Kazunori Yamauchi: Driving Simulators Shouldn't be Difficult
- Thread starter GTPNewsWire
- 270 comments
- 22,559 views
- 48,709
- Australia
In regards to the physics, How can anyone playing Gran Turismo, be able to match the Nurburgring time of the '07 GT-R, which was set by a professional driver?
It would be like me trying to match that time in the real world, by myself. I doubt I could possibly do it, even if I had coaching like this:
There are players who may know the 'Ring like the back of their hand, put them behind the wheel and they might not be able to match their virtual skill.
It would be like me trying to match that time in the real world, by myself. I doubt I could possibly do it, even if I had coaching like this:
There are players who may know the 'Ring like the back of their hand, put them behind the wheel and they might not be able to match their virtual skill.
- 28,471
- Windsor, Ontario, Canada
- Johnnypenso
A good sim driver should be able to beat that time by several seconds in a game with fairly accurately modeled physics. In a game you can take a chance riding curbs you might not take in real life. You can use more of the road without worrying about what happens if you push too hard and get a wheel into the grass. You can make endless repeated attempts to get an entire lap perfect. There are several advantages to driving in a game that real drivers don't get. Biggest one of all is you don't die or injure yourself when you make a mistake so you can push the limits every second of every lap.
- 219
Wrong. Simulators assume things. FFB fills gaps. There are many reasons the real car may not match the virtual one (weather, temp, wear, compound, setup, track changes, atmospheric pressure, driver style).
The facts absolutely matter, but just as important is how the games fill the gaps. Otherwise no game could be called a sim until 1:1, which is obviously not the case.
Unless of course you can point me to a game that gets all the facts right in all situations?
Well while you are not completely wrong what you say sounds (reads) very biased towards a 'feeling'/'opinion' based creation of a simulator.
And it is this that is the problem. It deminishes all the work put in by engineers the countless of hours of datamining their insights...
When their work is 'done' many people drive this thing and give feedback. Because as you say no sim is 1:1 currently. By having a significant set of drivers you 'tune out' the outliers who have a very diffent experience from the 'pack'.
Then the engineers look at the impressions of the 'pack' and adjust accordingly.
In other words not a single opinion matters...
If you're an outlier no one cares about your feeling or your opinion. As your personal opinion is invalid to develop a sim.
The only 'opinion' valid is that of the pack. It's just like this: most people will prefer to eat chocolat over tar if someone likes tar that's on him, his opinion is irrelevant to the general statement.
So that mezns a personal opinion is useless, yet the opinion of the pack is usefull. As that pack opinion shows flaws the engineeringdepartment made.
Edit: where is this constant need for people to try and validate peoples opinions as usefull facts to develop something?
Edited below as my expression was deemed abusive
Development comes from facts, please be less stubborn and try to accept information when passed to you... (This is a general statement not directed to a single person but all people who refuse facts because it doen't fit their worldview)
Last edited:
- 219
Indeed. There wouldn't be so many varied opinions if 'feeling' was not a large chunk of how people perceive (and argue about!) racing games
For instance, it's a fact that Project Cars simulates more physical elements and data than AC (at the time of PC1 release anyway), GT and Forza, but I think you'd agree stating it is more sim would not see agreement from all, not even close.
It's a matter of how much 'close enough' games get in fooling the mind. And there's the catch; I think it's not how many things these games simulate well that matter, it's the things that break the immersion that determine the opinion. It only takes one odd behaviour (physics bug, FFB, controller) to remove someone entirely from feeling like they're in the car, regardless of all else, and so the physics 'fail.' The game that can be most consistent (i.e. AC's praise) tends to be more convincing than the one that almost gets there, but then has a bug/issue that breaks it (i.e. PC1).
That's a very context sensitive statement. What if in 50 years tar changes to somehow actually taste better for the masses? What about racing games from 10 years ago, they don't 'hold up' now, yet people would argue endlessly about 'most sim' then, as they will for future racing games.
See the facts don't matter all that much because we're not close enough to simulating everything yet to determine conclusively the 'best' simulation. Hence the pack, as you say, which has not much to do with facts. I.e. 'the pack' is playing GT/Forza, not iRacing and rFactor
And Kaz isn't helping at all by making blanket statements (he's not alone in that) about driving being easy or hard, which has nothing really to do with what a game is simulating or not.
Indeed. There wouldn't be so many varied opinions if 'feeling' was not a large chunk of how people perceive (and argue about!) racing games
For instance, it's a fact that Project Cars simulates more physical elements and data than AC (at the time of PC1 release anyway), GT and Forza, but I think you'd agree stating it is more sim would not see agreement from all, not even close.
