The online racing simulator
An Open Letter to Sim Racing Developers
An Open Letter to Sim Racing Developers

A report on the accuracy and short comings of current sim racing titles and recommendations for areas of improvement.

In recent years enhancements in physics have transformed the gaming landscape, we've moved away from arcade games to home gaming with complex simulations with an emphasis on realism.

In regards to sim racing I feel the quest for realism has been grossly misplaced. Simulations are now very detailed realisations of the mechanics of a motor car, with suspension geometry moving and reacting realistically and considerable detail going into areas which effect a vehicles performance. Most noteably the work goes on improving the accuracy of the contact patch simulation (the area where the tyre touches the tarmac).

In terms of vehicle physics the interaction between the tyre and the road is one of the major significant areas for accuracy simulation, along with aerodynamics and engine performance.

However all current simulations focus on the physics of what happens on the cars side of the tyre. All current simulators treat the environment as relatively static.

This has created an atmosphere where recreation of the perfect lap is possible, lap after lap, with clinical precision and monotonous recreation of identical control inputs the perfect lap can be achieved, and there is such a thing in sim racing as the perfect setup - or atleast, close to it.

In real racing such precision is impossible to achieve because the environment is dynamic. The biggest physics factor in real racing is the track surface temperature which not only continually fluctuates throughout a day at a race track, but changes lap to lap, it changes across the width of the track, and it changes as one traverses across the lap.

No current simulation accurately recreates track surface temperature.

Low grip situations are also currently grossly overlooked, whilst some simulations have introduced wet weather driving, the realisation of these weather systems is still rather too crude, as in the dry on a real race track low grip situations still occur with such nuances as marbles on the track, and broken tarmac usually offers more grip when the surface is a little greasy under trees in autumn as the falling sap ebbs away better than on a smooth surface.

In real racing the dynamics of the environment result in drivers continually altering their car setup, such as taking tyre pressure out and adjusting front anti-roll bar torsion near the end of the day.

Trackside objects dissappear from view and are reset, stricken cars removed from the circuit. These are atmospheric killers which in real racing would be dealt with under yellow flag conditions or a safety car.

In short, the current focus on accurate simulation of the car itself is overshadowing the far more important area of environment simulation.

It should not be possible to recreate the same control inputs lap after lap to achieve great lap times. Real racing drivers are continually reacting to changes in the environment. Wind effects cooling rates, more than it effects straightline speed. A seering hot day with a gentle breeze is lap record conditions in real racing, in simulated racing midday without wind is ideal conditions.

There currently is not enough focus on gameplay dynamics, with better handling of environmental repairs (trackside objects and stricken cars) will come a change in driving mentality. With better modelling of environmental conditions will come a change in racing mentality. Both of which will bring sim racers closers to racing drivers.

Go to watch any single class race with cars that have downforce in club racing and you'll notice that lap after lap most gaps between cars remain at a fairly constant distance, even when positions swap. A few cars will pull out a small gap etc, but throughout the field there is a modal gap between the cars (which varies at different parts of the circuit). This occurs because of the effects of downforce. In sim racing field spread is much greater, which highlights that the aerodynamic simulation of the physics currently lags much further behind than the mechanical areas of simulation. With drastic effects on gameplay dynamics.

Currently sim racers focus on the perfect setup and recreating the perfect lap as often as they can. Real racers focus on continually adjusting their setup to chase perfection whilst reacting to the environment. Sim racers experience high field spread and enjoy few consequences in recovering a stricken car, real racers enjoy closer racing.

For sim racing to improve noteably from this point more emphasis must be placed on simulating environmental aspects (in dry conditions, not simply adding rain and a shiny/reflective road surface) and the dynamics of actual gameplay.

We must see the inclusion of better retirement handling - stricken cars sitting out a race by the side of the race track and their competitors bound to the pits, the introduction of a safety car (with automatic and manual control), trackside repairs being handled properly, yellow flags not being waved so wantonly that their effect counts for nothing and people do slow up.

Physics simulation should address current short comings such as aerodynamics and environmental conditions before worrying about competing with other simulations on tyre and suspension geometry which are currently over-simulated and the sole focus of sim racing physics. Too much emphasis has been places on tyres and suspension in the past, whilst not actually realising what racing is all about.

