Discussion in 'General Discussion' started by LokiD, May 7, 2020.
That's not really what was said. I think everybody is aware of the fact that more complex physical models can produce wrong or unreasonable outputs when you put in the wrong numbers and that they require more development time. The same applies for empirical models aswell though, when you put in wrong numbers. What he didn't mention is the fact that adjusting the slip and grip curves just based on driver feedback is a very subject way of tweaking your tire and from my perspective this is equaly as complex if you want to do it in an scientific way (lots of different drivers, testing different scenarios, testing different parameters).
The big advantage of physical models don't get touched in the videos wich is a bit of a shame. When you have very reliable data and can construct a very accurate tire in let's say rF2, your outcome will just be better. And I think it's pretty safe to say that besides the disadvantages that got mentioned here, that one of the strong aspects of rF2 are the tires for the most part. This chain leads further to the general handling of the cars and the FFB. I just use a Logitec GT Driving Force, but there is no other sim that gives me the sensation to feel the tire flex when I hit the apron on an oval or if I scrub the sidewalk at Monaco 66 with the sidewall of the tire. And that's just with a low budget entry level wheel.
To sum it up: if we were just to call it all guesswork we just could stop producing and using simulations for good. I think the idea behind the whole sim thing is to drive technology, research and science forward instead of stagnating where we are. That's one of the basic human instincts. Going by your logic we would still be playing GPL and be done with it for the rest of our lives. Luckily the trend has shown that developers put in quite a bit of effort to drive the tire topic forward so that we can enjoy even more dvnaced products in the future. Just comparing AC and ACC in terms of handling and FFB is well worth a look
Maybe approximation is a better word? Still they are only going by what they think feels right.
I love the progression of these tyre models too. But sometimes more numbers doesn't necessarily mean better feel. Rf2 is also a great example of this.
I'm not saying Niels is making this error, but many people seem to be linking the current criticism of the iRacing tyre model (it isn't forgiving at and beyond the limit) with the pros and cons of physical vs empirical tyre models. Iracing's unforgiving handling is not because they use a physical model. There are empirical tyre models which are unforgiving and there are physical tyre models which are forgiving.
The area I think Niels is wrong about is the motivation for developers to adopt physical tyre models. This is Dave K's explanation for why he switched from an empirical model to a physical model.
Why not just use an empirical model and be done with it? Wouldn’t that be easier? Well, it would be easier to code up the model, but it’s much more complicated to tweak and tune it so it has the right characteristics in all sorts of conditions (different loads, pressures, temperatures, etc.) Empirical models work well when the conditions can be considered to be fixed, as they might be for a passenger car with the recommended pressures travelling at highway speeds and below. But they become unwieldy in the racing environment, with large temperature changes, pressure changes, aerodynamic downforce and high loads from high-speed, high-banked tracks, along with the need to model curb hits well, and so on. There are just too many different things to measure, and it would be too expensive in terms of tires and time to test them in all necessary conditions. A decent theoretical model, though, should give reasonable responses even when the tire is doing crazy stuff, which on a race track is a lot of the time.
Actually I think that's the exact opposite of what 'they' are going by. I mean teams (or individuals) striving for accuracy. Niels doesn't tweak by feel, he uses data. Same for S397.
Approximation is the perfect word for it. Because that's what models are made for. They try to approximate reality. And if you worked with models - I've done this plenty of times for science and research at university, where you do this alot depending on the field (economy, environmental science ...) - you will understand that every good model isn't just made out of guesswork. And the people who build those models are mostly well aware of their strong and weakpoints. It just depends what you want to achieve and wich model fits your requirements best and if people find a way to overcome shortcommings they update those models.