It's a matter of how much 'close enough' games get in fooling the mind. And there's the catch; I think it's not how many things these games simulate well that matter, it's the things that break the immersion that determine the opinion. It only takes one odd behaviour (physics bug, FFB, controller) to remove someone entirely from feeling like they're in the car, regardless of all else, and so the physics 'fail.' The game that can be most consistent (i.e. AC's praise) tends to be more convincing than the one that almost gets there, but then has a bug/issue that breaks it (i.e. PC1).
That's a very context sensitive statement. What if in 50 years tar changes to somehow actually taste better for the masses? What about racing games from 10 years ago, they don't 'hold up' now, yet people would argue endlessly about 'most sim' then, as they will for future racing games.
See the facts don't matter all that much because we're not close enough to simulating everything yet to determine conclusively the 'best' simulation. Hence the pack, as you say, which has not much to do with facts. I.e. 'the pack' is playing GT/Forza, not iRacing and rFactor
And Kaz isn't helping at all by making blanket statements (he's not alone in that) about driving being easy or hard, which has nothing really to do with what a game is simulating or not.
First off I hope you take the time to read this, as you seem to have a mind that processes things in a logical way.
This makes the discussion interesting.
I think we might have a communication error.
What most 'feelings are important guys' will never understand is that, for about the 8th time in this thread, yes personal opinions on set things differ due to their perception of reality.
I really believe you showed why some people find some sims more immersive then others. And my statement that opinions don't matter in developping a sim do not reflect on wether peoples opinions matter in general. Off coarse we can have lengty converations about these simulators and how we perceive em.
BUT stop making the story up that it's development is very feeling based. This is NOT how simulation models are made. This is NOT how you'll ever come to a good model.
The tar statement? Context dependend? Well yes! Off coarse, when you discuss a certain subject the context defines exactly what we are discussing. But just curious, why would tar ever taste better?
Those old games/sims are outdated, yes. Yet we can still argue which one did it best and opinions will differ as some people will have break from immersion due to a certain 'factor'. Don't see how this is a problem? We agree they are worse simulators but can still compare them to their own generation.
The facts matter a lot! Even in defining the best simulation. People could argue and have personal opinions but the facts show this generation is more realistic then let's say the ps1 generation. So facts do matter, it's when I or you personally define a sim to be the best that mine or your opinion matter.
Also the reason for this diffrence of opinion is rather clear. We didn't set up what was the goal to sthe simulator. If we set a goal thise simulators need to reach we could objectively say which sim is better. Yet we never do. We all fill in the spots where the sim has to be 'perfect' this is out opinion of the perfect sim. This in turn makes us put one sim higher on a pedestal then an other sim.
About the pack statement. I'm specifaclly refering to the pack of people who are actively helping to develop the specific sim. So this gt/forza thing is irrelevant. I meant the general consensus amoung the testers of said specific sim.
Any thoughts are definetly welcome if propperly motivated as on your last post!
But yes when judging a sim opinions matter when developing a simulator facts matter. This while the other feeling guy told us it isn't, he said it's all perception. That would be a lie.
- 28,471
- Windsor, Ontario, Canada
- Johnnypenso
The world is not black and white. It isn't only facts that matter when developing a simulation style of game because, even if you get the game to simulate physics perfectly, those physics are being translated through imperfect devices. FFB wheels cannot perfectly provide all the feedback that a driver would get from a car, because some of the feedback is through his body, through g-forces etc. So some of the effects that a driver might not normally feel through the wheel are there, like enhanced road or kerb feel. Some games add in a "seat of the pants" oversteer effect in the FFB. To that extent, real driver opinions and feedback are the only tools available to align the feeling through the wheel with their real life experience. This does not in any way change the underlying physics of the game, which is the factual element you are referring to. We can evaluate the level of a simulation through the relevant telemtry data and by looking at the underlying code, but games will also be evaluated by players based on how those physics are translated through the devices we use.
In the bigger picture, both facts and opinions matter when looking at games as a whole.
ItsHim
Premium
- 1,315
- Bavaria/Germany
I track a 96 Celica GT-Four IRL. Been to Spa a couple of times and Spielberg once.
Things that hold be back on the track:
Wear and tear. The fact I'm actually driving a 20 year old, 170.000 km car. Physical fitness. Changing track conditions. Heat buildup. Traffic.
Things that hold me back in GT6: lack of depth perception (eased a bit by 3D glasses). Worst of all: lack of G-Force. In GT6 I go by (engine) sound instead.
Considering everything driving IRL is much easier than in GT6. I'm about 20 seconds slower IRL on Spa than in GT, but that's due to shifting earlier and braking sooner to reduce wear and keep the temps down. Corner speeds aren't far off.