When, as a sim racer, i'm forced to react to the same things that a real racing driver must react too, then I am playing a simulation.

For as long as sim racing focuses on tyre and suspension modelling it is nothing more than a vehicle simulator.
Nicely written. However, while I agree with many of your points, I think its more important to get the car dynamics right first, and only then, the environment, as much as it sucks to see stuff disappearing and perfect weather conditions every time.

Thats the direction LFS is heading for. When? Only Scavier knows
Yeah, all very good points.

I'm in agreement, immersion is what sim-racing is all about to me. Do I feel (as far as it is possible) like I am racing for real?

When a car is on it's roof/immobile after an incident, why is there no recovery-truck? When tyres are hit, why do they disappear after a short time?

I agree with Becky for what my opinion is worth. How closely do tyre dynamics need to be explored? The environmental aspects of the simulation are equally important in immersing us in the race, suspending our sense of reality - they should be given equal attention.
I don't think that sim racing developers have been persuing the wrong thing. Without strong driving dynamics, a car will never feel like it should and environmental dynamics won't help that. I do agree though that top end sims have developed the tyre physics to the extent that they shouldn't be the focus anymore, and the perhaps even more difficult areas of aerodynamics and much neglected dynamic environments need to become the focus from now on.
#5 - Gunn
Quote from Joe_Keaveney :How closely do tyre dynamics need to be explored?

Much more closely, if current sim tyres are anything to go by.
I agree that a dynamic environment would be the best way to improve immersion, but I feel that the current interface between driver and sim is far too limited.

Suppose LFS could simulate a dynamic track: marbles, oil, weather, rubber buildup, etc. What good would it be if you can hardly see, hear, or feel what it does to the car? I expect most drivers would only spin out more often, and many spins would come unexpected because the driver never saw it coming. That would not increase immersion, it would raise irritation.

The average simracing rig consists of a decent FF wheel, a single 22" monitor and stereo sound. That's it. No force feedback system on the seat, because it's either affordable-but-botched or good-but-bloody-expensive. Pedals: ditto. Directional sound is not available in any sim, and most sims are sample-based. And a 3-monitor setup is considered top-notch, but is still far below the capabilities of your 2 retinas.

Until there are more ways to get feedback from the sim, I think I'll prefer a monotonously perfect track.
It's all been suggested before and if it not on LFS's development radar I would be gobsmacked.

Tyre simulation has some way to go yet though, but that can go hand in hand with a more sophisticated environment.
Quote :I expect most drivers would only spin out more often

I believe the opossite. Whilst there is such thing as a perfect lap drivers will try to achieve it. With a more dynamic environment you HAVE to drive within the limit, perfection is a variable that moves moment to moment.

The mechanisms which cause a driver to spin already exist, only currently we aim for a knife edge margin of error. With a less consistent environment drivers will have to give a larger margin for error which is infinitely more realistic.
Totally agree with what you are saying Becky but I think you're a little ahead of time on a lot of it. I guess with LFS it's slightly different in that the tracks are not real but with iRacing, for example, it's all to be done in stages. First part is to get accurate representations of the track and then as real as possible dynamics for the cars.

I'd love to see a development where temperatures gradually started to play a larger part of a racing game/sim. I think it will come but for me environment needs to come after the tracks are right and the cars feel right.

I remember when you were still running and managing the CTRA (those were the days!!!) and you had wind set rather high on FE2 and everyone was being blown off the track on the back straight. That was funny!
Quote :Much more closely, if current sim tyres are anything to go by.

The accuracy of the physics simulation is only important whilst it is the only aspect that is simulated.

The tyre physics in LFS are more than adequate. Anyone who argues they are not 'perfect' is clearly missing the point. There is no such thing as perfect in motor sports, there's always another thousandth to be found somewhere.

In real racing my car can and will handle totally differently at the start of a stint to the end. Therefore the way the tyres 'feel' is a moot point. The car is a constantly evolving platform, this is more important an aspect to be simulated than how accurate that feel is, given the mechanism by which we interact.
Nice post but imho you are a bit missing the point. First of all, there is still a lot to be done with the tire physics before you can say that they are "good enough". Tires is the single most important aspect of a racing sim and as such they just aren't yet good enough. Even in iracing the tires are very simple at the momenth even if they do best what they do.

