Author Topic: Collision Model Change  (Read 12067 times)

Offline Shane

  • Plutonium Member
  • *******
  • Posts: 8474
Re: Collision Model Change
« Reply #135 on: September 18, 2026, 03:53:17 PM »
I have all that, I was looking for film of collisions so I could show the messages and such. I looked through near 100 films I have and found only 2 collisions and both showed both planes colliding. I rarely get a collision any more as I work to avoid running into people. So the video I posted is the best I could do. Check it out https://www.youtube.com/@AcesHighShorts



See what you can do with this one.  At the end, stepsis (47) claimed a collision which  he is clearly shown to receive one from my perspective (51).  Ntjunior in the seafire clearly flys thru Stepsis taking no damage, while Stepsis loses his entire starboard side.

I can only presume that Stepsis got a collision message that he collided while Ntjunior received no message on his end, flying off unscathed which implies he did not "see it" on his end, due to connection differences.

I posted a 16 second clip:  https://bbs.hitechcreations.com/smf/index.php/topic,413119.0.html

I accidently cut the clip too short where you could see Stepsis claiming a collision.

« Last Edit: September 18, 2026, 03:56:04 PM by Shane »
Surrounded by suck and underwhelmed with mediocrity.
I'm always right, it just takes some poepl longer to come to that realization than others.
I'm not perfect, but I am closer to it than you are.
"...vox populi, vox dei..."  ~Alcuin ca. 798
Truth doesn't need exaggeration.

Offline Wiley

  • Plutonium Member
  • *******
  • Posts: 8362
Re: Collision Model Change
« Reply #136 on: September 18, 2026, 04:12:04 PM »

I can only presume that Stepsis got a collision message that he collided while Ntjunior received no message on his end, flying off unscathed which implies he did not "see it" on his end, due to connection differences.


Correct.  And what you saw of both planes movement was only an approximation, because none of it was on your FE.  The damage or lack thereof gets shown to you if one of their FEs detected a collision.

Any other plane you see, its position is only an approximation of what it is doing on its owner's FE.  It's generally a couple plane lengths back from where it is on the owner's FE, and about a third of a second or so lag on their orientation changes.

Wiley.
If you think you are having a 1v1 in the Main Arena, your SA has failed you.

JG11

Offline Shane

  • Plutonium Member
  • *******
  • Posts: 8474
Re: Collision Model Change
« Reply #137 on: September 18, 2026, 04:20:39 PM »
Correct.  And what you saw of both planes movement was only an approximation, because none of it was on your FE.  The damage or lack thereof gets shown to you if one of their FEs detected a collision.

Any other plane you see, its position is only an approximation of what it is doing on its owner's FE.  It's generally a couple plane lengths back from where it is on the owner's FE, and about a third of a second or so lag on their orientation changes.

Wiley.

It infers that Stepsis and my connections are fairly comparable whereas NTJunior's has a discrepancy enough to not get the message.

The same thing applies to guns - I know when I'm taking "not from my view" hits that are valid from the other players perspective.  It's even evidenced on film where the shells are way behind from perspective, but still doing damage from the other player's perspective. 

I've been all over this for years, since even before you joined.
Surrounded by suck and underwhelmed with mediocrity.
I'm always right, it just takes some poepl longer to come to that realization than others.
I'm not perfect, but I am closer to it than you are.
"...vox populi, vox dei..."  ~Alcuin ca. 798
Truth doesn't need exaggeration.

Offline Wiley

  • Plutonium Member
  • *******
  • Posts: 8362
Re: Collision Model Change
« Reply #138 on: September 18, 2026, 04:34:19 PM »
It infers that Stepsis and my connections are fairly comparable whereas NTJunior's has a discrepancy enough to not get the message.

The same thing applies to guns - I know when I'm taking "not from my view" hits that are valid from the other players perspective.  It's even evidenced on film where the shells are way behind from perspective, but still doing damage from the other player's perspective. 

I've been all over this for years, since even before you joined.

Yep, sorry that was just for the lurkers to explain that there was what happened on each of their machines showing something different, as well as third parties seeing something different as well, but having no impact on what happened to either plane. That's all.

Wiley.
If you think you are having a 1v1 in the Main Arena, your SA has failed you.

JG11

Offline Shane

  • Plutonium Member
  • *******
  • Posts: 8474
Re: Collision Model Change
« Reply #139 on: September 18, 2026, 05:18:09 PM »
Yep, sorry that was just for the lurkers to explain that there was what happened on each of their machines showing something different, as well as third parties seeing something different as well, but having no impact on what happened to either plane. That's all.

Wiley.

It could also have been a situation where I as that 3rd party saw both fly away unscathed if neither of them saw it on their front end despite me seeing both on mine.  Or very likely both could have taken some sort of damage as they saw it on their front end with all us of having comparable connections.

