WHERE IS LIFT OFF OVERSTEER????????????

I'm not going to continue this because you're not worth the effort.

 

A better response would be to address the content of the discussion or @steelreserv's specific points instead of resorting to childish retorts. Both of you are sincerely interested in a serious topic. But if I may offer, there are technical industry/sector norms and standards for a reason. It's so technical discussions can occur with a shared language and definitions and not devolve into ridiculous circular or semantic debates, which are inevitable when someone doesn't respect the norms and insists on revisiting foundational aspects of a topic.

LOOS is and always has been a thing to be avoided--it can be dangerous for the driver and promotes unforced errors in a sport where errors can lose races more often than superior skills can win them.

Someone above mentioned we should be discussing rotation and other helpful subjects instead of wasting time and space debating semantics about LOOS--this was a good suggestion. @steelreserv has published a bunch of useful videos on his YT site that attempt to help all of us understand many aspects of set-ups and how they work in AMS 2. I suggest a retreat to those for most of us. @Andrew Hollom, if you have something useful to contribute to help us understand which cars are behaving or misbehaving in AMS 2 and why, now would be a good time to switch gears into that topic

 

Somebody find a padlock and lock this thread up

 

I don't think of it as a childish retort, more an expression of exasperation. I was just fed up trying to explain why I thought the two things (lift-off oversteer and lift-off rotation) are merely two shades of the same thing governed by the same physical phenomena, and that AMS2 seems to struggle to provide (I'd like to be able to tune the car to spin on a lift so that I can tone it down to the level I want, but when you can't get even a small amount of rotation on some cars (most of the ones I tried), then this is not possible). The discussion was made difficult by the bludgeoning style used with very little explanation and many absolute statements terminated with "PERIOD". I'm not a fan of the "that's it and all about it" style of debating, and there was really very little point struggling against this attitude, and it wasn't really even a debate.

I think some want to see this thread die, so let's let it do just that.

 

I dont want to get in between you guys "argueing" but it doesnt really matter in this case how someone interprets the terminology, the actual feeling/behaviour of the car is what we should focus on. Anyway, could you post please couple of examples of cars you are having trouble with? I have been testing this and there are cases where its obviously an issue, just want to compare.
Thanks

 

OK, as @Shadak suggests, let's move off the semantics and to reality.

Being able to rotate the car nicely is a result of having a balanced set-up. The elusive balance that every racer wishes they could get in every race car instead of some degree of under steer or over steer, often fluctuating from one temperature or tire compound or point in the stint to another making it even more frustrating. All race engineering is aimed at achieving balance so the car will be the most manageable (or balanced) across the largest range of conditions.

Here's something we can all do/try (whether a newbie or a set-up expert):

Take the Ginetta G55 Cup car to Brands Hatch Indy. Drive at least a dozen laps (they are short!) with the default set-up unless you are uber-familiar with the car and track combo already. What do you feel? When pushing the car, I experience fairly consistent and frustrating under steer...that is preventing me from exploiting the car fully and undoubtedly slowing me down.

Go to the set-up screen--we're going to stick with the grade school basic screen, not the university or grad school "Advanced" tabs--and notice something glaring about the anti-roll bars. We don't know every nuance of how the chassis and suspension of this car is designed, but assuming it is some form of normal vehicle (it sure looks like one from the outside), it is painfully obvious that the car is set-up for extreme stability and tendency to under steer just from looking at the very stiff front bar compared to the rear one...

Drop the front bar down two or three notches and then go drive some laps. Touch nothing else.

Report back whether you can feel a difference with that modest, basic change. Report how the car handles now. Stunning is my word for it. This is why Reiza has been tinkering with the physics for a year and a half--to get to this point where changes you make in the real world on a real car have the same effect in the "game."

Spoiler alert: what @steelreserv has been talking about and demonstrating in his videos is that it is all about set-up and that Reiza's set-ups tend to be tame/stable/safe/call it what you will. Imagine how this car could be transformed with something beyond the most elementary, 10-seconds of thought and effort, set-up improvement...

 

I was interested in m Collins comments earlier. Balanced in how cars arrive can be just as important as changing them on the day. At least in ams2 people will know how much they have to change. There's a lot of tracks after all

 

My long past aside, can people tell me if they think the situation has become a lot better or at least a little?

