Author Topic: Collision Model  (Read 1972 times)

Offline SKurj

  • Gold Member
  • *****
  • Posts: 3630
Collision Model
« Reply #15 on: June 28, 2003, 12:54:02 AM »
It isn't always the case like i said...

1 in 10 you may actually fly through someone.

The claim here is always that someone hit YOU.... when as far as your FE is concerned YOU made contact with them.

I've had a collision in a scissors before as have most of us I am sure but guess what....   I hit them, NOT the other way around.

Start looking at it this way and the current collision model makes perfect sense.

Even if you were parked on the runway and a vulcher appeared to run into you...  If you took damage YOUR fe saw your plane hit another.  If you didn't then your FE didn't see contact.


None of us want no collisions, if you seriously think about it, the reasons are obvious.

Mutual destruction upon any collision would bring you back here whining alot more than the current model, because you would more often die from NO contact than you would where you actually saw contact.  AND even then odds are your damage would occur up to seconds AFTER the supposed collision took place.


Suggest a better model (no collisions is not an alternative) that would work and rid us of these whines and I am sure HT would implement it right away.


SKurj

Offline MANDOBLE

  • Parolee
  • Silver Member
  • ****
  • Posts: 1849
      • http://www.terra.es/personal2/matias.s
Collision Model
« Reply #16 on: June 28, 2003, 05:39:25 AM »
Quote
Originally posted by GunnerCAF
If you hit some, you take damage.  If they hit you, they would take damage.


FALSE, totally FALSE, this is a pure myth.
I landed my 109 at a deacked enemy field and placed it near the end of a runaway heading the opposite side. An enemy P51 spawned at my back and rolled over my stopped plane. My plane was cut in half while the enemy P51 took off undamaged.

Offline Innominate

  • Gold Member
  • *****
  • Posts: 2702
Collision Model
« Reply #17 on: June 28, 2003, 05:54:22 AM »
I died the other day colliding with B17's.  I lined came in fast behind the lead one, shot him up, and got hit by one of those ultra-laggy random-warp drones.

More than once I've died in collisions when people have warped on top of me.

Offline whels

  • Silver Member
  • ****
  • Posts: 1517
Collision Model
« Reply #18 on: June 28, 2003, 08:32:55 AM »
Quote
Originally posted by SlapShot
There have been times when I would swear that I got under the oncoming plane with the correct angle as to avoid any gunfire ... only to hear bullets ripping thru my plane and the tail come off.

I was flying the other night, into a merge with an N1K, where I had at least 100-200 foot separation between him and I on my left. After the merge, I swung right, checked my rear view and to my amazement, he lost 1 wing and was going down completely on fire. I have seen this at least 4 or 5 times before.

If on the above scenario, I /You ended up in the tower, I/You would be pissed beyond belief.



slap,

i have had collision deaths were i visually saw  our 2 planes miss, but i still took damage from a ram. i have had a plane pass me
to the left(missing me visually) and taken right wing damage.

i have visually on my FE flown through a enemy plane and took NO damage.


the collision model is porks, it dishes out damage arbitaraly sometime.

did u ask the other pilot in ur situation if he saw a ram? he might have seen u miss also and still gotten arbitary damage.

Offline SlapShot

  • Plutonium Member
  • *******
  • Posts: 9121
Collision Model
« Reply #19 on: June 28, 2003, 11:00:09 AM »
Quote
Originally posted by whels
slap,

i have had collision deaths were i visually saw  our 2 planes miss, but i still took damage from a ram. i have had a plane pass me
to the left(missing me visually) and taken right wing damage.

i have visually on my FE flown through a enemy plane and took NO damage.


the collision model is porks, it dishes out damage arbitaraly sometime.

did u ask the other pilot in ur situation if he saw a ram? he might have seen u miss also and still gotten arbitary damage.


No I didn't ... there were other cons to deal with, so I didn't have a chance.

How many times, per night/day, does this anomoly really effect you ? I typically can run into this, at the most, 3 times per night and most days/nights I NEVER run into this. For me its a non-issue ... I just up another plane and get back into the action.
SlapShot - Blue Knights

Guppy: "The only risk we take is the fight, and since no one really dies, the reward is the fight."

