Online Physics vs Offline Physics --- Which is better?

  • Thread starter Thread starter xchristox
  • 53 comments
  • 12,065 views

Online Physics vs Offline Physics --- Which is better?


  • Total voters
    222
Messages
219
Messages
xchristox
Messages
F********** XBOX! :)
So we had our weeks of fun with GT now. Most people that bought the game from day 1 are above lvl 25-30 at the moment and going strong, including myself!

But, as you have probably noticed the differences in the physics engine online and offline is not only due to tire wear and fuel depletion, it's a lot more than that. The more I play online, the more I start to love it. The cars just feel way more lively and tail happy than in A-Spec, which I totally love!

ATM I'm stuck at A-Spec lvl 29 because I haven't played a-spec in a week or two, it's just too boring for me. I just can't be bothered to do the extreme events with 10 or more laps!

What bothers me in A-Spec:

- Offline physics --> extreme amounts of grip, you can throw the car around and it just doesn't break out. Takes all the fun out of the game for me.
- Crappy AI
- Being in 1st position after the first lap in like any event, even on hard tires. It's just way too easy.


For those who ain't convinced about the huge differences in physics online and offline, just take the test for yourself.


How to do this the easy way? Well, I took the Audi Pikes Peak Quattro, tuned it like a maniac, but kept it stock weight. Racing softs, and tuned the suspension to moronic settings. Ride height etc. on maximum and went for a spin. (a funny GT5 video gave me this idea, see below).



When on the track in practice mode, the car had so much grip. When throwing my steering wheel from left to right the car went on 2 wheels and jumped around like a snoop dogg lowrider, just like in the video above. If done to the extreme, it even flipped over with ease.

Now if you succeed in what said above, try doing this online! You just cant, your tires are way more slippery there and you don't get the required grip to flip the car over. It's nearly impossible if not completely impossible. The best I could do was get the car to drive on 2 wheels in a corner, but flipping it over and jumping around like a Snoop Dogg lowrider? No way!

Not only is it impossible to do this online, the possibilities you get offline are just weird and unbelievable. Shorty said --> it's not realistic.

So what is your opinion about the difference in physics? I've seen many people complain about the online physics being weird for them. The rear end breaks out and the car spins over and over. They are getting sick and tired of it because they can't drive the same lap times, therefore they hate the online physics. What I don't understand is that many people don't like a challenge, if something is harder I usually like it more!
 
Last edited:
I don't care which is better, I just want them to be the same.
If it's just a tires/fuel issue, I still don't care, and still want them to be the same.
 
I don't care which is better, I just want them to be the same.
If it's just a tires/fuel issue, I still don't care, and still want them to be the same.

I agree they should be the same. It's just that there is a huge difference, and the question here is which one you prefer :)
 
Personally i dont think thre is a difference between the 2 modes

I used to think there was until i did some investigating

I did what you said in preperation for an online championship a couple of weeks ago, ran sessions in practise and in online lobbies and i also noticed that in practise mode my car seemed to be planted to the track compared to online........in online i was nearly 3 seconds a lap slower, then i figured it out.

In Practise mode your car is running with no fuel and tires fully up to temperature compared to when you start driving in online mode with cold tires and a full fuel tank.

With this knowledge i ran in online for 30 laps to burn the fuel then put fresh tires on, within 3 laps i was nearly hitting the sort off laps i could do in practise mode. The differance is purely to do with tirea and fuel load rather than a difference in physics.

I further tested this by doing a one make race which allows you to set tyre fuel depletion and found that lapping in that compared to practise mode made my car 3 -4 seconds a lap slower due to the fuel load.

Alternatively a friend off mine thinks the handling online is affected by the lobby settings, race qaulity and voice quality, something to do with bandwith, its goes over my head so cant really comment, maybe someone more tech minded may be able to shed some light on that.

All in all online is a hoot lol
 
Personally i dont think thre is a difference between the 2 modes

I used to think there was until i did some investigating

I did what you said in preperation for an online championship a couple of weeks ago, ran sessions in practise and in online lobbies and i also noticed that in practise mode my car seemed to be planted to the track compared to online........in online i was nearly 3 seconds a lap slower, then i figured it out.

In Practise mode your car is running with no fuel and tires fully up to temperature compared to when you start driving in online mode with cold tires and a full fuel tank.

With this knowledge i ran in online for 30 laps to burn the fuel then put fresh tires on, within 3 laps i was nearly hitting the sort off laps i could do in practise mode. The differance is purely to do with tirea and fuel load rather than a difference in physics.

I further tested this by doing a one make race which allows you to set tyre fuel depletion and found that lapping in that compared to practise mode made my car 3 -4 seconds a lap slower due to the fuel load.