I drove the steelcast AMG GT3 at Bathurst today and it's a good drive. Felt it really needed to be driven like a car with minute inputs - it really felt like a Merc in spirit obviously not driven a gt3. And the Rocco had nothing to complain about really in terms of getting it round a track - interestingly reducing the ARB to 60 and making the diff be 55 and 35 (only) just like the AMG, it made a big difference in reducing its understeer and/or reaching a level where forces can be over come. Is loos really that big of a deal in the game just now? I think the debates lost me in which one was more true couldn't keep track of how it is overall. Also I put steering rack up to 20 from 18..everything else was default. Some of the steelcast type approaches.

I like it since you can confine the forces or unleash them. It's not a relatively short scale. Other games scale 2 5 7 9 but here it's 1 2 3 4 etc.

You can drive the Rocco so HARD around bathhurst now, didn't bother adjusting suspension beyond ARB at 60 and its got more grip than the AMG, so yeah...and there is a lot more detail in it swaying about as you drive it round, slipping and grabbing than ever before. More subtle/full complete and in multiple doses than I would expect from other sims maybe. It really felt low to the ground compared to the AMG in its movement with the physical forces, so thats only a tick in my book as to the debate. I am not so sure other sims can grab that kind of feel. Am I the only one who experienced that kind of thing? Note I used the amgs coast and power settings on it. I don't know if that's true to life for the Rocco or if it's the scale in the game but the turns at Bathurst liked it. Yes slower than nemesis haha but possibly the best driving yet in the title for the combination. It made coming first so very easy so would need to up the ai now.

Now I side on the side of the debate where I think LOOS is there anyway; but now in the rocco its definitely there, and if not LOOS - then its definitely Lift off "extra" steer. I like how it can be overcome with how you steer or invoked should you want it so its nice and nuanced; and this is with and without extra turn-in you may apply, so one can work the throttle way better and be in more control. It turned the car into more of a point and shoot machine I expect the little thing could be. This meant when holding down the throttle and wanting to remain in a particular gear it was a lot easier to place the car amongst the others, then change gears and advance the position. Yeah thats good. Like when you ride a motorbike and hold onto the gear, it was a good sensation to have in the race. So I don't know if that LOOS in terms of the debate, but how to say "its got way more avenues to race with than ever before"...on throttle, off throttle, managing throttle, or flat out, turning and accel or turning and decel or whatnot. Was really good.

To summarise: times when you wanted extra rotation and could do it under throttle - and times when you wanted to step off the throttle and there was a different kind of extra turn, which could be mitigated when back on the throttle. That was good and I think maybe theres more LOOS than some think, at least in that car(s). When on a motorbike its similar- you have 4000 revs to play with in a certain part of the overall rev-range and you can hold onto that gear for when you need the extra power...so when you get to a turn or just coming out of a turn, and really crank it hard, and then it turns even harder as you not only go faster but increase the turning angle/angle of velocity... in the past one would have to govern how hard to drive or you would drift around like there was an invisible wall in the game - so what I am explaining is it had all the features and transitioned between them almost? - And it could also be a knife and more purposeful.. as opposed to just a lump of metal I was hauling around the track.

Very nice. With that in mind, when you were close to the turn you could stay there or maneuvoure there, and when you wanted to be 'flung out' wider as you moved around cars you could accelerate and do that too. I don't think the game has ever been this pliable and with the under and over steer aspects of it too. And its not just straight up rotation, so I am interested to hear other thoughts on the debate.

Because about the debate I personally think it can be hard to miss since it's way more finely grained, it was not there one patch then it was depending on car, and there's more settings to achieve the level which may explain why.

 

3 patches ago I needed the custom ffb to load up the front feel coming down the hill at Bathurst. Then some was there a patch later. Always had to adjust race room ffb to the front slightly to get more wheel feel. It's there in ams2 with telemetry meter you see it shift forward. The steering fundamentals change and the car is lighter in the rear... You probably do not notice because you're doing 100kms an hour down a hill, but you DO notice the whole thing when you spin out.

It's definitely there. Focus on the breaking and speeding down the Bathurst hill, with the lighter rear and loaded front turning (suddenly, important distinction) becomes much faster and breaks get tested. But weight transfer is not as apparent in a wheel.

Not sure about actuals but there is a divorced feeling still maybe due to that where AC might shove it in your face. But the car behaves well enough, it seems like the wheels are loaded, and it's consistent through braking. If you brake too hard it gets lighter but what I think people look for is the turning getting harder in ffb feeling. I'm not sure at high speed this would be an actual thing. So reiza probably have it right. We can not simulate the feel of the tyre friction as it turns so closely, think about the friction being overcome. One would need a movement rig.