For example, you will have noticed that Kunos implemented their 5-point tire model in ACC, because cars just freaked out while hitting curbs, wich was and still is an issue in AC and in the early days of ACC. That's a pretty important aspect in racing sims that ISI and later S397 never had to worry about, because the model is good enough to cope with those issues. And the further the whole digital industries evolves, the easier it will get to get hands on more reliable data (quality and quantitiy) making physical models propably alot more reliable going into the future. Where it get's difficult is especialy stuff like historic tires, where good data is difficult to get. But you won't get around this with empirical models and getting reliable driver feedback in such cases isn't easy either.
But not to be negative, doesn't rf2 to have issues that the ac tm doesnt. Ie, pressure's and temps in some cases.
I think no tm is perfect and all do one thing better than the other. If these guys could work together and collaborate the ultimate tm would be good nah?
Hehe, if it was so easy they propably would have done this allready. But as with so many things in life people have different visions of what is more important compared to other aspects and how priorities are set. I also don't think that pressures per se are really an issue in rF2 even if people allways refere to it as being THE issue. Using the lowest possible preassures is a common practice in real racing even if people question it on a regular base and there is a reason why tire manufacturers and racing regulations give allowed minimum preassures for racing tires. The death of Manfred Winkelhock at Mosport is just one quick example of this.
The issue with the tires in rF2 and why preassures work as they work is the lack of environmental effects such as temps and the lack of tires that simply collapse or get destroyed due to stuff like underinflation, wear or debri - wich let's be honest isn't peroperly implemented in any of the sims. But as often mentioned, it isn't as easy as simply triggering a switch and that's it. Enabling ambient and track temps will require a huge amount of testing of what is allready there. If it was easy there was nothing to talk about and everyone could do it. In the case of Kunos for example, they have the luxury to only have to worry about a few sets of tires right now compared to their previous product or any other title right now (rF2 and iRacing propably being the most extreme opposites right now).
Straying slightly off topic...
I have a theory that iRacing are keeping their tyre model unforgiving on purpose. It's the same reason they have overly strict penalties for track limits.
If the cars/tyres are difficult to drive, then more people will spin/crash. If more people crash then they lose their iRating/Safety Rating. So then they need to get their rating back up, and how do they do that? Spend more money.
If the track limit detection is overly strict, then more people will be penalised for putting two wheels off the track (which is common practice in the real world when driving on the limit). If more people dip two wheels then they lose their iRating/Safety Rating. So then they need to get their rating back up, and how do they do that? Spend more money.
So after 8 years of criticism, do they: make the tyre model more realistic and risk losing money? Or, ignore the complaints and maintain a steady cash flow?
The endless pursuit of a imaginary licence and safety rating in a video game is something that has always baffled me.
Anyway, back on topic... Yes tyres are questionably the most difficult bit of any serious driving simulator.
I never tried Iracing, but I had the same thought: if driving beyond the limit was somewhat "easy" this would disrupt the phylosophy of Iracing itself. Considering the consensus that tire models on the fringe of grip limit is a very complex matter, they probably decided that fitted better their narrative a more punitive approach. On the contrary the risk was that this limit could be exploited in a unfair way... without reducing everything to a matter of money, that I don't think they lack of. I don't think I can blame them for choosing, they built a structure that work on pilots behaviour, and good pilots are those that can stay consistently a hair below the final limit, those that stay a hair above often lose more than they gain, and crash too often. RF2 if full of very fast people... for a lap or two... but what is the purpose if you don't finish your race?
Beside this in this period IR is 50% discounted.. I'm somewhat tempted to try it but considering that I keep coming back to RF2 for FFB I fear it would be a waste of GB's (downloading 25GB for my connection is not a quick task).
On the topic. To complete the circle. Evaluating car physics by feel is also guesswork, and all of us do that.
@dylbie Had similar thoughts of iR trying to understand why it is so. It is also very likely that choosing a tire that is insanely hard to master, but yet with super high performance peak creates a system where simracer has to spend ton of time to master it, and time is money.