Alternatively a friend off mine thinks the handling online is affected by the lobby settings, race qaulity and voice quality, something to do with bandwith, its goes over my head so cant really comment, maybe someone more tech minded may be able to shed some light on that.

All in all online is a hoot lol

Hmm. So how do you explain the huge difference in the jumping of the car online and offline? Even when the tires are perfectly warm, it's impossible to re-do that video online. And the more weight due to the fuel should only help, as the car is tuned to maximum ride height and is really top heavy. The more weight the better if you wanna jump and flip it over...
 
I'm pretty sure that there are a different physics between my private lounge and the private lounge of my friends!
With the Arta gt300 I can easily make 2.02 on Suzuka in my private lounge but with the same car I hardly make 2.04 in the private lounge of my friend!
 
Hmm. So how do you explain the huge difference in the jumping of the car online and offline? Even when the tires are perfectly warm, it's impossible to re-do that video online. And the more weight due to the fuel should only help, as the car is tuned to maximum ride height and is really top heavy. The more weight the better if you wanna jump and flip it over...

Except with the fuel weight you won't be doing the same speed, so the results don't really mean much if more than one variable is changing.
 
Except with the fuel weight you won't be doing the same speed, so the results don't really mean much if more than one variable is changing.

I don't really understand what you mean. But if you say it's due to speed, I think you're wrong. I'm able to drive the car on 2 wheels, jump it around on it's suspension and flip it over between 30 and 150 mp/h. offline that is....
 
I don't really understand what you mean. But if you say it's due to speed, I think you're wrong. I'm able to drive the car on 2 wheels, jump it around on it's suspension and flip it over between 30 and 150 mp/h. offline that is....

I assumed you were lapping and chucking the car into a specific corner both online and offline. I was suggesting that if you were doing this you'd be going slower online when you reached the corner due to the fuel-weight, and therefore the comparison wasn't like-for-like.

Maybe you didn't do this though.
 
I prefer the more complex offline physics, I just want all the options turned on, or at least be able to toggle them (tire wear and fuel) myself
 
There is definitely a difference because I take my cars online (my lounge) to go for lap times and tests all the time.

I love the Subaru 307bhp with Comfort hards equipped cos they go so sideways and are easy to recover. Well after messing about for a long time on Nordschleife (especially the first sector) I wanted to make a video replay to watch the great slides I'd been doing but of course you can't watch replay in Lounge mode or other online modes.

So I quit out and took the same Subaru to Nordschleife, but this time in Practice mode. I started and couldn't get the car anywhere near as sideways as I could just 5 minutes earlier on Lounge mode. I figured I must be on Sports Hards or Soft instead of Comfort hards. Nope, after checking I was still on Comfort Hards. So I went back in and tried again and the same thing happened again I couldnt get the car dancing like I could before.

I'm quite experienced in GT5 and also use a wheel setup, so I know where and when I can get the cars to slide when I want to. But the difference on and offline in physics is quite a lot indeed, which is a shame.

For the record I prefer Online Physics :)
 
Don't prefer one over the other really. Just wish they could be consistent. I like the fact that your tires wear online, but at the rate they ware and the amount of grip you lose so quickly, it's not really realistic.

Example: Take the BMW M3 GTR race car to the Nurburgring. Put any race tire on the car (they all wear the same, I've tested it). By the time you get to the karussell corner on your first lap, your tires are already worn to the point that it's almost a struggle to not spin out through the turn (Im using a DFGT, and I know this track well). I really don't think the tires should be 1/4 worn already half way through the first lap.
 
Oh seriously why do you have to defend it at all cost?

Running a game, host a online session, doing voice chat, provide left processing power if someone eventually wants to play 3D is alot you are asking to do for a Playstation. It`s alot for a powerfull PC.

Your best bet is online physics are dumped down to share processing power with everything else the system has to calculate. The game is not able to display other drivers in a smooth way on a straight for many gamers - because of lag.

I also had some strange circumstances during online races, i couldn`t see some other players, ppl told me this should be normal for a lower bandwith. I saw ppl driving a car they never picked more than once, it could be the game is choosing to display a car model with less polygons rather than freaking out.

At the end it is hard to believe it could manage what you see in the video above on top of everything.

My guess is online physics are more simplistic to keep the whole thing running in online mode.
 
I haven't noticed any differences in the two that can't be attributed, for the most part, to tire wear. I'll definitely be looking now, tho.
 
So where is the video showing online vs offline?... That video in the first is just offline isnt it?
 
I wanted to make a video replay to watch the great slides I'd been doing but of course you can't watch replay in Lounge mode or other online modes.

watch, no. But you can save them, hit the "results" button and then it gives you possibility to save your online race replays...
 
