Aces High Bulletin Board

General Forums => Aces High General Discussion => Topic started by: aknimitz on October 02, 2001, 09:24:00 PM

Title: Collision Model Inquiry
Post by: aknimitz on October 02, 2001, 09:24:00 PM
Ok.  I just would like a few things explained/clarified, if possible.  I could have swore today I heard HT say something along the lines of the slower your connection, the more likely you will survive a collision.  Maybe I heard wrong, but I thought I heard this.  If this was said, why is this the case?

Second, why have collisions?  If we are being "realistic" and that was a "realistic" part of fighting, then it certainly seems fair to have friendly collisions on as well.  Perhaps there would be a little less steal killing going on then.

So, I am sure there is a good reason why we have enemy collisions on, but could someone tell me?  

What sparked my frustration is I am flying along fighting a 109 and his engine quits.  I was in a vert scissor and came back down and obviously he was able to come up quicker (albeit briefly).  We collided, and of course, I die.  I ALWAYS, literally ALWAYS lose on collisions.  Then, HE gets the kill!  I put around 50 rounds into him, he none into me.  

In any event, just curious about the collision model.

Nim
Title: Collision Model Inquiry
Post by: moose on October 02, 2001, 09:43:00 PM
i think that if you see the collison on your FE, you die.

a faster connection will collide before a slower connect will. slower connects usually have the planes like 100-200 yards from where they really are?
Title: Collision Model Inquiry
Post by: streakeagle on October 02, 2001, 09:46:00 PM
Why can't the server get a message from the FE indicating the collsion and mirror the results to the other server? A little laggy, but better than the one-sided events that usually happen with the current system. Is it any more laggy than sending bullet collsion information? After all, your aircraft is just one big slow guided bullet.
Title: Collision Model Inquiry
Post by: Boozer on October 02, 2001, 10:02:00 PM
*So, I am sure there is a good reason why we have enemy collisions on, but could someone tell me?*

  So the HO dweebs pay for their lack of ACM
Title: Collision Model Inquiry
Post by: Wotan on October 02, 2001, 10:19:00 PM
I know fer a fact the lower the ping quicker ya die in colision. I have never survived 1. I ping 30 to at highest 60.

Others have said that as well but get the old "shut whiner treatment".

Also there seems sometimes to be a delay when I get good pings on certain people. Especially with gvs. I'll put my bomb dead on and as I pull out and away then the gv blows up. This doesn't happen all the time so I assumed it was bad connection on the other guy. Sometimes they blow right up.

As for friendly colisions being off I understand that or you would die everytime 2 guys spawned at the same time same spawn point or on the rearm pad.........
Title: Collision Model Inquiry
Post by: Blue Mako on October 02, 2001, 10:38:00 PM
Full discussion in this (and many other) threads.  http://www.hitechcreations.com/cgi-bin/ultimatebb.cgi?ubb=get_topic&f=8&t=001293 (http://www.hitechcreations.com/cgi-bin/ultimatebb.cgi?ubb=get_topic&f=8&t=001293)

--------------------

  (http://home.iprimus.com.au/melandgreg/AH/sig-blue-3.jpg)   (http://webpages.charter.net/davidlj/)
Click on image to go to the 412th home page...

[ 10-02-2001: Message edited by: Blue Mako ]
Title: Collision Model Inquiry
Post by: whirl on October 02, 2001, 11:04:00 PM
this would explain why i have such a terrible time getting my rounds to hit--this has to be fixed.  i have virtually no delay or lag.  1.4 meg connect.  i'm constantly maneuvering for kills only to not getem cause i'm seeing where he is supposed to be, even thoughhe's not gotten there yet. heh.. my ammo is hitting empty space because of anothers laggy front end.  there has to be a way to balance this out so faster connects can deal with slower connects.  or raise the min requirement to dsl connects<BFG>
Title: Collision Model Inquiry
Post by: Blue Mako on October 02, 2001, 11:26:00 PM
Quote
Originally posted by whirl:
this would explain why i have such a terrible time getting my rounds to hit--this has to be fixed.  i have virtually no delay or lag.  1.4 meg connect.  i'm constantly maneuvering for kills only to not getem cause i'm seeing where he is supposed to be, even thoughhe's not gotten there yet. heh.. my ammo is hitting empty space because of anothers laggy front end.  there has to be a way to balance this out so faster connects can deal with slower connects.  or raise the min requirement to dsl connects<BFG>