I think it is very premature to suggest that the sim devs should take their eyes off the tires and start impelementing changing weather variables, dynamic track conditions and more elaborate engine and damage models. The tires just aren't good enough. What needs to be done is to get developers who want to understand the tires and teh data and are able to make a physics simulation that can handle it well. LFS has very good physics system as far as I can tell and the parts of the simulation that is a bit off is very much wip (heating/wear/longnitudal grip etc.). Just like iracing or nkpro. Even the ISI sims seem to have at least average good physics engine even if the data put in is something a tire engineer wouldn't wipe his... desk.

Imho, now it's the perfect time for LFS to make a stretch in the physics side of things. Better tires and more advanced aero with the bad sides in the tires developed better. As much as I see it, with LFS the destiny rests on few things: better tires and new tracks. What it already has is an active and dedicated community and large userbase with respectable online usage figures.

To make a good racing simulation you need to have excellent tire simulation. A race simulation with average tire simulation won't be anything more than average racing simulator. Dynamic track temperatures are all nice and that but the visible effect is just few percent unless you add moisture or start playing with different tarmac types.
Quote from Becky Rose :The accuracy of the physics simulation is only important whilst it is the only aspect that is simulated.

The tyre physics in LFS are more than adequate. Anyone who argues they are not 'perfect' is clearly missing the point. There is no such thing as perfect in motor sports, there's always another thousandth to be found somewhere.

In real racing my car can and will handle totally differently at the start of a stint to the end. Therefore the way the tyres 'feel' is a moot point. The car is a constantly evolving platform, this is more important an aspect to be simulated than how accurate that feel is, given the mechanism by which we interact.

While developing the tyres will make the game more realistic and enjoyable, developing the environment will do the same things at a much faster rate imho. I agree, this is what needs to be focussed on now. Though I have to agree with wsinda's post (if that isn't too contradictory!) When making a race sim from scratch, you should focus on the car physics first. But we already have very, very good physics. Sure, there is a lot to improve, and this should still be developed pretty heavily, but I think the environment is the way to go right now. But, again, wsinda...I'm torn!

Wait - I've decided! Chuck everything I just said out the window! The environment should be thought about a little more - 'retirement handling' as Becky puts it and a bit of track degradation - but, in the short term, car/tyre physics should be #1 in the 'To develop' list. Also, just to stir things up a little bit: "3rd party track mods".
I very much agree with the OP, sims are way to static to offer a well-rounded recreation of real racing...

HOWEVER, I do believe a good static model is needed first, as the dynamic environment may mask issues that need to be adressed...
I agree that in the end a dynamic environment is needed to properly simulate racing, however a strictly static environment helps greatly in identifying core issues with the physics that would otherwise be masked. Sure, a fully dynamic simulation would help hide some of the most blatant flaws that came to light during the years of hotlapping and searching for every way to exploit the system, but hiding problems instead of fixing them is not the way to go, IMO.

Besides that, that the tyre simulation is "good enough" might be true for other sims (I doubt), but unfortunately not for LFS. It is very good indeed, but still lacking in core aspects such as tyre heating, temperature effects, pressure effects, longitudinal grip, flatspots, etc.

Another thing I'd put before the environment simulation is a proper damage model. The amount of abuse LFS' cars can take is phenomenal and leads to a lot of the far too risky and gung-ho driving behaviour we see in LFS.
Quote from BurnOut69 :Nicely written. However, while I agree with many of your points, I think its more important to get the car dynamics right first, and only then, the environment, as much as it sucks to see stuff disappearing and perfect weather conditions every time.

Thats the direction LFS is heading for. When? Only Scavier knows

Pretty much agree with him.

However, the requirements hardware wise cannot allow many of these features to be propperly implemented for the majority of home users, because of the amount of processing power required. (JUST GUESSING)
Quote from Becky Rose :With a less consistent environment drivers will have to give a larger margin for error which is infinitely more realistic.