My rear wheel drive car is fairly powerful for a road car and I much prefer turning and going faster down a long incline turn as the front is loaded up and the rear is held in check, because the cars systems are handling keeping the car in balance, so the mechanical grip of the wheels is more due to the shift in weight. If that hill levels out at the bottom after the turn, the weight goes back and the wheels at front lighter... Only then does the car feel lighter at front and by that point is not using the lsd but the stability control system as the turn straightens out. I could try that hill and turn one time with the systems turned off I bet you the wheels would slip .. But the transition is smooth.

It's probably not in your head, it's possibly more to do with the ffb not replicating that. But as of this patch there is definitely at least forward and reverse weight transfer because the cars go down hills with the right setup so much better. The throttle control for the AMG GT3 and Rocco down hill is great. I didn't turn on telemetry when racing but the turn feel was a lot better, just that I don't think the custom ffb was necessarily making it seem that way.

I did slide sideways over the mountains top run off strip, felt the ffb was amazing and felt the car holding and the weight was DEFINITELY there on to one side, the car slowly spun back to the correct way and I drove down the hill at Bathurst.

Maybe it's just not pronounced but neither is the feeling of going down the hill too light.

I bet it's definitely there but maybe more would be done on it? The small jump hill on the front Bathurst straight is definitely a thing. Also the uneven Bathurst road makes the car shift at the steering wheel.

It could be the next step for reiza I'm not sure. A game like AC simply over emphasises the weight for many cars I would think. Ams2 does have it in there when things go wrong or otherwise. But it's not pronounced. I would think at 220kms an hour in a car and I've only done 180 lol on a motorbike the forward weight did not make steering hard in my experience. In the car with no stability control it made things skittish but that was not a race car only a sports suspension.

On a motorbike the steering conversely was harder owing to wind resistance even while leaning was easier. With 4 wheels cars are so much more stable. I also think it's hard to quantify in game as so much aero differences. I will guess that 4 wheels in a race car would have better forward momentum rather than weight shift as far as feedback. Not sure if this explanation is right. But forward momentum at speed makes up for so much or over comes a lot and I know ams2 has that pretty right for those cars I mentioned. The rotation with the throttle seemed pretty good in order to position the cars in and out of those turns.

So referencing your post those cars had enough about them to turn as needed in various ways, to work over the under steer, so under, neutral and over. It was good to massage it out of the car. Other games it's just one style maybe. The ams2 throttle control was so much better imho because of it, could do it purposefully.

 

I saw a video this morning showing proof of lift-off oversteer in AMS2, and it involved setting the car up to oversteer as much as possible in a general sense, a full throttle lift before turn in followed by cranking the wheel rapidly in an unrepresentative manner. It was quite a funny video to watch (and reminded me of the accident report I read of an Aeroflot Airbus (I used to work at BAe, and some of the accident reports were quite a read!), where it said that child-like inputs were evident, I recall a result of the plane being flown by the pilot's son).

This video would have been more useful if done with a setup tuned to set the fastest lap time possible and then show the mid-corner lift-off handling characteristics of each of the cars using a long bend, the throttle being adjusted up and down to show it transitioning from pushing wide to tucking back in. Once this has been shown, move to ever larger throttle adjustments until you get to a full lift (and possibly a loss of control - true LOOS apparently). A test track with concentric circles would be good for this.

If the car in this state is still unresponsive to throttle lifts, then this should be addressed in a manner that doesn't ruin its general handling, and if it proves difficult to do this, then you have to question why. I got to this stage and was left scratching my head.

The cars I tend to like are the P1 (AJR), P3 (I tend to drive the Duratec, I don't know why), Brazilian Stock car (I've only tried the older one) and a few others I can't remember right now (I should be more methodical and record which I like/dislike). The ones that I didn't warm to (in terms of lift-off response) were the FV10, 1967 F1, BT44, BT46B (quite a peculiar car), and again a few others I also can't remember. I drove these cars at Guapore, which has some nice long bends, so if you want, load my ghost and and fiddle with the setup until you get to handling nirvana, and if you do then let me know what you did. It is possible that it is all setup related and the physics are bang on, or something could be amiss, perhaps the differential is locking in coast too quickly, I don't know, but I'd like to.

