been my theory for a good while now, though I do also think the sims people grew up on also play a massive role. People may be used to a way a certain engine feels, so other games ran on that (or similar) engine will feel a lot more comfortable compared to one that isn't. It's a subconscious thing, and a bias that is incredibly difficult to look past that I think all simracers have.

I'll use myself as an example. The two most important sims for me growing up were Project Cars 2 and Nr2003. Today, my two favorite racing sims are AMS 2 and Iracing. I have a hard time believing it is a coincidence, especially with Iracing because I had a very negative bias against Iracing until I tried it, it turned out that I gelled with the sim very well. Compare it with GTR2, a game that so many say is one of the best sims out there... even after multiple times of trying to give it a fair shake I still struggle with getting to grips with it. Despite the positive bias towards GTR2 I had before going in, I still didn't particularly enjoy my time with it. It's telling that I had a harder time with GTR2 when compared to GPL.

 

In general yes. Less clutches means less locking which results in wheel torque (and therefore acceleration) being limited by the (unloaded) inside wheel.

 

Let us know when you find one then.

 

Well a real competitive racing driver is kind of an expert for car handling and general feel, characteristics compared to random simracers, or do we pretend that we know just as much? Or that drivers know nothing?

 

Most folks are not saying he "doesnt know anything" just that there are a few things that would improve his experience. Not gonna sit here and pretend many AMS 2 fans aren't overly-defensive at times (they are), but that doesn't automatically invalidate the criticisms being made against the video. Honestly, two things can be true here... that James is feeling something fundamentally wrong and that he could've given AMS 2 a fairer shot, they're not mutually exclusive.

 

I suggest everybody would simply stop putting so much weight on oneoff quick test from a Youtube person, even if he is a real GT3 driver and has generally more worthwhile comments when it comes to both sims and real racing. (Yes, I watched the video because he is a real driver and it is always interesting to hear comments between sims from more credible people.)

Get LeMans Ultimate and try it yourself to make comparisons in handling between these sims. Sims are so close to each other these days that default (and any other) setups matter much more than their small differences in modelling different things.

If you have VR, it is a great thing to get perfect field of view for most sims, and give you a fully comparable view of yaw and small changes in the movement of the car. If not, make sure your FOV is calculated correctly for both sims.

In LMU pick any GT3/GTE car, and choose Circuit of Americas as a test track. It has fast and slow undulating turns and it is generally good for testing car behavior.

In the car setup turn off TC and ABS. Not to brag, but to find the limits for everything.

Decrease preload to half or one third of its original value.

Increase dampers' fast rebound value quite considerably from the defaults (in McLaren the value are reverse but use the right arrow). Set also normal rebound few clicks stiffer. These help with snap oversteer and give time to react to tank slappers you are bound to get otherwise with very stiff bump damping and stiff springs and racing tyres.

Hit the track to burn some rubber. Try driving turns normal and and very aggressive ways.

I really don't see that much of difference between AMS2 and LMU. Both are lively, both slide to all directions quite a lot, both can be driven very well without traction control, even flooring it most of the time, both have lots of room for error until you clearly cook it, and in both you can affect greatly how you go in and out of turn, when the setup allows it, in very similar manner.

If you have driven rFactor2 in the past LMU is pretty much same but in a neater package. Which is good because rF2 is all over the place when it comes to UI design and content, and its Steam workshop based store is horrible to find and get coherent experience.

All are great sims! All deserve your attention. Above all their handling does not differ so much from each other that one could pronounce that "this is realistic sim and that one is not" without stretching it greatly. Where the differences are these days is in live track systems, tyre heat modeling, weather modeling, AI, online etc. and even these are unifying in features.

Also what is important in sim racing is subjective. For instance I give credit for AMS2's FFB for exactly what James Baldwin is not happy about; that it is busy, giving out a lot of information. He is fully entitled for his opinion that AMS2 puts lots of extra info through the wheel, and it is realistic from a steering wheel point of view, there is no denying it.
For him it just breaks the illusion of reality. But AMS2 bridges the gap for those who still miss G-forces and suspension load and bumps, and they are needed particularly with the cars that do not have any assists.

 

I think the biggest difference, honestly, is LSD modelling between AMS2 and other sims. Nils Naujoks is the only person who appears to have taken note of this, and even has a video explaining how to use the LSD to your advantage, and HAS to use AMS2 to show the biggest difference in operation because it appears to be the only sim that actually lets you change the diff as much as you can. The difference in on-power behaviour between 6,4 and 2 clutches (i occasionally have gone up to 8 on indianapolis) is so marked that anyone stating 'the cars in AMS2 do X, the cars in LMU do Y' or whatever, without taking into account the diff, is essentially not credible. I even think that the years of complaints about slidey cars would 90% have gone away if Reiza had set the default to be 4 or 2 clutches.

 

Could you please explain what this setting does?
I'd be very grateful, I really would because I don't understand the explanation in the setup menu, I'm afraid.
I experimented with clutches the other day in the BMW M3 E30 (Group A) and I had the impression that with higher values the inner wheel (not loaded) gets all the torque but after the description in the setup menu - or better my understanding of it - with higher values more lock should be applied what would mean that the inner wheel should get less torque.

Lowering the value made the car more stable in bends and made it easier to get out of them.
So as you wrote: It would be benefitial if the default value was 2 or 4 clutches instaed of 6.

Sorry for derailing this thread.

 

The way I understand it is as follows (and then we'll get back to the thread).