Sorry whirl, it doesn't work like that.  You shoot at people on your front end (FE).  If you hit them, your FE sends a hit packet to the server and then the con getting hit is sent the hit packets by the server.

If you are not hitting people I'm afraid it means you aren't a very good shot  ;) <G>

If you are seeing hits on your end and the con is not getting damage, you may have other "rubber bullet" problems...  In any case, hit calcs are done on the shooters FE...
Title: Collision Model Inquiry
Post by: SKurj on October 02, 2001, 11:56:00 PM
If collisions were mirrored so that both players took damage then ramming would raise its ugly head in the game, we can't have that.  I am on a fairly decent connect (cable) and I too quite often suffer the worst in a collision.
 Just tonight Shane and I were in a vertical scissor both hovering on the edge of stall when we collided yet he took no damage and down I went a flaming wreck.
Friendly collisions bad idea, even if they only became active once a plane's wheels left the ground, we would have to deal with friendly rammers... or at least the accusations.
We need enemy collisions, and as far as i am concerned, for the time being anyways, I think HTC have the best solution.  It does happen where the 2 players involved do go down, but lately it seems to be a rare occurence, its one or the other mostly.  I can accept this, because I've seen some of the alternatives...


SKurj
Title: Collision Model Inquiry
Post by: whirl on October 03, 2001, 02:52:00 AM
i guess thats the point mako, i snap for a point blank range canopy shot, that should kill anyone, and nothing happens.....  there's a problem somewhere and it needs to be fixed.
Title: Collision Model Inquiry
Post by: FDutchmn on October 03, 2001, 03:38:00 AM
Hiya!

First off... on friendly collisions...

I think we wouldn't be able to take off or land decently with the way the game is played right now... think about it, ya get on a zeke and go to the run way and a Lanc or B17 mows you down   :D

I am not the designer of this game so I am only speculating... but I think that it is not just realism that we are dealing with.  Playability also counts.  :)

I don't know about survivability in a collision due to connection speed.  I have thought that there was some sort of algorithm in the program to decide who the winner is.  I have also wondered how this was programed.  Like when I was following a spit in a scissors, I was pinging away on the guy and his wings were about to fall off.  Except, he banks right into me and we collide.  He won the battle according to the message.  

Flying Dutchman
Title: Collision Model Inquiry
Post by: ~Caligula~ on October 03, 2001, 04:05:00 AM
Collisions in AH suck.
If You collide both should take damage...period
Title: Collision Model Inquiry
Post by: Apar1 on October 03, 2001, 04:07:00 AM
I have a cable connection with good connection to HTC servers and I'm always the one to die during collision while my opponent has'nt got a scratch. When I asked the opponent whether he saw the collision as well I often get the answer that he didn't collide (he didn't experience a collision), go figure. Now that is not a good thing for playability at all.
Title: Collision Model Inquiry
Post by: FDutchmn on October 03, 2001, 04:25:00 AM
oops i think i was bit unclear about the word "playability".  I meant that for Friendly Collisions on this thread, not for collision survivability.

Flying Dutchman
Title: Collision Model Inquiry
Post by: fats on October 03, 2001, 06:00:00 AM
Shooting:

Your FE detects the hits on an object, the location you must shoot at corresponds with what you see on your screen. Once there is hit, a packet is sent indicating such event, and after awhile you get a return packet telling what was destroyed if anything from the object. Then you will see wings fall off or the object explode, albeit with a little delay.