Offline whels

  • Silver Member
  • ****
  • Posts: 1517
Collision Model
« Reply #20 on: June 28, 2003, 11:10:49 AM »
Quote
Originally posted by SlapShot
No I didn't ... there were other cons to deal with, so I didn't have a chance.

How many times, per night/day, does this anomoly really effect you ? I typically can run into this, at the most, 3 times per night and most days/nights I NEVER run into this. For me its a non-issue ... I just up another plane and get back into the action.



i rarely collide with anyone.  maybe 1  time every couple months.
the fly  through without damage happens alot from my FE standpoint, especially  Buff killing. the arbitrary damage, id say hard to  tell how often but i have noticed it.   i have cable connect
so i usually die in the collisions.

Offline GunnerCAF

  • Silver Member
  • ****
  • Posts: 946
      • Gunner's Grange
Collision Model
« Reply #21 on: June 28, 2003, 11:39:54 AM »
Quote
quote:
--------------------------------------------------------------------------------
Originally posted by GunnerCAF
If you hit some, you take damage. If they hit you, they would take damage.
--------------------------------------------------------------------------------



FALSE, totally FALSE, this is a pure myth.
I landed my 109 at a deacked enemy field and placed it near the end of a runaway heading the opposite side. An enemy P51 spawned at my back and rolled over my stopped plane. My plane was cut in half while the enemy P51 took off undamaged.


Relitively speaking... In this case, you hit him, but you were not moving in relation to the ground :)

The model is not perfect, if you call real life perfect.  Dealing with delay and collisions, the model is brillant in my opinion.  Collisions are needed to make head on passes dangerous.  Many of the quirks come from delays.  Figure out a way to eliminate all delay, and I bet HTC will have a perfect model.

If you think about it, it is truly incredible that people from different parts of the world can play a real time game like this at all!

Gunner
Gunner
Cactus Air Force

Offline Murdr

  • Platinum Member
  • ******
  • Posts: 5608
      • http://479th.jasminemaire.com
Collision Model
« Reply #22 on: June 28, 2003, 12:22:30 PM »
This isnt the only scenerio, but it is the one that causes complaints.
Problem: Pilot A on broadband sees collision and reports damage before Pilot B on dial-up sees the inevatible collision.  Result Pilot B flies merrily on his way because A is either dead, or had a drastic flight path change.

Possible solution:
Delay damage pending collision confirmation via both FE.
Did A and B see collision?
No= No collision
Yes= collision, assign damages.

"A" may not take damage until after the merge from his pov, but at least "B" also takes the hit now.

Just a thought.

Offline Martlet

  • Platinum Member
  • ******
  • Posts: 4390
Collision Model
« Reply #23 on: June 28, 2003, 01:02:10 PM »
Quote
Originally posted by Murdr
This isnt the only scenerio, but it is the one that causes complaints.
Problem: Pilot A on broadband sees collision and reports damage before Pilot B on dial-up sees the inevatible collision.  Result Pilot B flies merrily on his way because A is either dead, or had a drastic flight path change.

Possible solution:
Delay damage pending collision confirmation via both FE.
Did A and B see collision?
No= No collision
Yes= collision, assign damages.

"A" may not take damage until after the merge from his pov, but at least "B" also takes the hit now.

Just a thought.


EXACTLY what i'm talking about.  That's a good idea.  I'd rather get delayed damage, than watching someone fly off.

Offline MANDOBLE

  • Parolee
  • Silver Member
  • ****
  • Posts: 1849
      • http://www.terra.es/personal2/matias.s
Collision Model
« Reply #24 on: June 28, 2003, 02:10:39 PM »
Quote
Originally posted by Murdr
Delay damage pending collision confirmation via both FE.
Did A and B see collision?
No= No collision
Yes= collision, assign damages.


This is raw common sense, why it is not implemented after years? who knows.

Offline ccvi

  • Gold Member
  • *****
  • Posts: 2074
      • http://www.carl-eike-hofmeister.de/
Collision Model
« Reply #25 on: June 28, 2003, 02:39:58 PM »
I think most unfair collisions happen when you crash into a plane that is doing a hard maneuver just before the crash.

Don't disable collisions completely, just disable them, when the G load of the enemy plane changed largely shortly before crash.

