When you try out a new driving game and you say, "the physics suck!" What are the objective criteria you're using to make that determination?

Braking distance? Grip when making turns? Weight transfers?

When comparing games I always hear people say things like, "the physics are no where near live for speed" but they rarely make objective statements about what it is that game x is lacking.

What criteria do you use when comparing sims? And what are the minimum qualifications a game requieres to be considered a "sim"?

For me, it's the general "feel" of the game, does the physics engine translate the movement of the car, the grip of the tyres etc to the player, does it "feel" like you are driving a real car.

It's very hard to quantify, but when something is "right" you just know it and feel it, LFS does a top notch job of translating tyre grip and wieght transfer to the player.

TBH, I find it VERY difficult to believe that anyone without a little bit of fast driving experiance could make a judgement call on the physics in a driving sim.

Minimum qualifications, first and foremost, it needs to feel right and perform realisically, can you turn and brake at silly speeds.

When you turn, it feels like you have some kind of aadvanced rubber that glues you to the road.

When you accelerate, it feels like you are slipping on the spot and all of a sudden, 6G's of acceleration...

For me the big tell is when you step outside of the grip limits of the tyre. Are what happens when you push beyond the boundary right and repeatable every time you try.

Not many "sims" are tbh.

For me its LFS, RBR and NKP (Although NKP should be considered a joke until they start to support it. All bugs and no dev support)

Many rate NR2k3 but is never did anything for me.

It's mostly the force feedback. It gives me the first cues about the ingame physics without making an extensive analysis. For example the wheel going 100% limp on understeer is a good indicator that the devs have a wrong idea about tyre grip over the limit, even though it's possible the effect could be hardcoded and not physics driven. However, even that tells me a lot about how the physics were approached. Chances are most devs try to generate the FF purely from the physics at first (it would be the least amount of work needed to get FF working), but the presence of hardcoded effects is a good way to tell that this approach obviously didn't work satisfactory, whatever the reasons might be.

The next point to try is to drift the car in a controlled fashion. If it stays glued to the road or spins around right away you can be pretty sure the sim is rubbish. Same is true when you need to use a weird driving technique that exploits physics glitches to make it look like you're drifting, or when you need a million tries to juuuust get it right one time.

These two points are for me the basic tests a sim needs to pass before I even consider deeper analysis and testing of the physics. If you have a brilliant physics engine but crap FF, then it's just as worthless as having extremely well faked FF on a NFS-like physics engine.

Does that mean LFS is the holy grail of physics? No. LFS has several flaws - some directly related to the tyres, others caused by not yet implemented/completed features. But you actually have to dig very deep to get to these issues, and LFS just gets so many more things just right. What's probably missing the most are some immersion creating tidbits that make you actually feel like you're sitting in a mechanical potpourri of metal, screws, rubber, oil, gas, etc. - LFS has a rather sterile feeling whereas netKar seems to be much better in that aspect.

Quote from AndroidXP :

But you actually have to dig very deep to get to these issues...

Not too deep... flooring the throttle and dropping the clutch produces a bit of squirm only when you get up to the FO8 or BF1.

I think, Somasleep, you know what to look for. Force feedback is irrelevant if you're not using an FF-wheel or a mouse, but even with a second-choice controller you'll be able to clearly judge the nature of the grip available and the way that the sim deals with weight transfer.

Basically, drive the car in a circle; if the car feels 'dead', doesn't understeer at any point, doesn't give much feedback at all, then there's a problem.

Quote from danowat :

For me, it's the general "feel" of the game, does the physics engine translate the movement of the car, the grip of the tyres etc to the player, does it "feel" like you are driving a real car.

It's very hard to quantify, but when something is "right" you just know it and feel it, LFS does a top notch job of translating tyre grip and wieght transfer to the player.

TBH, I find it VERY difficult to believe that anyone without a little bit of fast driving experiance could make a judgement call on the physics in a driving sim.

Minimum qualifications, first and foremost, it needs to feel right and perform realisically, can you turn and brake at silly speeds.

Totally agree with that Dano said. It`s the generally "feel" of the game. Some games I just feel that it`s too much grip, and it`s like impossible to do a slide without using the handbrake. A other thing too is how cars react when you hit them, in some games you just can`t wreck the AI, or, you can bump them some but they are like impossible to spin, that is a negative thing too.

What I like to do, apart from judging by feel, is watching the cars from the outside. If you've spent countless hours of your life watching motorsports of all forms, at the track and on TV, then you get a pretty good sense of whether car behavior looks real or not. For example GT4 just looks dead while ISI looks very strange and snappy whenever a car gets into oversteer. LFS is still the only one where you can really see the car squirming for grip.

I am fairly sure that in ISI sims, the replays have very different physics representation to the actual "driving" of them.

I can't expand on what Danowat said either. It's definitely a seat-of-the-pants feeling that you know when a car is behaving the way you expect it to when you perform an action.

That's the reason I can't play things like CMDirt or Race with a wheel - the cars don't feel like I'm controlling them at all. It's also why I can't play them in cockpit view. It's like they're designed for a joypad and chase view.

Quote from J.B. :

What I like to do, apart from judging by feel, is watching the cars from the outside.

trouble is the only sim i know that has properly simulated rigid body physics (lets just ignore the collision bit for now) is lfs
funny thing is it had that edge over all other sims out there right from the beginning (the gti on a soft suspension rolling like a punto almost made me buy s1... too bad the tyres were rubbish back then)
its probably the easiest bit of physics to get right yet theres only 1 sim i know of that does

Quote from J.B. :

What I like to do, apart from judging by feel, is watching the cars from the outside.