I agree fully with having more dynamic driving conditions and that that would bring about more realistic driving practises. I'm sure plenty have heard me whine on about it

But as to not spending more time on tyre physics I don't believe Scawen is spending much time on tyre physics he has only focused on tyres physics once or twice since I've been arround (early 04), so I wouldn't suggest he is spending an inapproproate amount of time on them.

While I agree that the tyre physics are pretty darn good within certain parameters, I do believe they need refining still. Which probably means that over the next few years another couple of month long stints of focusing on the tyre physics again would be a good thing. On the other hand I'd think introducing dynamic track/driving conditions etc would require a much larger effort.

I would hope in the next few years we would see both
Great post Becky.

Requesting permission to post this on the nKPro and iRacing sections of RSC.

It should produce some interesting discussion.
Very interesting post Becky. Well thought, well argumented : well done

However, I have to agree to those saying a static platform is more important to begin with. Simply for the fact that a dynamic platform, as other stated, can hide physical issues (if a dynamic track was implemented, then developing a perfect tire model would be impossible).
I, personally, think that the devs are thinking about a dynamic environment and will start working on it "soon". The tire model is good for the moment, and with a little more attention, it will be way, way better, to handle even closer to IRL. Then, the next patch is most likely to see those dynamics : change of the wind direction and strenght, dynamic weather, oil on track and track heating.

However, for some aspects, it might need a long adjustment, since, for the track heating for instance, to make it realistic you would have to make it heat with squares with sides as close to 0 as possible, which will be a real pain for older computers .
But this is what LFS will go toward quite soon, don't worry

@ the guy who said LFS' display was bad : Dude, I have a 17" (I mean inches) monitor on my laptop, and with appropriate settings in the views' options, you can acheive a pretty nice effect : increase the FOV to 110, inclination to -5, head moving to G forces, lower the driving position a bit and make the driver stand a bit forward to prevent the cockpit from being distorded and you get a pretty decent view, IMHO
Quote :Requesting permission to post this on the nKPro and iRacing sections of RSC.

Yes sure, post the linky too please. It's an open letter to all sim race developers which is why I posted it on the LFS forums (seeing as authors of 4 major sims hang out here to some extent...).

Regarding the static environment for development issue. It's a moot point, the programmer can set any environmental condition they please whilst working. Infact testing the car in a greater range of environmental conditions is more likely to achieve a realistic result, in my opinion. However my point was that this area of physics is substantially ahead of the others already, the other 2 critical areas of vehicle simulation lag way behind, and no sims yet have any environment simulation that recreates real racing, though some are less aweful at it than others.

Quote :to make it realistic you would have to make it heat with squares with sides as close to 0 as possible,

Actually no LFS' engine already has all the internal workings needed to introduce variable track temperature, using the existing path and picking a small number of points laterally at each node point along it would be more than adequate.
Quote from becky Rose :
Actually No Lfs' Engine Already Has All The Internal Workings Needed To Introduce Variable Track Temperature, Using The Existing Path And Picking A Small Number Of Points Laterally At Each Node Point Along It Would Be More Than Adequate.

S3?
#21 - Gunn
Quote from Becky Rose :The accuracy of the physics simulation is only important whilst it is the only aspect that is simulated.

Does that mean that you think the accuracy of the physics becomes unimportant once other things are simulated? I'm having trouble believing that is what you are trying to say.

Quote from Becky Rose :
The tyre physics in LFS are more than adequate. Anyone who argues they are not 'perfect' is clearly missing the point.

There is no argument here, the tyre physics in LFS are not yet adequate. Of course they will never be perfect, but a tyre that heats up too fast then deteriorates too fast is not adequate, especially when the alternative choice is a tyre that doesn't ever reach operating temp and therefore never realises a decent level of grip in the first place.

Quote from Becky Rose :In real racing my car can and will handle totally differently at the start of a stint to the end. Therefore the way the tyres 'feel' is a moot point.

On the contrary, the way the tyres feel is the whole point if you want dynamically changing conditions and tyre wear.