 

Quick question: do you happen to have onboards of these cars demonstrating off throttle rotation or on throttle understeer (which is what often creates contrast)? The reason I ask is, because I watched a couple GT3 and GT4 onboards and I was surprised by how throttle oversteer biased they were. Off throttle rotation was more neutral than on throttle rotation. And even then you gotta figure that they are likely trail braking. In one video the GT3 driver is trail braking while throttling (~15%) at Magny Cours T3.

Of course this doesn't apply to all cars. Copa Uno has liftoff oversteer (real Formula Uno races are nuts, I recommend checking them out on YT). Caterham Academy has some off throttle rotation. Formula Vee has a lot of liftoff oversteer. Group C has power understeer at times. And that's my point: you can't state that all cars have this dynamic. Whatever some road car does has nothing to do with a race car especially when it has downforce that affects grip way more than a tiny shift in weight from lifting. Whatever one type of race car does has nothing to do with another type of race car.

Just to be clear: I'm in the camp that believes that the clutch LSDs are suspiciously sticky when coasting. You can feel how it snaps open rather than "slips" here and there. Smoother operation might give the perception that you have more liftoff rotation. I dunno if how it works now is actually correct though. But this is a minor thing for me since hairpin handling and trail braking was improved a lot. Although the low speed diff stiction is still pretty nasty in some specific cars (whatever is actually causing it in the driveline including tyres and suspension).

 

I don't have any onboards, but I can get some for you when I get a moment. I'm pretty busy with league racing, and funnily my next race is GTE at Magny Cours, and turns 1, 2 and 3 can require line changes using throttle lifts as they're all interlinked (you don't want to be too far wide after T1 as you can't get tight to T2 for a good entry into T3, and T3 starts open and tightens a little (not sure how accurate the track is!), so you end up gassing it before easing a little and gassing it again). All this felt nice and comfortable in rF2, and with the lowest wing settings didn't even require a lift from T1 to T2.

I agree, some of the cars in AMS2 handle really nicely, while others don't, so it's quite variable (like the quality of mods in rF2 I suppose, but with one author for all the cars in AMS2, you'd expect them to be less patchy). I haven't said they're all bad and I hope I haven't given the impression that they're all bad, as I know there are some good ones!

Some cars seem spookily easy to control with excess power. The person braking while throttling was probably stopping the rear from stepping out because doing this effectively shifts the brake bias forwards (if you really want to scare yourself, do this in FWD car, as it does the reverse!). Either that or he was being lazy with his feet!

 

And the hammer drops again to squash the revolt - ironically Soviet!

 

I really think most pre existing sims work backwards from a perfect racing line without physics, then add in physics. A standard is made before any values are added. Generic values first.

How any one can believe ACC for example with it's simplified physics needed rocket scientists to work on it when the madness engine simulates a lot more is beyond me.

I mean sorry that they claim so. They did good work on that game, Kunos is very open.

I'm pretty sure having seen the madness engine YouTube videos and listened to interviews that over time the SMS people made the engine to be complex

Complex nonsense incoming. Think about ue4, bullets and vectors. Which includes velocity and angles. Imagine bullet fall off calculations. But make them horizontal only.

Think of a perfect driving line and simplify it's programming round an artificial course.

Next think about just one car class and adjusting it on a detailed enough but simplified spectrum to illicit different behaviours. And you have the beginnings of ACC in unreal engine. Very simplified. And you can see why Kunos went for what they did.

When I watched videos on how SMS did the madness engine I only have an incomplete understanding of them.. but they spoke to drivers and mapped individual systems into the framework that rather than working with any single or few reference points, they worked with individual systems meshing independently. So they did it a little different befitting a multi car sim. They added in complex things which they could get a hold of, and simulated more and more within them, it was just never teased out of their games and car setups to be visible.

It's almost as if there's two opposite ways of going about it. But we can consider madness to have 2 layers as a conception point.

It's no wonder ams2 took some time to tease this out. Because they're adding values and also going into things SMS seemingly didn't go into. I'm guessing SMS had a change of heart half way through their project.

In one project they shrink down a fixed almost ultimate value within one car class and it's pretty darn good.

In the other project they need to add complexity and it may even have multiple paths by it's end to arrive at a similar on the surface but more difficult end result.

In two cases they determined not to deal with any extra fluff. Just one car class and simplifying physics.

Reiza have chosen to deal with some of that stuff is my impression and they should be commended for it. But they may not like the praise as I am naive, moreso they may have indicated beforehand that they're up to the task and will take the time.