Lag (latency) does come into play in all aspects, HO/rams/collisions, gunnery, and even gv's. HTC has done a great job of smoothing it with coad to the extent they have.

And tbh, it's not that frequent where only one plane takes damage, the more common situation is where one has their wing ripped off while the opponent only clipped their prop and got smoked.  Each plane took the damage they "saw" on their own front end.
« Last Edit: September 18, 2026, 05:29:00 PM by Shane »
Surrounded by suck and underwhelmed with mediocrity.
I'm always right, it just takes some poepl longer to come to that realization than others.
I'm not perfect, but I am closer to it than you are.
"...vox populi, vox dei..."  ~Alcuin ca. 798
Truth doesn't need exaggeration.

Offline Wiley

  • Plutonium Member
  • *******
  • Posts: 8362
Re: Collision Model Change
« Reply #140 on: September 18, 2026, 05:41:44 PM »
My experience for whatever reason, I am very rarely am involved in a 2 sided collision. I would say around 90 percent of collisions I have is one sided, usually me.

The third party thing is also why every once in awhile you will see a plane pull out of a dive and go under the ground for a few yards. On his end he pulled up at tree top level.

Wiley.
If you think you are having a 1v1 in the Main Arena, your SA has failed you.

JG11

Offline Dadtallica

  • Gold Member
  • *****
  • Posts: 2571
Re: Collision Model Change
« Reply #141 on: September 18, 2026, 06:43:04 PM »
My experience for whatever reason, I am very rarely am involved in a 2 sided collision. I would say around 90 percent of collisions I have is one sided, usually me.

The third party thing is also why every once in awhile you will see a plane pull out of a dive and go under the ground for a few yards. On his end he pulled up at tree top level.

Wiley.

No, that’s wrong. Planes can be submarines too! Stop spreading falsehoods!
October Player Poll Results

Back in 2022 after a loooooong break from 2010. Old name Ratpack, same for the BBS. Old squads, Excaliburs, 172nd Rabid Dawgs.

Offline The Fugitive

  • Plutonium Member
  • *******
  • Posts: 19202
      • Fugi's Aces Help
Re: Collision Model Change
« Reply #142 on: September 18, 2026, 08:06:17 PM »



See what you can do with this one.  At the end, stepsis (47) claimed a collision which  he is clearly shown to receive one from my perspective (51).  Ntjunior in the seafire clearly flys thru Stepsis taking no damage, while Stepsis loses his entire starboard side.

I can only presume that Stepsis got a collision message that he collided while Ntjunior received no message on his end, flying off unscathed which implies he did not "see it" on his end, due to connection differences.

I posted a 16 second clip:  https://bbs.hitechcreations.com/smf/index.php/topic,413119.0.html

I accidently cut the clip too short where you could see Stepsis claiming a collision.

Thanks, I have my video done now so Im thinking its good for now. I have the film should I need to add anything.

Offline hazmatt

  • Gold Member
  • *****
  • Posts: 2592
Re: Collision Model Change
« Reply #143 on: September 19, 2026, 10:37:18 AM »
Because it was created to show people who didn't understand the collision model so they could easily see how it works.  With both planes maneuvering it gets messier and harder to follow.

Wiley.

I guess the sarcasm wasn't obvious. Sorry. I was goofing with the questions.

Good video

Offline CptTrips

  • Plutonium Member
  • *******
  • Posts: 9971
Re: Collision Model Change
« Reply #144 on: September 19, 2026, 10:53:21 AM »
I guess the sarcasm wasn't obvious. Sorry. I was goofing with the questions.

Good video

It's a good video.  It explains it well.  I wasn't sure what the original was showing me but it makes good sense with the explanation.

Anyone that could look at that, and say they think that Jug should have taken collision damage from a plane 6 plane-lengths behind them in their view is just stupid, brain-dead and moronic.

But it still won't make a dent on the knuckle-heads that need that info most.

You can fix uninformed, but you can't fix stupid.




« Last Edit: September 19, 2026, 11:17:42 AM by CptTrips »
Toxic, psychotic, self-aggrandizing drama queens simply aren't worth me spending my time on.

Offline Wiley

  • Plutonium Member
  • *******
  • Posts: 8362
Re: Collision Model Change
« Reply #145 on: September 19, 2026, 10:54:37 AM »
I guess the sarcasm wasn't obvious. Sorry. I was goofing with the questions.

Good video

Woops...  :)  I've often said sarcasm needs its own font.

Wiley.
If you think you are having a 1v1 in the Main Arena, your SA has failed you.

JG11

Offline LCADolby

  • Platinum Member
  • ******
  • Posts: 7954
Re: Collision Model Change
« Reply #146 on: September 19, 2026, 01:07:13 PM »
Woops...  :)  I've often said sarcasm needs its own font.