Hmm. So how do you explain the huge difference in the jumping of the car online and offline? Even when the tires are perfectly warm, it's impossible to re-do that video online. And the more weight due to the fuel should only help, as the car is tuned to maximum ride height and is really top heavy. The more weight the better if you wanna jump and flip it over...

So many of us here love the GT series and so badly wanted GT5 to be worth-the-wait that I think it is hard for many of us to accept that this product has a huge array of flaws, of which this is but one. Fanboys seem to be able to convince themselves of nearly anything in defense of their chosen product. In this case many people here want GT5 to be perfect badly enough that they can overlook any amount of proper evidence in the pursuit of this delusion. I sure wish GT5 was perfect. Heck, I wish every item I every purchased was perfect, but I'll never be a fanboy for any product period. I'll talk about the positives and the negatives. In this case, your are absolutely correct, GT5's online physices are different than the offline physics in ways that can't be explained away (in any rational way) with tire temperature, tire wear, or fuel weight.

However, in defense of those people who think that it is just tire wear: This problem varies incredibly from car to car and track to track. There are a few cars I drive online which are close enough that, were I so inclined, I could easily mistake the handling difference as being tire wear, temperature, and fuel weight. However, this theory falls completely apart as you experiment with different cars and tracks. Trial Mountain is the worst one I've found. It is almost one of my "tuning tracks", and one of the very few tracks I still like in GT5 (PD eliminated nearly all my favorite tracks between 4 & 5) so this problem rose up and slapped me in the face almost immediately.

So, on to the offline vs. online physics question. There is a thread in the online racing subforum where we discuss this a lot, so I'll try to keep it brief here. However, the question appears to have a complex answer, so I don't know how short I can be. First, I'm going to talk about Rear Engined, Rear Wheel Drive cars only (RR), because those are the only kind of car I drive in motorsports (autocross, ice-racing). I've also driven RR cars as daily-drivers now for (I feel old) 25+ years in all conditions (my winter beater is a classic RR porsche).

At least in the case of RR cars, the OFFLINE physics is, as you mention in the original post, a bit too easy. However, it is also vastly more like reality. The problem is that under the online physics model the RR cars (especially the Yellowbird and the Alpine 1600s, the BTR etc.) simply do not respond to weight-transfer and throttle-control properly at all. It is very nearly impossible to gently lift the throttle to get the nose down, initiate that very small amount of rear-engine (throttle-off) drift, and then squeeze the pedal back down to settle the rear-end of the car. Go watch videos of how the best drivers drive classic 911s in motorsports. Or go to a PCA autocross or track-day event. The best drivers do what I'm describing, and I can honestly say that I'm very good at it also. I can to the lift-rotate-squeeze-settle thing all day long in real life and all day long in GT5 offline physics. Go online though, and everything in these cars goes to ****. They develop 2 problems which make these cars drive in a unrealistic way. 1) They get absurd quantities of snap-back even under the most mild of slip-angles. 2) The rear of the car will not settle properly when the throttle is squeezed back down.

So, at least for those cars I feel I can legitimately judge: The offline physics is too easy, but much more accurate in the way the cars respond to steering and throttle inputs: It is possible to use real-life driving techniques to drive the RR cars offline. Online these real-life methods fail to work in these same cars. (However, there are many other cars which I drive where I prefer the online physics. More on-the-edge, less understeer, etc. But since I don't do motorsports in those cars in real-life I don't feel I can judge which model is more realistic. But my favorite cars are all the classic RR models, and I'm not exactly happy with PD for bungling their online handling so badly.)

Here is a video I made comparing offline and online handling of the Alpine 1600s on Trial Mountain. This is one of the cars with the most severely-altered online/offline personalities, on a track which exhibits this problem severely as well. I'm purposefully exaggerating the amount I'm allowing the car to rotate to demonstrate the problem (video is annotated in my own weird way).

Offline and Online handling of the Alpine 1600s
 
Last edited:
In the online gameplay the devs use a 3D Cartesian plane to send information like x=153728 y=245378 z=315 x and y determine the Position of the Vehicle and z determine the height, and that use another variable to determine the direction, a=157 if the car could flip, they will need another variable, b=324
And the severs will need manage more data, and slow connections will fail, soo they change the Physics to not need another variable and make the online gaming viable to slow connections.
 
How to do this the easy way? Well, I took the Audi Pikes Peak Quattro, tuned it like a maniac, but kept it stock weight. Racing softs, and tuned the suspension to moronic settings. Ride height etc. on maximum and went for a spin.

How do you do this?

Hardest spring setting, highest ride height and 0 dampers??
 
Anyone who thinks that off-line the cars are "way too grippy" and that the game is "way too easy", needs to put down the crack pipe, and back away from the race slicks - in fact, maybe even from the game itself as they are just too obsessed with gaming the game.