Me too. How realistic cars look when simply going around the track was the first thing about LFS that made me go "wow".

I know this may seem like a very unscientific approach to determining realism, but the human brain is extremely good at identifying things that look slightly "off" (danger! run!), either in behaviour (physics, movement) or looks (graphics). A physics engine that fugdes a lot of details will just not look as realistic in a replay. You can usually tell instantly.

they are probibly refering to that they don't like the way the car handles and they think its the game, not setup/car etc. or the game doesn't permit NOS @ 600kmh on a turn and they think it sucks because they can't do that.

Quote from wien :

Me too. How realistic cars look when simply going around the track was the first thing about LFS that made me go "wow".

I know this may seem like a very unscientific approach to determining realism, but the human brain is extremely good at identifying things that look slightly "off" (danger! run!), either in behaviour (physics, movement) or looks (graphics). A physics engine that fugdes a lot of details will just not look as realistic in a replay. You can usually tell instantly.

Same. I can tell a lot about the realism of a game by watching the cars do a lap of the track from the outside. I've always had a knack for that kind of stuff, and LFS just looks so REAL - even though I've never really raced a car - I can tell what looks real just based on my own experiences IN a car.

i would say that i am basing my judgment on real driving experience... if i had driven as many cars as Jeremy Clarkson, but, no i haven't...

I am just waiting for someone to hand me the most realistic sim and i am going to base other games off of that.

and by the way, about maneuverability of the cars... is there a point when it's just too much? where there is an unrealistic amount of understeer... like how, in GT, the cars just feel so heavy in "Pu-ro-fe-shi-en-oh-ru"(the word "professional" in a japanese accent) mode due to the lack of grip and the inability to put the power down to the pavement exiting a turn...

Quote from atlantian :

i would say that i am basing my judgment on real driving experience... if i had driven as many cars as Jeremy Clarkson, but, no i haven't...

I am just waiting for someone to hand me the most realistic sim and i am going to base other games off of that.

and by the way, about maneuverability of the cars... is there a point when it's just too much? where there is an unrealistic amount of understeer... like how, in GT, the cars just feel so heavy in "Pu-ro-fe-shi-en-oh-ru"(the word "professional" in a japanese accent) mode due to the lack of grip and the inability to put the power down to the pavement exiting a turn...

not sort of picking up what ur trying to say there?

Atlantian - are you talking about the GT cars in LFS? Or about a game called GT? Or what? I'm thoroughly confuzzled...

the main reason i dont like GTR2 (the only other proper sim ive played) is when you do clutch kicks and other manuvours to force the ass end out, it just didnt respond as it would in real life..

edit: and yes i did turn off ALL the driving aides..

Quote from wien :

Me too. How realistic cars look when simply going around the track ...

Really not sure whether this is a good test... I'm no programmer, but I'll bet that game replays are often based on positional data rather than thorough physics calculations.

Take GPL as an example. I know its old school compared to LFS, but it still gives me a very immersive sense of feedback from behind the wheel. So much so that I often get the impression that I can feel the tyres (and the general attitude of the car) through the wheel, especially in limit situations. And this is despite my wheel being a cheap, non-FF item.

The replays however do look slightly weird and floaty at times...

Quote from danowat :

I am fairly sure that in ISI sims, the replays have very different physics representation to the actual "driving" of them.

But that's because the "physics engine" isn't even close to being capable of rendering a realistic look, so they fudge everything 10x as much to acheive what still looks crap. I maintain the only saving grace of ISI sims is the day/night cycle and the squeeling brakes.

You have a point there. I haven't looked too closely at how the Papyrus/ISI/Simbin sims do replays, so they may give a completely wrong impression. RBR and LFS (the IMO best overall sims out there) both look extremely realistic compared to other sims in replays, so in my head replay realism and realistic physics seem to correlate fairly well.

You're right that an interpolated keyframe replay wouldn't give proper results though, so it may not be accurate for all sims.

Quote from Stang70Fastback :

Atlantian - are you talking about the GT cars in LFS? Or about a game called GT? Or what? I'm thoroughly confuzzled...

oh, i ment GT as in Grand Turismo...

Quote from atlantian :

oh, i ment GT as in Grand Turismo...

Here is a viper doing donuts(??) IRL

http://www.youtube.com/watch?v=QBW1FC7OyiY&NR=1

Here is a viper in Grand Turismo doing the same thing

http://www.youtube.com/watch?v=qLii_uu1I9s

Here is the same thing in LFS

http://www.youtube.com/watch?v=cTqanKs2CUo

I can't tell the difference just be looking at replays. Well, I see a lot more smoke IRL

well starting from(IRL)0:15 and (GT5)0:34 they look identical for a few seconds... the LFS is... okey... i am assuming that traction modeling is a relatively widespread algorithm and very easy to replicate... since i am assuming that ALL 3 looks the same, and it's just that LFS has crappier graphics then GT...

Quote from somasleep :

I can't tell the difference just be looking at replays. Well, I see a lot more smoke IRL

Well, personally I think that the Grand Turismo clip DID look 'fake' compared to the other two... BUT. That doesn't say anything. You can tell the difference if you watch the cars pitching on their suspensions in turns and over curbs doing a lap of a track. Doing doughnuts doesn't show anything. There is no suspension work in that - just spinning wheels...