Offline AKIron

  • Plutonium Member
  • *******
  • Posts: 12772
Collision Model
« Reply #26 on: June 28, 2003, 02:45:04 PM »
We need a "dead horse" forum, Skuzzy?
Here we put salt on Margaritas, not sidewalks.

Offline Karnak

  • Radioactive Member
  • *******
  • Posts: 23047
Collision Model
« Reply #27 on: June 29, 2003, 12:21:42 AM »
Quote
Originally posted by MANDOBLE
Quote
Originally posted by Murdr
Delay damage pending collision confirmation via both FE.
Did A and B see collision?
No= No collision
Yes= collision, assign damages.

This is raw common sense, why it is not implemented after years? who knows.


Because it is not raw common sense.

This is an idea that has not been fully thought through and would be a disaster.

Basically this idea would have the same results as the "just have no collisions" idea.  There is no practical difference except that once in a great while this idea would produce a collision in an HO situation.  Do to the rarity of this event it would shock both participants.

I would guess that 99+% of the time the collision only happens on one player's FE (Mind you, if it happens on both player's FEs both players get killed, its just that the poor ping player gets the kill.  Whoopdee doo.) and would under this proposal result in no collision for either player.

Under this proposal I could fly right through a B-17 at the fuselage/wing juncture blazing away with a guaranteed chance of killing him and I would never suffer a collision.



Here is HiTech's response to a similar suggestion by lemur:

Quote
Originally posted by lemur
There are 4 possible ways of handling this:

1) 'Collider' takes damage, 'Target' doesn't. This is the way it's handled now. Downside: It's exploitable (barely), it's unrealistic. Upside: You know you've been killed right away (No delay from your FE telling you you're dead)
2) 'Collider' takes damage, 'Target takes damage': Downside, the 'target', which didn't see a collision on his FE, takes damage. He may have manovered to avoid the collision, but he gets screwed instead. Also, the damage to the target shows up a second AFTER the collision
3) Target takes damage, collider doesn't. This is a stupid idea (I'm just throwing it in here for completeness sake)
4) Neither takes damage unless both sides see the collision on their FE:
Downside: Long delay before damage takes effect (since both sides have to tell the server about the damage, then the server has to come back with a 'Yup, the target agrees that was a collision'), occationally you'll fly right through a guy.
Upside: Not gameable, realistic (up to a point), completely fair.

I'd say options 1 & 4 are the only 'workable' and fair solutions.

In the case of 4 we'd all have to live with collisions that took effect a certain time after the fact ('Collider' + 'target' ping time later) Still, this time shouldn't be more than 1 second later, on average.

So it boils down to this: Either one guy might get screwed in a collision, or nobody gets screwed but we have to live with 'late' damage from collisions.


And HiTech's response:

Quote
Originally posted by hitech
Lemar, only thing I disagree with your analiss is #4,on anything but a head on collision both FEs will never both see a collision.

HiTech


That about sums up why Murdr's idea wouldn't work.
Petals floating by,
      Drift through my woman's hand,
             As she remembers me-

Offline ccvi

  • Gold Member
  • *****
  • Posts: 2074
      • http://www.carl-eike-hofmeister.de/
Collision Model
« Reply #28 on: June 29, 2003, 08:07:15 AM »
Quote
Originally posted by ccvi
I think most unfair collisions happen when you crash into a plane that is doing a hard maneuver just before the crash.

Don't disable collisions completely, just disable them, when the G load of the enemy plane changed largely shortly before crash.


I hate to quote myself, but... Any ideas why this might not work?

Offline SlapShot

  • Plutonium Member
  • *******
  • Posts: 9121
Collision Model
« Reply #29 on: June 29, 2003, 09:43:09 AM »
Quote
Originally posted by ccvi
I hate to quote myself, but... Any ideas why this might not work?


For the same reasons we have now ...

The report of High-G load data from both planes is no different than any data elements that are being sent now to determine if a collision happended or not. So, with netlag, the collision will still happen.

Your just changing the criteria not the environment ... you need to make the ping for all under 10ms ... got any suggestions ?
SlapShot - Blue Knights

Guppy: "The only risk we take is the fight, and since no one really dies, the reward is the fight."