In short - it's not. Pushing things to known ridiculousness doesn't really prove anything because all software can be pushed outside of it's intended bounds. It may well be that PD just didn't do that great a job of reigning in jack-assery like that, but that doesn't say anything about how the modeling is when things are treated properly (well within expected parameter bounds).

Also, anybody can put slicks on a car and blow the field away and turn in good laptimes. Use normal tires and actually drive the car and learn it's quirks. Put comfort softs on a car with no driver aids and learn how to actually drive, rather than just ace a game.

You'll find plenty of challenge then, and not too much grip - guaranteed.
 
Anyone who thinks that off-line the cars are "way too grippy" and that the game is "way too easy", needs to put down the crack pipe, and back away from the race slicks - in fact, maybe even from the game itself as they are just too obsessed with gaming the game.

In short - it's not. Pushing things to known ridiculousness doesn't really prove anything because all software can be pushed outside of it's intended bounds. It may well be that PD just didn't do that great a job of reigning in jack-assery like that, but that doesn't say anything about how the modeling is when things are treated properly (well within expected parameter bounds).

Also, anybody can put slicks on a car and blow the field away and turn in good laptimes. Use normal tires and actually drive the car and learn it's quirks. Put comfort softs on a car with no driver aids and learn how to actually drive, rather than just ace a game.

You'll find plenty of challenge then, and not too much grip - guaranteed.

I don't run racing slicks on every car, but I do put them on race cars. And the difference is there.
 
I finally discovered that you can change tires prior race start, LVL 18 spec-A, and I just found out now. Now I can win 1st place by almost lapping the last cars. Why the hell can't we change our tires IN THE GARAGE??!

Staying on-topic, The ONLINE rearend snapback, tail spins HAS GOT to be a BUG in the game. Every car from various players from various tracks experiencing the same thing?
and people think this is only because of fuel loads?

what a load of crap. clearly PD needs to address this bug. Still, GT5 is worth $500 , and not just $50 as a video game. I just love the variety in game.
 
I don't run racing slicks on every car, but I do put them on race cars. And the difference is there.

"I don't always run racing slicks, but when I do...I put them on race cars. Stay grippy, my friends."

dos-equis-guy1.jpg
 
Personally i dont think thre is a difference between the 2 modes

I used to think there was until i did some investigating

I did what you said in preperation for an online championship a couple of weeks ago, ran sessions in practise and in online lobbies and i also noticed that in practise mode my car seemed to be planted to the track compared to online........in online i was nearly 3 seconds a lap slower, then i figured it out.

In Practise mode your car is running with no fuel and tires fully up to temperature compared to when you start driving in online mode with cold tires and a full fuel tank.

With this knowledge i ran in online for 30 laps to burn the fuel then put fresh tires on, within 3 laps i was nearly hitting the sort off laps i could do in practise mode. The differance is purely to do with tirea and fuel load rather than a difference in physics.

I further tested this by doing a one make race which allows you to set tyre fuel depletion and found that lapping in that compared to practise mode made my car 3 -4 seconds a lap slower due to the fuel load.

+1

That and weight distribuition because of the fuel load

thats why I voted online because taken in account fuel load and tyre wear makes it a lot more realistic
 
+1

That and weight distribuition because of the fuel load

thats why I voted online because taken in account fuel load and tyre wear makes it a lot more realistic

Just as an FYI, for every poster out there who keep asserting that it is only tires and fuel, there is another one (like me), who can demonstrate with 100% certainty, via. reproducible experimentation, that it isn't just these things. You need to try a wider variety of cars on a wider variety of tracks because this issue varies tremendously from car to car and track to track... That's one of the primary reasons we keep getting diametrically opposed viewpoints.

At the very least this indicates some serious bugs in the system. That might be a good thing, at least if it is really down to just some software bugs they could eventually fix it. If PD would just go back and do whatever tweaking is necessary to fix those cars which just "flip out" when driven online that would go a long way towards helping the situation.
 
To make things clear: the video is not mine. It just gave me the idea to test it both online and offline. I'm not able to make decent videos except for a handycam, which sucks :P

Also I'm not the guy that is in love with racing slicks. As I stated in my original post, I love slippery cars with tail happy rear ends. I prefer sports tires the most, but to fit comfort tires in an extreme event to make it "challenging" just doesn't cut it for me. I just think the AI should take wayyy more risks and push the cars to it's limits. The way it is now it's that they simply cruise around the track, not losing 1% of grip!

Therefore I love to play online because a) humans always find the edge of the car, and go over it most of the time :P and b) because cars feel a lot more lively online than they do offline.
 
Back