Collisions unrealistic

[Steve Worrell]

I have to say I'm absolutely loving ams the and beta with the updated AI. Very special indeed. I think I would go as far to say it's my favourite SIM now.

One thing that really bugs me though is the collisions between vehicles. They sort of bounce off each other and impacts are unrealistic. Does anyone else find this to be the case?

 

[replicaz1427]

Damage simulation should encompass all aspects of damage, some examples could be, tyre blow outs from aggressive cambers, over pressured tyres or running over sharp curbs or debris on track, gearbox or clutch failures from poor rev matching, engine valve train damage from downshifting to quickly and over revving the motor, this would be more appropriate on older and historic cars. But the list of issues for modern cars could be similar due to tech failures, this would make the hobby so much more realistic and also bring that air of unpredictability to sim racing, as in real world motorsport a failed $2 part can cost you plenty on race day!!

I definitely second this opinion, gearbox failures, clutch failures, punctures we need this to be more of a factor in future titles. It seems nigh on impossible to get the engine to mis-shift in games like RRRE and non existent in AMS, which over the course of a long race losing gears would add so much too the challenge side of things and change the entire face of a race.

 

[Gringo]

What seems to be lost in this conversation is the limits of real time simulation. You simply cannot make the (practically) infinite amount of calculations in measurement required to achieve the level of detail you are looking for. There are only so many things you can do in a microsecond.

No doubt one day this will be possible, but keep in mind the engine used today is 10 years old. Even this old timer is hard to beat by the more current crop of contenders. I suggest the problem is not in the wanting for more, but it is simply not possible with the current limitations.

 

[Steve Worrell]

What seems to be lost in this conversation is the limits of real time simulation. You simply cannot make the (practically) infinite amount of calculations in measurement required to achieve the level of detail you are looking for. There are only so many things you can do in a microsecond.

No doubt one day this will be possible, but keep in mind the engine used today is 10 years old. Even this old timer is hard to beat by the more current crop of contenders. I suggest the problem is not in the wanting for more, but it is simply not possible with the current limitations.

Yeah this is true. But if we didn't say anything how would Devs ever know what the customers want. Consider it feedback rather than a demand.

 

[BillGuy]

What seems to be lost in this conversation is the limits of real time simulation. You simply cannot make the (practically) infinite amount of calculations in measurement required to achieve the level of detail you are looking for. There are only so many things you can do in a microsecond.

On a related note, you need to consider that improvements in collision modeling might not necessarily “align” with the physics and screen presentation of damage modeling.

Hypothetical Example:
Say cars implement outer, middle, and inner bands that have defined energy absorption properties. This might increase the realism of how cars interact with each other in collisions. But what if these bands don’t align (physically or conceptually) with how you would expect a car to suffer damage in collision? Whose priorities should win? The people who would prefer more realistic collision behavior or the people who place more value in damage modeling?

Again, that’s just a hypothetical example. I do think, however, that it illustrates the point of collision behavior and damage behavior being a real balancing act with the potential of upsetting people’s priorities just as much as pleasing them.

 

[Steve Worrell]

On a related note, you need to consider that improvements in collision modeling might not necessarily “align” with the physics and screen presentation of damage modeling.

Hypothetical Example:
Say cars implement outer, middle, and inner bands that have defined energy absorption properties. This might increase the realism of how cars interact with each other in collisions. But what if these bands don’t align (physically or conceptually) with how you would expect a car to suffer damage in collision? Whose priorities should win? The people who would prefer more realistic collision behavior or the people who place more value in damage modeling?

Again, that’s just a hypothetical example. I do think, however, that it illustrates the point of collision behavior and damage behavior being a real balancing act with the potential of upsetting people’s priorities just as much as pleasing them.

Yeah I see your point. Personally I wouldn't mind if the damage model wasn't as visually impressive as the likes of wreckfest or beamng. What I would like to see is realistic collision detection with a moderate visual damage and a realistic mechanical damage simulation.

What I think desperately needs to happen is players need to be punished for crashing. The SIM that does this best is rfactor2 but even that is rubbish.

 

[pennywiseguy]

just spotted this line in my .plr...

Alternate Collision="0" // Temporary variable as we develop new collision system. NOT AVAILABLE IN THIS VERSION


I wonder.