But anyway, I think your OP has a 'duh factor' of ten. I can't see anything that you have suggested that any serious sim developer hasn't thought of maybe a thousand times already. Most of the things you have suggested seem to be things that any dev would include in their sim if they had the means to do so, but not neccessarily with the priority that you demand.
Quote :But anyway, I think your OP has a 'duh factor' of ten. I can't see anything that you have suggested that any serious sim developer hasn't thought of maybe a thousand times already.

You might think that, but then again no sim developer has created a sim that simulates anything more than a fancy suspension model up to now. It's all about "physics" the buzz word, not about looking at what racing is really about.

For years games have been focusing on ever more detailed suspension and tyre models, but they're not even the biggest factor of racing - they are just the easiest to recreate with standard programming concepts.

What i'm saying, is to think outside the box and stop giving more of the same. Have a rethink of just what the biggest factors of a simulated motor race should be, to capture what racing is about, stop thinking about the car and start thinking about the race track, and the mechanisms of a motor race.

It might seem obvious now, but seriously, show me a race sim that has these things?

Quote :the tyre physics in LFS are not yet adequate.

There is more variance in the handle of my car between morning and midday than there is between a hard set of tyres and a soft compound. The ommission of remotely adequate aerodynamic simulation in any race sim has created huge field spreads that dont happen in single formula racing, seriously, go watch some...

Quote :Does that mean that you think the accuracy of the physics becomes unimportant once other things are simulated?

No, i'm saying that given our interface is inherrently flawed that the most important aspects to simulate are those which racing drivers must react too in the course of a days racing, not what mechanics happen the other side of an inifinitely variable collection of different control interfaces. I'm saying that to make sim racing a realistic simulation you must create an environment that lives and breathes like a race track, so that sim racing drivers have to make the same decisions as a real racing driver, and reacts to the same things.

When I race for real i'm constantly scanning the track surface. In sim racing the track surface is dead.

Thats not to say that suspension geometry and tyre physics are not important, but it would be possible to make a better simulation with tyre physics far inferior to LFS'. Tyre physics determine if the 'feel' is right, but the 'feel' is something that constantly changes in a real race car, where one change of castor, or a rise of a few degrees in the track surface, creates an utterly different feel anyway. How can we ever have a car that 'feels' right when there is no fixed thing that determines how a race car 'feels'.
Quote from Becky Rose :When I race for real i'm constantly scanning the track surface.

That is why I argued for a better interface between sim and driver. You need to see small objects on the track: bumps, oil, debris, puddles, etc. The human eye has a small part -- the area of focus -- where you can see in good detail, and the rest is peripheral vision with a wide FOV. Compare that to an LCD monitor, which has constant detail on a smallish FOV, and non-stereoscopic. This difference could become a pain if the ability to see track details becomes crucial. (With nondynamic tracks it's not so important, because you can drive by memory: you don't so much see where the bumps are, you remember them.)
Re: FOV ingame and RL
Quote from wsinda :The human eye has a small part -- the area of focus -- where you can see in good detail, and the rest is peripheral vision with a wide FOV. Compare that to an LCD monitor (...)

Think about putting on a full-face helmet with a pull-down screen. I only once in my life (only recently) sat in a kart going for my fist ever couple of RL-laps and the first impression was: "What's up with my vision/FOV". Luckily I could leave the screen open since it was just a small indoor track. I seriously doubt that anyone wearing such a device will ever get a vision close to what you get in a street car. I think a reasonably modern PC-rig with a large wide-screen display running at 1680x1050 or even 1920x1200 -tuned at the right viewing angle- will give you a much larger fov (in lfs) than what you can experience in RL when wearing a helmet.

Quote :(With nondynamic tracks it's not so important, because you can drive by memory: you don't so much see where the bumps are, you remember them.)

