Author Topic: Collision Model  (Read 25125 times)

Offline Shuffler

  • Radioactive Member
  • *******
  • Posts: 27091
Re: Collision Model
« Reply #390 on: January 13, 2012, 03:36:43 PM »
Lets say the other computer tells you where he is once every second (the send rate is irreverent for this discussion). Your computer now simply draws him at that position each time it receives the packet. What you would see on your computer is his plane jump from 1 position to another every second.

Obviously you do not wish to see this so instead of just position he also sends his current speed and direction.

Now instead of seeing him in 1 position every second your computer continues to move him at the speed and direction he received.

Now say he starts a turn immediately after sending you a packet. Your computer would be drawing him in a straight line, but when the next packet came in he would not be where your computer expected him. So his new position would have to be drawn.

Now as long as everything is consistent, you will receive a packet every second no mater how long it takes for it to get to you , Lets say a 5 second lag, you will get the 1st packet 5 seconds later, but after that you will receive a packet every second and things will be just as smooth as if there was a 0 second lag.


Now what happens if some times the lag is 1 second and some times 5 seconds. Or if you receive 1 packet and then 4 packets are lost , and then you receive a packet again. Your computer will continue to display him using the last packet received information. When the new one comes in many seconds later it must correct for the error where it is drawing him.

This is what causes a warp. Dropped packets or inconstancy latency/lag. Ah is extremely tolerant of both latency changes and dropped packets. But when those changes are in the many seconds range, things start to warp.

What I have described is a very simple version of what AH really does, the total complexity and precision needed to make planes look smooth is far outside the scope of this discussion.

HiTech









That was very well put into laymans terms.
80th FS "Headhunters"

S.A.P.P.- Secret Association Of P-38 Pilots (Lightning In A Bottle)

Offline Chilli

  • Platinum Member
  • ******
  • Posts: 4278
Re: Collision Model
« Reply #391 on: January 13, 2012, 04:29:24 PM »
That was very well put into laymans terms.
:aok

Especially, since I hadn't even considered that information about speed so forth and so on, being sent with each packet.  That was an AHAH!!!  moment  :bolt: and answered many more questions that I even thought to ask.  It all is still magical to me, even when you show me some of the smoke an mirrors.

Thanks for the input.

Offline Daddkev

  • Silver Member
  • ****
  • Posts: 1727
Re: Collision Model
« Reply #392 on: January 13, 2012, 05:40:41 PM »
 :O :O :O Wow..! I,m dumb and Naked! :bhead :bhead :bhead
God Bless America
Go tell Momm, im flying! and make me a sandwich !
EvilKev

2012 68KO Cup 1st Place finisher

Offline Hopper

  • Silver Member
  • ****
  • Posts: 1421
Re: Collision Model
« Reply #393 on: January 13, 2012, 05:42:33 PM »
So, I guess I should give you some test time before explaining why speed alone isn't a valid criteria?

There are a couple of obvious reasons that come to mind immediately, that I think would render much effort towards data-gathering a waste of time.  Then again, gathering your own data would probably be the easiest way for you to understand why it wouldn't be a "fair" criteria to use in an effort to assign "fault".

I'm all ears, I hate wasting time.
Hopper


JG/11

Offline Ack-Ack

  • Radioactive Member
  • *******
  • Posts: 25260
      • FlameWarriors
Re: Collision Model
« Reply #394 on: January 13, 2012, 05:44:03 PM »
That was very well put into laymans terms.

Pyro's been posting under HiTech's name again.

ack-ack
"If Jesus came back as an airplane, he would be a P-38." - WW2 P-38 pilot
Elite Top Aces +1 Mexican Official Squadron Song

Offline Hopper

  • Silver Member
  • ****
  • Posts: 1421
Re: Collision Model
« Reply #395 on: January 13, 2012, 05:45:14 PM »
Well, in a multiplane engagement, the probability of getting in a collision naturally increases just as the chance of getting shot down increases.  You are ultimately in control of where your plane goes, but even if you enable perfect ACM against 3-4 bogeys, you will still die a high percentage of the time because ultimately it is out of your control.  There is an element of luck.  Same goes with collisions in a multiplane environment.  That's not the game's fault, imo.