Things that hold be back on the track:
Wear and tear. The fact I'm actually driving a 20 year old, 170.000 km car. Physical fitness. Changing track conditions. Heat buildup. Traffic.
Things that hold me back in GT6: lack of depth perception (eased a bit by 3D glasses). Worst of all: lack of G-Force. In GT6 I go by (engine) sound instead.
Considering everything driving IRL is much easier than in GT6. I'm about 20 seconds slower IRL on Spa than in GT, but that's due to shifting earlier and braking sooner to reduce wear and keep the temps down. Corner speeds aren't far off.
I tried to write only a few lines of text about these things. It is not possible. You must study everything.
You can start learning by reading these:
https://en.wikipedia.org/wiki/Computer_simulation
https://plato.stanford.edu/entries/simulations-science
- 28,471
- Windsor, Ontario, Canada
- Johnnypenso
ItsHim
Premium
- 1,315
- Bavaria/Germany
A conservative lap is around the 3 minute mark on Spa in GT6 and a swift but safe lap IRL around 3:20.
I was planing to do a side by side comparison video, but at the moment can't be bothered by the technical side of capturing video games.
On Spa fuel consumption is around 1 litre/minute. A set of rotors and pads 300€, a set of tyres 400€.
After 7-8 laps the brakes start to fade slightly. The whole car gets hot like seriously hot. Not only engine oil but diff, clutch, water. You can actually feel the heat coming through the tunnel. The body starts to ache, the mind gets tired and I'm a bit of a lazy fart so I feel the physical stress.
I'm not made of money and i must drive the car 800 km back home. So I shift at 6000 rpm, brake 50 metres early and don't go flat every time on the Camel straight. I'm very cautious through Eau Rogue because if I screw up, given my talent, it's game over for real.
Ad this all together and 20 seconds is actually quite close as I lose mainly under braking and on the Camel straight.
A track day is 8 hours and given the insane cost, i maximise fun time on the track.
So yes, to me GT6 Spa is a lot harder than IRL Spa because there's no feedback from the tyres, my body gives no feedback, there's no feel in the brake pedal and the FFB of a G25 is nothing like the real deal.
I really understand what Kaz is trying to say.
- 219
Ah sorry I see you were referring to development of simulations. Yes I agree, the facts are very important and having real data respond accurately is an ideal confirmation of simulation.
In terms of testing, other than telemetry data and other mathematical verifications, it still comes down to opinion unfortunately because of the gaps (particularly behaviour over the limit) and even in PC1 we had Ben Collings and Nic Hamilton offer opposite opinions on the same virtual car and they're accomplished racers.
Overall I think the developers while increasing the complexity and realism of the models have the biggest challenge not in matching data, but in preventing the wrong assumption ruining the experience.
Oh and tar: http://africansinfinland.tdv.co/2011/08/finnish-food-tar-ice-cream
- 23,800
- Philippines
There's a racer I respect who tracks a Mazda2 in competition locally. He's much, much quicker than me on track, and often has two wheels off the ground in corners, as he's learned to extract the absolute maximum out of the car. I don't know how close I could come to him IRL, but I know that if you put him and me in the same car around that track, I'd get my ass kicked.
But on GT6? He can't even get out of the blocks. He's hopeless on it (and that's with a full play rig with a Logitech and a big screen). It's different, of course, for professionals in modern teams, who've grown up on simulators. Those guys are wickedly fast at both. But a simulator demands acclimatization and a different skill set from track driving.
Eliminating that gap, bringing the "organic feel" to simulators, to allow any driver to jump right in and drive as they would in real life. Kaz is absolutely right. The driving simulator itself should not be difficult to master. It should facilitate easy transfer from real life to virtual and vice versa, to allow you to concentrate on learning more advanced skills.
Sorry, I wast clear enough. I agree this is a simulator and this how it feels is important. And the opinion of multiple drivers is nescecary to tweak the feeling through the imperfect hardware. This was what I meant with the 'pack' statement. As a general rule the opinion of multiple testers is valid the opinion of 1 tester is next to useless if not within the context of other opinions.