Isn't this partly the core of racing on a circuit: getting a quick and mostly accurate grasp of the track and actually remembering the potential go's and no-go's so-to-speak?
I say you do have a point there saying the level of difficulty controling a car on an ever-changing surface / dynamicly open piece of track will increase by some margin. But that would apply for all of the players. And that would mean for them to adjust in the most obvious ways:

-> they'd begin to use more all-round capable setups with less limited handling characteristics, thus making it possible to use a different line here and there in order to circumvent occuring inconveniences when racing in a more dynamic environment. Put in simple terms: the number of drivers using hotlap-setups in a race will decrease
-> they'd change their driving style to being more alert to forthcoming dangers & take less risks in on-the-edge dogfights, because racing-lines and track conditions are less predictable. It just might take some time for this change to happen. Cleaner racing - I agre with Becky on this - would be one of it's by-products.
If this projection is correct, then I honestly can't wait for it to hit lfs. It'll be well wroth the time waiting for.

status-quo:
But on the other hand: When I read that the sim-makers really spend too much thought on the car and it's mechanical systems then I can think of a perfectly good reason for this:
It's mostly the teenagers and car-lovers that are attracted to this kind of computer-game/sport up until this day. And of course THEY tend to focus on exactly that: "What does the car do when...?" "How am I going to master this or that car?".

wher to aim at?
What simulating actual motor-racing does or will be doing is going far beyond that. And here I do agree with Becky 100%: If the vehicl dynamics are at a point of refinement that is -judged by itself- this good and at the same time the overall racing feel is this far behind in almost all other aspects: then to take away the focus from the cars might really mean a giant step in the right direction!
That's not to say: "Don't finish the job on the tyres mate, they're good-enough"
To go with this example: Actually I think they are -quite frankly- good enough for Stage 2, now. As perfection is and will always be hard to achieve, stepping away from them for a bit and focusing the attention on other parts of the sim might just be a good idea once in a while.

After all the question is: (might drift off the main topic a little)
With what product will the more driving-focused user stick with? Or is it that this type of simrace does only exist in small numbers and the big money will be with those that just pose to be a driving sim? If so then the future of LFS or any other seriously thought-through sim is in BIG TROUBLE. Because there is no point in working on a product that isn't going to sell at all.
I for one waited with trying the GTR2 game until it became part of the low-budget sell-out mass at our local supermarket. The thing that put me off there was the domonstration version that didn't allow me to try the simulation-mode although it was purely out there to promote a driving-simulation. That's sort of paradox, isn't it? And it is the same story with rFactor. Just that I kind of read the other day that there now is a version out there allowing some mod-content to be tested for a limited time. Great. Just now I really lost interest for far too long in order to aproach that product again. To cut a long story short: I didn't like what I saw/felt/turned-my-wheel-at, so that was that (admittedly it has to do with driving-physics, there)

But then again: all the major magazines and journalists alike rather point at those mainstream-titles with official licenses of car-models, tracks and whole RL racing-leagues. And they do it for a reason: Those at the end of the supply-chain gain public awareness with it. And the more people play, the better they and their products will be recognised by the crowd. And connecting a new product with an "image" that has spread the world before never was a bad idea: called 'Marketing'. And it works because it is easy and thereby cost-effective to satisfy those who do not have too high expectations. Serious driving simply is not for everybody.


When I turn my head again and look at lfs which is still there, taking on this competition, I feel amazed - and immensely lucky at the same time...
...it gives me something I really like: a racing simulation that does what it should better and better, every time I start it. With enough players online that I can have a good race or two when I feel like it. A very straight-forward interface to all its settings that no other program I ever run could offer that even lets me acutally see what I do to the car when tuning its set-up.
Hope it'll survive the competition of that money- crazy i-Racing stuff and the advances of the ever more photo-realistic, "plastic-fantastic" arcade-genre that we are facing these days.

It will be interesting in what way improvements on dynamic environments will have an impact on lfs - when they come. I think one of it's real advantages has always been the small demand of system-specs and resources allowing for almost anybody to join end enjoy lfs online-racing. I can still use my 4y old laptop (P-M 1.4/ati-M11-64M) with lfs. What competing product even comes close?
Quote from Becky Rose :
Actually no LFS' engine already has all the internal workings needed to introduce variable track temperature, using the existing path and picking a small number of points laterally at each node point along it would be more than adequate.

Yes, but then it would not be *that* realistic. To have the most advanced simulation of track heating, the patches of tracks would have to be as small as possible, otherwise you might find yourself driving in a patch of 50x50cm that is heated well, then take the corner too wide, and put your tire on a patch of 50x50cm that is too cold, resulting in a very shapr loss of traction

FGED GREDG RDFGDR GSFDG