<edit> It's equivalent to saying that ACM's are bogus because I employed them perfectly against 3 cons and still died.

I'm not blaming the game, thats been my entire problem throughout the thread I'm blaming the other players or even me.  I want to assign blame to who caused the collision.  I don't like being in a 3-4 con engagement getting rammed by 1 of them and they fly away with a missing elevator and landing a kill.  
Hopper


JG/11

Offline FLS

  • AH Training Corps
  • Plutonium Member
  • *******
  • Posts: 11617
      • Trainer's Website
Re: Collision Model
« Reply #396 on: January 13, 2012, 05:48:16 PM »
Pyro's been posting under HiTech's name again.

ack-ack

When you have a kid everything changes.  :D

I'm not blaming the game, thats been my entire problem throughout the thread I'm blaming the other players or even me.  I want to assign blame to who caused the collision.  I don't like being in a 3-4 con engagement getting rammed by 1 of them and they fly away with a missing elevator and landing a kill. 

It's clear that you want to assign blame to people who didn't collide.


Offline Babalonian

  • Platinum Member
  • ******
  • Posts: 5817
      • Pigs on the Wing
Re: Collision Model
« Reply #397 on: January 13, 2012, 05:50:06 PM »
:O :O :O Wow..! I,m numb and Naked! :bhead :bhead :bhead

fixt   :aok
-Babalon
"Let's light 'em up and see how they smoke."
POTW IIw Oink! - http://www.PigsOnTheWing.org

Wow, you guys need help.

Offline Ack-Ack

  • Radioactive Member
  • *******
  • Posts: 25260
      • FlameWarriors
Re: Collision Model
« Reply #398 on: January 13, 2012, 05:55:37 PM »
I'm not blaming the game, thats been my entire problem throughout the thread I'm blaming the other players or even me.  I want to assign blame to who caused the collision.  I don't like being in a 3-4 con engagement getting rammed by 1 of them and they fly away with a missing elevator and landing a kill.  

Remember when you point a finger in blame, you have three fingers pointing back right at you.

ack-ack
"If Jesus came back as an airplane, he would be a P-38." - WW2 P-38 pilot
Elite Top Aces +1 Mexican Official Squadron Song

Offline Bear76

  • Platinum Member
  • ******
  • Posts: 4161
Re: Collision Model
« Reply #399 on: January 13, 2012, 06:06:44 PM »
Remember when you point a finger in blame, you have three fingers pointing back right at you.

ack-ack

How do you fly with one finger and a thumb??









Had to sorry  :D

Offline hitech

  • Administrator
  • Administrator
  • *****
  • Posts: 12398
      • http://www.hitechcreations.com
Re: Collision Model
« Reply #400 on: January 13, 2012, 06:30:56 PM »
Pyro's been posting under HiTech's name again.

ack-ack

Even a broken clock is correct some times.

And if you want some details of one way to do it.

http://patft.uspto.gov/netacgi/nph-Parser?Sect2=PTO1&Sect2=HITOFF&p=1&u=/netahtml/PTO/search-bool.html&r=1&f=G&l=50&d=PALL&RefSrch=yes&Query=PN/6042477

HiTech

Offline Ack-Ack

  • Radioactive Member
  • *******
  • Posts: 25260
      • FlameWarriors
Re: Collision Model
« Reply #401 on: January 13, 2012, 06:32:41 PM »
How do you fly with one finger and a thumb??









Had to sorry  :D

I use Kinect, though my arms get tired quickly

ack-ack
"If Jesus came back as an airplane, he would be a P-38." - WW2 P-38 pilot
Elite Top Aces +1 Mexican Official Squadron Song

Offline FLS

  • AH Training Corps
  • Plutonium Member
  • *******
  • Posts: 11617
      • Trainer's Website
Re: Collision Model
« Reply #402 on: January 13, 2012, 07:17:10 PM »
Even a broken clock is correct some times.