Here are some words from another expert
Interestingy I came to the same point and just said to myself: let's try iRacing for a tenner and see how it's like so that I can make up my mind for myself. And I got to say the entry level stuff has nothing to do with ice racing or any of those myths. It's much more people running out of skill or patience than anything else and it's def. not harder than AC or rF2. That was all my experience with three month in the MX5 races and I can just recommend that experience so that people see how the game works, how the FFB and everything else is. FFB is pretty much top notch with that car btw and you can have some good online races from the get go. Not saying it will be cleaner than anywhere else, because people have some wierd tendencies, but you just get races on a regular scale. And that's where other developers just s*ck. In that regard I got to say, that even sim raceway was further advanced.
So for the people who think that it is hard to keep the car on track in iRacing - it's certainly not harder than anywhere else.
I disagree, I found the safety rating quite frustrating to be honest and it actually limited me from buying DLCs, as you are restricted to compete against cars in your own license class. There is almost certainly more cash flow to be had from making the cars easier to drive, simple because you will attract more beginners. This is precisely the reason why games like CS:GO etc. are popular in esports, they are easy to jump into, but hard to master. I often hear the theory that sims are made harder to drive by purpose, but I never found any good reasons for doing that. It just seems like something somebody once mentioned and now it became some dogma repeated all over without any evidence to back it up.
iracing for me was the nearest of all sims that makes you feel like a seat driver on a grid in actual real life!
Yes the points system can be unfair but its a great experience. Certainly better than any ai can offer.
So what is rf2 doing trying to chase the online experience, its nowhere near. or even the offline experience.. its just stuck inbetween!
But no wait... the tyre model is SUPPOSED to be the best.?!
Ok, at this point, after the various discussions on this (tyre model philosophies, potential, shortcomings), I can only assume you don't want to say or think anything other than what's already in your head. Hence posting this video and the incorrect thread title.
Act like a troll, get treated like a troll.
Report & Ignore, it's the only way to weed out them out!
Hey isn't that what you all do. Only see a fan boys point of view?
And in my opinion it is still guess work in a round about way what Neil's said. So no thread title isn't wrong, your perfect assumptions of rf2 is what's wrong!
I thought you loved rF2 enormously, despite these countless flaws.
For some time now, you have given the impression of hating rF2.
It's a shame but it doesn't matter.
Just a word of advice :
-> If you hate rF2, don't waste your time here.
Too bad you no longer see the immense qualities of rF2.
Despite the many flaws in rF2, no other simulations have the necessary qualities (as rF2 has) for me to use them.
IMO, rF2 is a monstrously frustrating simulation, ..... But also and above all monstrously good !!!!!!!!!!!!
Loved Niels' recent videos. I actually didn't know about the distinction between rF 2 using a physical tyre model vs the standard empirical one. I just assumed they ran an empirical one with a ton more simulated contact points underneath the tyre. Other people have suggested that rF 2 runs a lot of its data 'offline' and still uses hugely complex look-up tables to get data, so it's only a semi-physical model. Does anyone have any clarification here? I'd like to actually get the truth for once!
All in all, it explains why the moment upon playing them both rF 2 and BeamNG became my favourite car sims instantly. All the things I was missing in AC, which I thought were due to my lack of skill, but actually ended up being due to lack of information and resolution in the driving model, were so quickly resolved and I immediately began to connect with the virtual cars on another level.
If we're talking about the way into the future, BeamNG is clearly the most ambitious car sim, and if they pull it all together finally they will make an engine that is unstoppable. I truly hope that happens. I want to see developers license code which can not only render things such as force transference and chassis flex naturally as a byproduct of physically modeled car structures, but also render accurate, unforgiving damage, so for once in our lifetimes we may see something resembling a realistic rally game!
PS. While finding ACC's mandated jump to the 5-point tyre model to deal with curbs was amusing, I found the 'chassis flex' patch equally amusing. Those two things showcased that those sims are just full of the canned effects some of us always suspected they were. A proper tyre and chassis model wouldn't need these 'added on'. They would be emergent effects springing from the physics engine itself.
Separate names with a comma.