Damage and Collisions

I'm all for a collision model with broken parts and debris etc but think about the implications of this.

If we got a working collision model then the next natural question would be "when will the collision damage affect the aero??" Aris has mentioned in the aero thread that the aero is not dynamically calculated and becomes fixed once you make setup changes (my understanding anyway). Collision damage would require aero to be dynamic, e.g. front/rear spoiler breaks off.
Not sure if CPUs are even up to wind tunnel simulation for 20+ vehicles in real-time yet lol

I think collision damage is probably very low on the todo list (ATM at least), unless KS decide to add it without aero impact, which would imo be a complete waste of time/energy as the collision damage would just be cosmetic then and I for one would prefer the focus to be on mechanical damage which affects the physics and handling behaviour of the car

 

To be honest: I purchased ACC the very first day, and I would have done it everyway, because I am very confident in Kunos capability of delivering a good simulation.
But
Any kind of sim should IMHO attempt to reproduce EVERY aspect of real counterpart: this is the purpose of a simulation.
I am aware that trade-offs must be made BUT on the other hand market provides € 1000+ Gpu and very powerrful processors, that rarely are 100% loaded by software.....
So my opinion is: some degree of visual effects in case of collisions, damage, mechanical failures (for example a tiny trail of white smoke from a tyre scratching against a damaged chassis..... or from a broken radiator, or from an occasional oil puddle....or for example some bigger debris layng for a certain amount of time on track after a collision....) could be useful ...... and very effective.
So I only ask DEVS to go ALSO towards this direction.....
PLEASE

 

First there was night, then there was rain, and now we have damage... hahaha. At least we aren’t back to grass and trees...

 

I understand your POV (respect), but night & weather are really game changers, in fact Kunos implement them into ACC. Maybe damage is a game changer too ?

 

The delicate GT3 racing car is very strange as if it is as hard as a tank.
I do not think,kunos will leave as a brick that run GT3.
Let's see the situation.

 

I think you just Need to Count the Cars with Tape on it After a Endurance Race, to See that a good collision Model is a big plus. Separated Carparts add to immersion (while if they just dissapear...) but it also adds depth, if you Need to avoid them.

Those Little Details and some stuff goin in around the Track can make it like a real Race instead of just the Cars being convincing.

Bumpercar physics can Be Really Frustrating so that is clearly more Important.

 

IN FACT: it is so obvious that it' s hard to understand how a Simracing Developer could miss how important it is having some kind of modelation into sim ....
Just watching a GT race is enough to understand that damage is a relevant part of the equation and off course of .... the show

 

Understanding is not the problem, it's just devs always have to prioritize certain things over other things.

 

Question is what would you be willing to sacrifice to have it? Tyres? Physics? Track accuracy? For prety much all dev teams it’s impossible to do everything so choices have to be made.

 

I understand, BUT IMHO some degree of damage modelization is possible/worth the effort or tradeoff for sure: We are not asking for destruction derby

 

I'm not following this. There *is* some degree of damage modelling, it's just not represented visually, and the collision physics exist but are fairly basic.

It's not that the game lacks any of this entirely... it's just that Kunos don't prioritize gimmick-levels of damage and collision modelling, because they have more important things to do.

I think people are underestimating the VAST investment of time and resources it takes to develop a complex FEM-based material deformation system (This is basically the entire development focus of BeamNG.drive, and theirs *still* looks and behaves a bit goofy) and the additional vast amount of time it would take to make such a system remotely stable and predictable, rather than... "interesting".
You're also underestimating how hugely resource-hogging simulations like this are... a game like AC/ACC is already running on the ragged-edge of what CPUs can handle.

They already made their tradeoffs... you're just not recognising them as tradeoffs.

 

The last 10 or so replies tell me that there is still a confusion between damage and collision model.

Collision model = The way cars bounce off each other, or act when they hit the barrier. Damage model = Influence on car handling and behavior when damaged, or - visual simulation of damage to the car.

It's really two different things, and i argued above that we need a good collision model in the first place. It's really vital for good online racing.

 

I don't have high hopes for the damage model, I think it would take a lot of ressources to have something good. But I hope that crashing at 200kph in a wall = massive damage or DNF. In AC it was nearly impossible to destroy your car.

 

Absolute BS, engine was toast when you hit a barrier at 120+ kph

 

I know they're two very different things... but the thread is about both.

Yes, good collision modelling is important, but again, is an area of huge complexity, especially when factoring network lag into the equation. I would guess it's a system that Kunos are constantly trying to improve, but the tradeoffs we see currently are probably the result of a huge amount of research and development already, and will invariably favour a conservative approach over a detailed one. The more detailed a dynamic system like that becomes, the more it's open to random glitches and weirdness... where in a game about endurace races, a single physics glitch-out could be the end of someone's race.

 

It was definitely very hard to trash the car though. Unrealistically hard.

 

Again, when the slightest lag-induced glitch can spell the end of someone's 2-hour race... it probably pays for the developers to lean on the side of less-punishing. The alternative would be an endless procession of bile-spitting hate-threads clogging up their forums telling them what incompetent programmers they are.

 

Incorrect, it's only true for some cars (F2004) and you had to go a lot faster than 120kph. I don't have time to lose to make a video to prove it so try it in AC yourself.

 

Nope.
120 or 125 head on, every engine was blown.

P.S. You need to consider 100% damage settings. If you palyed on serevrs with lower damage settings then it was of course a very differentr scenario.

 

That's why i argued for a good collision model, including a good working netcode.

I wouldn't sacrifice damage though. After all, we are trying to simulate, right? A visual damage model, on the other hand, is pretty low on my wish list.