So yes opinions are a tool and it's not just fact, but the way they process those opinions is more then opinion and is based on facts. (Or should be based on facts)
- 182
- con-sistant
Project cars is a good sim. as far a driving, and the way the vehicle drives when its actually moving. my biggest problem with project cars is the tire physics, when you lock up your brakes, or go into a spin. the vehicle behaves so unnatural. but in driving, their isn't much of a difference tween Assetto and Pcars. my other big problem with Pcars is the subjective feel to the wheels settings. if you don't have a good settings, it just feels awful.
and because you actually have to go through this to make the game feel right. is why IMO, i feel that the physics in Pcars is subjective to how an individual likes it. and doesnt necessarily simulate what that car may actually feel like
i had many complaints about the early version of GTS beta. covering almost every aspect of the game from sound,suspension,tire physics ect. i even stopped playing it after a week or so. and played it on occasion when new version where released. and sure enough with each updated version the game has gotten so much better. the last version being the greatest improvement overall.
the sound is perfect now. just need to increase the volume in general, to get that deafening sound you get from Pcars
the gear whine is subtle now and not preset among all vehicles. IE transmission whine is only heard when you install custom tranny.
tire can be heard now, needs a little more adjustment as it may be to loud
tire and overall physics has dramatically improved. especially in the tire department. however the suspension physics needs some work. but none the less, much better than it was.
as far as wheel, they finally fixed the back and forth sway of the wheel that plagued every gran turismo title. wheel feel, is almost identical to Assetto.
- 12,300
- paracuru
- praiano63
I m not a specialist and i don t have any notion of the real sensation to drive a racing car. BUt, if it s possible to do this in real life, you ned a permissive game physics that can allow some close racing contact. This is not possible on AC or PC for example, any touch like this will have a lot more dramatic consequences or at least a lost of speed and grip.
- 12,300
- paracuru
- praiano63
Outside in a turn, at least you loose your line, your speed and your position. From what i have experimented with both games. Just my opinion.
EDIT: Just a sample for a PC game crazy race, PS4 seems even more unstable, you can t pass over low kerbs like this staying inline this way.
Last edited:
- 15,067
- Melbourne
- ScottPuss20
- CheetahsMeow
- 1,820
- Ireland
- turnupdaheat
A bit of a thread resurrection here but it was interesting to hear @IanBell making comments mirroring Kaz's statement during SpotTheOzzie's Youtube livestream today. He said driving sims "should be too easy" and that they are "too difficult". He also mentioned that the racing drivers, or key drivers as he calls them, that SMS consulted were of the same opinion regarding PCars.
According to Ian, "pleasing the hardcore sim guys makes the driving less realistic" and the fact that "sims are harder than reality doesn't make sense". He also said the in order to placate the hardcore community PCars1 ended up being "too harsh".
According to Ian, "pleasing the hardcore sim guys makes the driving less realistic" and the fact that "sims are harder than reality doesn't make sense". He also said the in order to placate the hardcore community PCars1 ended up being "too harsh".
Driving a car isn't difficult, and arguably even easier with modern race cars thanks to TC, ABS, paddleshift etc.
The difficulty comes from extracting that final second (or two).
PCars was unnaturally difficult with snap oversteer, Assetto Corsanmuch more progressive, reminding me of GT Sport.
The difficulty comes from extracting that final second (or two).
PCars was unnaturally difficult with snap oversteer, Assetto Corsanmuch more progressive, reminding me of GT Sport.
- 28,471
- Windsor, Ontario, Canada
- Johnnypenso
A generic statement about how difficult it is to drive racercars is pretty meaningless without a ton of context. A modern GT3 car is much easier to drive than a Lotus 49 for example, and in turn, a 200 hp saloon car is easier than both. All cars are difficult to get to limit but it would be easier to get a saloon car within 2 or 3 seconds of the fastest lap than it would be a Lotus 25.
A true sim should reflect these various levels of difficulty well. Any game that makes it relatively easy to drive all cars of various eras with equal ease wouldn't really be classified as a sim to me.
A true sim should reflect these various levels of difficulty well. Any game that makes it relatively easy to drive all cars of various eras with equal ease wouldn't really be classified as a sim to me.
- 4,806
- Sweden
I was going to ask for a link since I've never heard of that channel before.
But I went and found it myself. Here it is if anybody else is also curious of what was said;
Scaff
Moderator
- 29,025
- He/Him
- ScaffUK
In general terms it's (depending on the car) relatively easy to drive up to the limit and also to catch over the limit. The far trickier part is to find the limit in all parts of a changing track and to be consistently on that limit (which is actually with the tyres slipping.
It does of course also need a ton of context, as it assumes that tyres and brakes are up to temperature, that the weather is also consistent and settled, etc.
It does of course also need a ton of context, as it assumes that tyres and brakes are up to temperature, that the weather is also consistent and settled, etc.
- 15,529
- Cairo, Egypt
- GTP_SEMS
- 327
- ::1
I'm going to play PCars2 with the "authentic assists" setting, i.e. with only those aids that the real-life car does or does not have. (Except for clutch-shifting, since I only have paddles on my wheel).
Will GTS have such a setting?
- 28,471
- Windsor, Ontario, Canada
- Johnnypenso