When turning, the weight of the car shifts and the two sides of the car have more or less weight (and therefore net grip) on them. Also, turning means that the inside wheels are on a smaller 'circle' than the larger ones, and so the outside wheels need to go faster than the inside ones in order to turn. If the axis is locked, i.e. the wheels are forced to spin at the same speed, then the wheels on the inside of the car MUST spin more than they should since they are forced to rotate at the same speed as the outside wheels, which have to go faster to go around the turning circle at the same time.

Therefore, we invented the differential, which allows us to send torque to the wheels in a way that is dependent on how much they want to spin. Doing this means the torque can be spread more appropriately across the wheels. In an open diff, i.e. one where the wheels can spin independently, this torque will go to the wheel that spins most easily; the inside, unloaded, wheel. This means the wheel on the outside, which is most loaded, will spin less, wasting power as this is the loaded side of the car.

To combat this, the LSD was invented. This essentially balances this spread of torque between the two wheels, based on the difference in torque that WOULD be sent to the wheels if the diff were open. Intuitively, when the LSD 'notices' a torque difference, it locks the wheels together 'more'. The times when you get a torque difference are when you're on power, and when you're coasting, hence the different settings for power ramp and coast ramp. The clutches also multiply the effect.

So, to understand it, when the clutches are HIGH, and when i power out of a corner, the LSD acts more strongly when it sense a difference in torque. So, when you power out of the corner with high clutches, lots of power is sent to the outside wheels, sending you turning. The example always coming to mind is the hairpin at the end of the back straight at Barcelona. Take e.g. a nice GT3 car ethat you like, set the clutches to 8, and see how it behaves as you power round the hairpin. If you drive like I do (poorly) you will find yourself spinning around as, as the weight transfers tot he outside as you turn, and push down the throttle, the diff will lock, your outside wheels will get all the power and you will oversteer into the corner. Do the same with 2 clutches. You will find this effect much less pronounced

Overall then, I use high diff clutches when I want to use the throttle to turn me more, like on circuits with tight corners when I want to be accelerating and turning a lot (for me, e.g. Indy road course). I use lower diff clutches when I don't want this effect to be so pronounced so I spin less.

I hope this was clear and useful, and we can go back to the thread now. Any corrections, if I have explained something incorrectly or wrong, are welcomed.

 

To drive fast doesn't require you to think about maths or physics. We generally start someone driving whose parents are hoping they'll get free flight and tickets to F1 races eventually, around 6-8 years old. Long before they have started algebra.

It's like if I throw a ball and you hit it with some kind of bat or catch it you're not the guy I need to explain aerodynamics.

When drivers talk to sim racers that's marketing. You might hope there's some conversation and exchange of information between people who develop computer games and people who work in motorsport - but, you know, me being on the phone to Lewis Hamilton because I've played a computer game and he's driven a car? That's not going give anyone any insight into anything.

 

The amount of ego displayed by some here is just unbelievable, a reality check is needed for some. I hope you all went in and tested that TT setup before claiming any of your opinion here and discrediting his feedback.

I did and ended up 2s off of his pace, without ever driving that combo before, of course it is slow, but I am no esport/professional driver and will not even pretend to be. That should just put some perspective here. I did not struggle getting around the track and went even one step beyond by adding just click on the wing to be able to take eau rouge flat, so yes I agree he could have been better. But guess what, his time might be very limited, way more than most of us.

But guys, we are talking about a person that drives the car in real life, got multiple wins in it, got wins in different sims in different series including in this very same car, if any of us is able to adapt to a setup, even with fault, it's him. The same person that just jumped in random TT session in the LMP2 and got world record on Spa and Le Mans in his live Stream. So stop claiming his opinion does not hold any value because he used a "random" setup. He also did his test prior the BOP balance, at a time where both Audi and Lambos were busted, without that he would probably be at the top (by the way we still need a reset of those GT3s gen2 leaderboards...).

As i stated initially at the end of the day, he was smiling and having fun with the game, so what's wrong with his opinion on realism that rub people the wrong way ? I see the same pattern in each and every post that touches any topic related to how the game feels or cars behave at/over the limit.

Feel free to provide him a "proper" setup if you feel it would make justice to the game.

 

as I’ve already made my submission to this elsewhere,I will just copy and paste
But basically I don’t have a problem with his views , but the poor process he let influence it , rather than giving it the same process he uses for other reviews he’s done on other games .
And yes I tested his set up and have done laps in same car in LMU , and they don’t feel that different apart from significant understeer on throttle in Lmu, and the way everything feels in slow motion due to lack of speed feeling Lmu has .
—————————————————————————
Copy and paste from another discussion about this video .

Just remember it was made with previous tyre model for these cars, he used an unknown Random Set up that had max engine breaking and over sensitive steering and an unknown and probably not ideal FFB configuration.
the engine braking information guide even says in its description, this is largely user preference… and is more or less set somewhere in the middle .

I tend to increase ( reduce effect ) a couple of clicks myself, but even with Default set up I was able to just about keep up with his time -1.3 secs and lose 0.6 of that in final sector without low drag .

With a set up bmw I’m only a few tenths away. (considered the slowest car )
And definitely wasn’t over driving , more trying to be as precise as possible right on the edge of grip and trying to not induce much yaw that would overcome it .
I’m not an Alien or E-sports guy , I race weekly with guys faster than me .

the crux is , if he actually tried to set the car up himself from default and sorted his FFB out , ( even though E-sports guys don’t tend to rely on it ) he should be good enough to set a better time by driving it properly.

I like James , and his reviews, and passing knowledge of experience , but he really didn’t do a great job with this one .