Wiley.

It does
YouTube+Twitch+KICK - 20Dolby10

Offline xanax

  • Silver Member
  • ****
  • Posts: 805
      • Awful's Youtube Channel
Re: Collision Model Change
« Reply #147 on: September 19, 2026, 01:55:20 PM »

Sarcastic text generator:

https://varunpatil.github.io/Sarcastic-Text-Generator/

YoU ArE CoOl!

Offline Animl-AW

  • Platinum Member
  • ******
  • Posts: 7658
Re: Collision Model Change
« Reply #148 on: September 20, 2026, 05:50:08 PM »
Beat me like a dead horse, nothing ventured, nothing gained.

Digital Combat Simulator
A mid-air collision can shear off wings, destroy control surfaces, or trigger a catastrophic explosion for both players depending on the angle and speed of impact.

IL-2 Sturmovik Series (Great Battles / 1946)
If two players collide during a dogfight or formation flight, the sim calculates the structural impact on both airframes. You will routinely see both planes lose wing sections, trade paint, or go down in flames together.

War Thunder (Simulator Battles Mode)
"Ramming" is completely physical. Ramming or accidentally clipping another player mid-air will break the wings, tails, or propellers of both participating aircraft.

Aces High uses Netcode

Netcode Variables: Because Aces High relies on server-side authority to determine real-time positioning, a slight latency difference can occasionally make a collision look strange on your screen (e.g., seeming like you barely touched but taking fatal damage). However, mathematically, the engine always applies the kinetic energy and structural impact limits to both hitboxes.

To make netcode produce mutual damage during a mid-air plane collision, you must implement deterministic collision resolution that overrides the local authority of a single client. Because network latency creates "network lag" (where each player sees themselves ahead of the enemy), relying on a simple "if I hit you, you take damage" system will often result in only one plane exploding or delayed, unfair deaths.

Here is how to structure netcode to ensure both planes take damage reliably.

1. Server-Authoritative Verification (The Gold Standard)
The most robust solution is to let the server make the final decision based on a synchronized timeline.

Client-Side Prediction: Both clients fly smoothly and detect the collision locally for immediate visual feedback.

Timestamped RPCs: When a client detects a collision, it sends a Remote Procedure Call (RPC) to the server containing the precise simulation tick or timestamp and the IDs of both planes.

Server Lag Compensation: The server rewinds its simulation history to that specific timestamp. If both planes were close enough to collide at that exact moment in game-time, the server applies mutual damage to both entities and broadcasts the explosion to everyone.

2. Handling the "Dead Man Flying" Problem

A common bug in mutual collision netcode is when Client A kills Client B a split-second beforethe collision happens on Client B's screen. On Client B's screen, they rammed a dead plane but took no damage because the server already registered the enemy as destroyed.

To fix this, implement a "Grace Period" for Latency:
If a plane is destroyed by projectiles or a collision, keep its physical hitbox active on the server for an extra 100–150ms (matching the average network round-trip time).

Even if Plane A is technically "dead," if Plane B rams into its burning wreckage within that tiny window, the server rules it a mutual collision and destroys Plane B as well.

Implementation Breakdown
Feature   How It Helps Mutual Damage

Deterministic Ticks: Ensures both clients and the server agree on when the planes occupied the same coordinates.

Lenient Hitboxes: Slightly inflating the collision boundary only for ramming calculations helps account for minor positional network sync errors.

Local Visuals, Delayed Logic

Explode the plane locally immediately so it feels responsive, but let the network health sync happen a few milliseconds later.


A useful rule is:

if server_collision(A, B):
    relativeVelocity = A.velocity - B.velocity
    impactSpeed = length(relativeVelocity)

    damageA = calculateDamage(A, B, impactSpeed)
    damageB = calculateDamage(B, A, impactSpeed)

    applyDamage(A, damageA)
    applyDamage(B, damageB)

    broadcast(A, damageA)
    broadcast(B, damageB)

Preventing common multiplayer problems
Don't do this:


Client A detects collision
    → damages B

Client B detects collision
    → damages A

With latency, one client might see the collision before the other, or both might register it at slightly different positions/times.

Instead, use a single server collision event:

collisionId = generateCollisionId(A, B)

if collisionId not in processedCollisions:
    processedCollisions.add(collisionId)

    damage(A)
    damage(B)
« Last Edit: September 20, 2026, 05:53:57 PM by Animl-AW »
AH Tech Hangar
Windows Gaming Tweaks

Offline CptTrips

  • Plutonium Member
  • *******
  • Posts: 9971
Re: Collision Model Change
« Reply #149 on: September 20, 2026, 05:59:56 PM »
 :rofl
Toxic, psychotic, self-aggrandizing drama queens simply aren't worth me spending my time on.