Collisions:

Your FE detects wether you have collided with anything, leave out ( enemy or friendly ) bullets, bombs and rockets ( ? ) since that is done by others' FEs. This implementation makes the collision parameters consistent. When you see ( your FE detects ) your self flying into something, you die/get damaged.
Why many of you think you 'lose' every collision is that you have no idea when the other guy collides with you most of the time.

Imagine a common HO, where you both come in guns blazing and connect a few hits. Then you switch to six view and see the enemy disintegrate mid air, and get the feeling that burst you got in must have hit fuel or something important. Have you ever considered that he just _collided_ with you and your hits in fact did no mortal damage?

To anyone smocking crack and playing AH... err thinking that _both_ should die/get damaged if either FE detect a collision. Imagine the previous sittuation again. You would have now died without any reason. Also because of varying pings ( positions ) there would be no consistency when you would collide and when you wouldn't. At times you could fly just 10yds from an other plane, but  the next time it might be 200yds and you still die from a collision for no obvious reason. This implementation would occur in most frequent collisions, so think carefully what you wish for. If you think you die too often in a collision now...

If you say that both die/get damaged when _both FEs_ detect a collision you still fail to grasp how varying pings would affect the sittuation. Collisions would be less frequent, but they would still happen under inconsistent circumstances. It would mean at times you can fly straight through the other player and at times you couldn't. Most would agree this isn't very good sollution either?


Ping time and collisions:

player A has ping 10.
player B has ping 400.

A and B start flying towards each other in a HO. At the last second B pulls up, A doesn't because he figures B is chicken and he can plough through. On B's FE B misses A slightly. On A's FE B is still flying _straight_, because the fact that B has pulled up comes 410 units later, and A now collides with B. Before you say A died unjustly for having a better connection, remember A did very well see what was about to happen. A didn't do enough to avoid the collision and died/got damaged justly.

A and B fly in a tail chase. A is chasing B's   plane. Both fly perfectly straight. A will again die in this sittuation before B. But again it's not that he was penalized for better connection. A saw that range indicator to go to 5, or how ever long his plane's nose is, before receiving damage and would have been able to at least maneuver at some point prior to that.


The point being, with the current implementation when you die in a collision it is your own fault. The other guy didn't ram you or anything, because that is simply put quite impossible. Ramming someone _on purpose_ requires you to know both player's ping, and then be able to calculate in your head how to maneuver for your aircraft to hit on the other player's FE.


// fats
Title: Collision Model Inquiry
Post by: Apar1 on October 03, 2001, 06:24:00 AM
Quote
A and B start flying towards each other in a HO. At the last second B pulls up, A doesn't because he figures B is chicken and he can plough through. On B's FE B misses A slightly. On A's FE B is still flying _straight_, because the fact that B has pulled up comes 410 units later, and A now collides with B. Before you say A died unjustly for having a better connection, remember A did very well see what was about to happen. A didn't do enough to avoid the collision and died/got damaged justly.

In a realtime scenerio player A would have been alife BECAUSE of the fact he saw B pull away and A decided to fly straight because he would be ok then, hence they wouldn't collide, but in AH with different pings the players A's FE decided he collided despite the fact that players B's FE decided they didn't. That is the whole point (as far as I'm concerned).

If both FE's would be correlated correctly by the HTC server that wouldn't happen. If this correlation would take a reasonable average lag time into account it would rule out (to a certain amount and dependend on the choice for that correlation interval) one players FE detecting collision and the others FE not.

IMHO, if you try to simulate a realtime process you should try to do that as close as possible (all within the limits of computation power, server capacity and internet conection issues ofcourse)

How do you explain the following; I'm trying to B&Z somebody that is lower than me and also has lower E than me by the time I'm getting into gun range. He sees me and decides to pull up. I see that I will collide with him and try to avoid that by pulling to the left or right. Because I'm at a much higher E state than him my turn radius is much bigger than his, i.e. I can't avoid the collision. Now guess what, I die and he didn't, he didn't even damage. This happened to me several times.

Maybe HT can comment on this whole issue.