And if you want some details of one way to do it.

http://patft.uspto.gov/netacgi/nph-Parser?Sect2=PTO1&Sect2=HITOFF&p=1&u=/netahtml/PTO/search-bool.html&r=1&f=G&l=50&d=PALL&RefSrch=yes&Query=PN/6042477

HiTech

Well that explains it. Thanks for the link.  :D

Offline mtnman

  • Gold Member
  • *****
  • Posts: 2438
Re: Collision Model
« Reply #403 on: January 13, 2012, 10:43:25 PM »
I'm all ears, I hate wasting time.

Ok, I honestly didn't expect that response...

If you model collision "fault" based on speed alone, it becomes an unbalancing factor and will have effect far beyond the fights where collisions occur.

If you decide that collisions are the "fault of the fast guy", you'll destroy any chances of "fair" close fighting between a fast plane type and a plane that's inherently slower than that.  An extreme example would be between a P51 or 262 and a Spit1.  If the P51 pilot knows his stuff, he'll seldom (if ever) be as slow as his target, so he'd have a much better chance at being found "at fault" simply because of the plane he chose to fly.   And if the pilot of the slower plane type knows that he'll be held faultless for any collision that occurs (simply because he's slower) why would he avoid colliding?  What would keep him from using collisions to his advantage?  This would put an unfair responsibility to avoid colliding on the faster pilot, while removing that responsibility for the slower one.

Conversely, if collisions were always the "fault of the slow guy" imagine how reckless that high P51 could be when he dives in on your slower plane...  He'd have no fear of colliding, because it would always be the slower guys fault, right?

Amongst identical plane fights, why should the guy who's a few mph faster than the other be at fault?  Why should the guy who's a few mph slower be to blame?  Energy conservation is a deciding factor in fights, but in the case of collisions it would reward the guy with less energy. 

A collision model based on speed would shift fights further away from reality.

You cannot place blame on a pilot because they have a different speed than someone else.  You also cannot place blame based on angle of intersect, or even based on who's airplane entered someone else, or one what part of a plane entered what part of the others.

MtnMan

"Those who hammer their guns into plows will plow for those who do not". Thomas Jefferson

Offline des506

  • Nickel Member
  • ***
  • Posts: 674
      • The 354th FG
Re: Collision Model
« Reply #404 on: January 14, 2012, 07:32:48 AM »
another case study scenario...

i get chased by someone... i am flying 800 infront of him... and i decide to go vertical to lure him up for friendlies to finish him off...he follows me up... i forget at the moment that he has a more superior climb rate than my aircraft...i maintain a straight and steady steady climb, i have maybe 100 more knots before i stall...and he closes in behind me rapidly 400, 200, ... i dun ever bother to struggle to roll or turn...i close my eyes in disappointment in my stupid tactic and the next thing i know i hear a strange bang( not the expected gunfire sound) when i open my eyes... i  saw that i got a collision msg... and i miss a wing, and engine oil... he collided into me but i got the msg.. and he flies home with a free kill( no doubt he deserves the kill for getting me into this sorry predictcament, :salute not sure if he was awarded perkies or not although i think he did... the system basically gave him that kill)... i found out later that he had run out of bullets, and decided to ram me...he was def not trying to avoid me thats for sure.

these are the descrepancies that happened many times to me....

so when you try to relate that to what you guys are saying about that my computer sees, it is definately not what actually happened.. and then it becomes my fault for not avoiding that collision? how ludicrious is that????  
« Last Edit: January 14, 2012, 07:43:40 AM by des506 »
DES 354th FG
The men dying out there have no choice... i have..i cannot order them into battle... i can perhaps lead them...Help them....Die with them
Manfred von Richthofen