[ 10-03-2001: Message edited by: Apar1 ]
Title: Collision Model Inquiry
Post by: fats on October 03, 2001, 07:40:00 AM
--- Apar1 ---
In a realtime scenerio player A would have been alife BECAUSE of the fact he saw B pull away and A decided to fly straight because he would be ok then, hence they wouldn't collide,
--- end ---

In real time scenario the A pilot reacted to what he saw: he saw B pull to a side and decided he can fly straight to avoid colliding.

In AH scenario the A pilot reacted to what he saw: he saw B fly straight and decided to pull to a side to avoid colliding

Both scenarios result in no collision. If the AH scenario _does_ result in a collision it is pilot A's fault, since he didn't apply correct maneuvering according to the sittuation he saw.

--- Apar1 ---
If this correlation would take a reasonable average lag time into account it would rule out (to a certain amount and dependend on the choice for that correlation interval) one players FE detecting collision and the others FE not.
--- end ---

How would _totally inconcistent_ collision circumstances improve playing or realism? You could never know if flying through an enemy will result in a collision or not. The way it is implemented now gives you consistent parameters for collision: your plane comes in contact with an enemy plane will cause your plane damage.


--- Apar1 ---
How do you explain the following;...
--- end ---

You screwed up. How bad is it to admit that you are the fault in sittuations like this? If you had flown the sittuation correctly you would have left enough room to avoid a break maneuver in any direction - including flying straight.

Now with the system you propose you would never be able to learn what are the safe parameters for a high speed bounce, because there would be no solid parameters to base your learning on. At times you would survive a fly through and at times you wouldn't. And I am 100% sure you wouldn't identify all those times you in fact came in contact with someone this way.


// fats
Title: Collision Model Inquiry
Post by: Westy MOL on October 03, 2001, 07:42:00 AM
Just don't collide. It's not as if it occurs ever flight - well, if it does then maybe online flight school time is warranted.       ;)
 AH collisions is one case of the very few times in online gaming where a faster connection hurts the LPB type.

 As for why enemy collisions are on? Well for one it makes the AH virtual pilot fly cautiously so they do not ram thier opponant (or opponants debri). That is more realistic if you think about it. IMO the way AH does it is simply the best to date and you should not be able to fly through your opponant while holding the trigger down shooting at them. If you want an example of a completely bogus setup go fly AW for a week or two.

 As for Apars example as to why the enemy did not suffer damage? It's simple, he rammed the enemy bogy. Sure, in RL there were always two victims in a collisions. Well, this is not real life. This is the internet. With car bombing, pork and auger dweebs, the friendly fire frag fest seen in the TA the other week and other "gameing the game" baloney that occurs online that if mutual destruction occured in collisions there would be a thousand fold increase in ramming. The "win at any cost" type of attitude. If HTC could force us all to be in the military, make us all complete real pilot training, introduce court martialialing for those who do not follow orders and implement permanent death then maybe the human survival instinct would suffice to cut down on  any collision problem. But that will never happene. So in the mean time because some idiot has a problem controlling his "aircraft" and rams me some folks are saying I should "die" too? fediddle off with that. I do not want to be penalised because of someone elses ineptness or bad internet connection.

 Westy

[ 10-03-2001: Message edited by: Westy MOL ]
Title: Collision Model Inquiry
Post by: Tilt on October 03, 2001, 07:55:00 AM
Quote
Originally posted by aknimitz:
Ok.  I just would like a few things explained/clarified, if possible.  I could have swore today I heard HT say something along the lines of the slower your connection, the more likely you will survive a collision.  Maybe I heard wrong, but I thought I heard this.  If this was said, why is this the case?

Second, why have collisions?  If we are being "realistic" and that was a "realistic" part of fighting, then it certainly seems fair to have friendly collisions on as well.  Perhaps there would be a little less steal killing going on then.

So, I am sure there is a good reason why we have enemy collisions on, but could someone tell me?  

What sparked my frustration is I am flying along fighting a 109 and his engine quits.  I was in a vert scissor and came back down and obviously he was able to come up quicker (albeit briefly).  We collided, and of course, I die.  I ALWAYS, literally ALWAYS lose on collisions.  Then, HE gets the kill!  I put around 50 rounds into him, he none into me.  

In any event, just curious about the collision model.

Nim

It seems that collisions are FE modelled. Therefore the fact that one can collide and the other apparantly not is obviously a lag issue.

The greater the combined lag of the two FE's via the server the greater the error and so the more likey that one survives and the other not.

Am I right in assuming that we do not actually recieve packets at the same lag as we send them

As to whether it favours a faster connect over a slower one (or vice versa) seems to be a function of the angle of attack so to speak.

Head on I guess it would not make much difference.

Nose to tail is the laggier AC behind or in front?

If I fly across the nose of a bomber can I time it so he hits my  lagging self on his FE when on my FE I have passed his line of flight. In this case the larger lag is in my favour.


Tilt
Title: Collision Model Inquiry
Post by: aknimitz on October 03, 2001, 08:37:00 AM
Thanks guys, seems to make sense.  And seems to be clear that although it is unfortunate, HTC has the best collision model possible all things considered.  Thanks for the input.

Nim
Title: Collision Model Inquiry
Post by: fats on October 03, 2001, 12:36:00 PM
--- tilt ---
As to whether it favours a faster connect over a slower one (or vice versa) seems to be a function of the angle of attack so to speak.
--- end ---

If you think about it, it really doesn't favour either in any major way. High ping and low ping players both play according to what they see.


--- tilt ---
If I fly across the nose of a bomber can I time it so he hits my lagging self on his FE when on my FE I have passed his line of flight. In this case the larger lag is in my favour.
--- end ---
It is possible to fly in such manner, but in any kind of repeated manner I doubt it. Think the sittuation the other way around. You are player with 10 ping, flying accross a bomber's nose ( or even HO ) with 400 ping. On your FE you make a move on the last second ( the point that places you in his flight path ), it still takes 410ms for the lagged bomber to see your plane change flight path. The lag works both ways. Gee - his own lag worked against him.


// fats
Title: Collision Model Inquiry
Post by: Apar1 on October 03, 2001, 12:49:00 PM
Westy, I'm not even gonna bother commenting on the last paragraph,   :)

fats:

 
Quote
In real time scenario the A pilot reacted to what he saw: he saw B pull to a side and decided he can fly straight to avoid colliding.

In AH scenario the A pilot reacted to what he saw: he saw B fly straight and decided to pull to a side to avoid colliding

Both scenarios result in no collision. If the AH scenario _does_ result in a collision it is pilot A's fault, since he didn't apply correct maneuvering according to the situation he saw.

Your explanation makes sense, you react to what you see, but couldn't that situation lead to advantages in the rest of that fight? I.e. I see player B fly straight where on his end he has turned to (for instance) the left. On what I see (him flying straight) I decide to avoid collision and turn right. Both will end up in crossing scenario or other HO in the next merge.
Now if I would have seen him turn left in time, I prolly would chosen a different maneuver after that first pass (follow his turn and try to get the lead?).
I realise this is highly hypothetical but still. The extreme of this situation is when somebody really lags, where you see him in one place and in a split second he jumps 100's of yards to whatever direction, it is almost impossible to fight that player because you cannot predict his maneuvers.

 
Quote
How would _totally inconcistent_ collision circumstances improve playing or realism? You could never know if flying through an enemy will result in a collision or not. The way it is implemented now gives you consistent parameters for collision: your plane comes in contact with an enemy plane will cause your plane damage.

Yes it is consistant but to me it gives the advantage to the player with a bad internet connection.

On the B&Z scenario, yes I screwed up (I do that many times, part of the learning curve) but, would you as player B have tried to turn your plane into the flightpath of a high E approaching enemy (in real life)? And should you as player B not suffer the consequence of that choice?
Title: Collision Model Inquiry
Post by: hitech on October 03, 2001, 12:51:00 PM
There is one way it favors a slower connect. If you both collide on your fe's at same time, and both die in the collision. The packet from person with the faster connect arives at the server first therfore the other guy is awared the kill before his packet arives and he dies.


This is what I was refering to online.


HiTech
Title: Collision Model Inquiry
Post by: hitech on October 03, 2001, 12:54:00 PM
Quote
Yes it is consistant but to me it gives the advantage to the player with a bad internet connection.

Incorect because the lag you see is the sum of both your lags. Therfore you both see eachother with the same amount of lag.


HiTech
Title: Collision Model Inquiry
Post by: AKSWulfe on October 03, 2001, 12:56:00 PM
If people with a slower connect always win HOs, why am I always dying from collisions when I've got a 56K?

Oh, that's right.. because it's totally FE dependant.
-SW
Title: Collision Model Inquiry
Post by: Apar1 on October 03, 2001, 01:01:00 PM
HT,

In a scenario with both planes in a HO, both planes fire at eachother and both planes blow up, is it correct to say that the kill goes to the player with the worsed ping because his "killed message" comes at the server the last?

[ 10-03-2001: Message edited by: Apar1 ]
Title: Collision Model Inquiry
Post by: hitech on October 03, 2001, 01:14:00 PM
Aper1: Isn't near that simple. In one specific case that could happen but you also need to consider the lag on sending the damage. The slower connect hits will arive at the server later than the faster connect, so things get very complex as to what the outcome will be.

BTW: Don't plan on much changing in the basic system, typicly the cure is worse than the problem.

HiTech
Title: Collision Model Inquiry
Post by: Apar1 on October 03, 2001, 01:18:00 PM
Thx for the explanation HT and fats,    :)
Title: Collision Model Inquiry
Post by: Am0n on October 03, 2001, 01:19:00 PM
Apar1,

Ive had that happen to me when i was fighting a bomber, he got my wing and as my air craft flipped over to the right my gun fire ripped his wing off. I died instantly as his wing shattered and we both got credit for killing each other.

Aknimitz-
Friendly collison is a bad idea, no offense. If used responsibly it would add a more realistic effect to the game, you must remeber though it would not be used in that manner. Many players would exploit it and cause all kinds of trouble with it.
Title: Collision Model Inquiry
Post by: Maverick on October 03, 2001, 02:21:00 PM
My $.02 on the collision idea.

Friendly collision. I see no need for it in game play adn it would make it hard to get on or off the ground at a congested airport. No I do NOT want some player being a ground controller telling me when I am cleared for takeoff or landing.

Enemy collision. ONLY fair way I see to handle this is to make it a collision for BOTH players. If one FE senses a collision then both go down. It's a nod of the head to the "realism" crowd and promotes fair play.

Mav
Title: Collision Model Inquiry
Post by: eskimo2 on October 03, 2001, 03:16:00 PM
Well said fats.

<S>

Folks who want both guys to die...
Remember, this is not 100% real time.

You are not seeing the same thing!
One person could screw up and collide while the other avoids the collision.

Imagine:
2 planes, 1 going North, the other West.
The Westbound guy misjudges his attack and collides with Mr. Northbound, but, Mr. North bound views it as a miss by 75 feet.  Why should Mr. Northbound die?  He didn't collide.

In WB, I lost my first 10 kill streak to an unfair collision.  I bounced a P-47 with a P-51.  I tor him up and then peeled off to avoide the bogus collision model.  Even though I missed him by an entire wingspan on my FE, we collided.  

I see no problem with only one person dying in a collision.  I would be very ticked, however, if I got killed 300 feet behind an enemy just because they have a lousy connect.

eskimo
Title: Collision Model Inquiry
Post by: milnko on October 03, 2001, 04:02:00 PM
HT,

So with what you've explained in mind...

Wouldn't it be a more fair situation if you set up the awarding of kills like you did in WB? That if both die in HO or collision each person is awarded a kill? Wouldn't that remove any advantage/disadvantage of netlag?

Or do you feel that setup promotes HOs more often?
Title: Collision Model Inquiry
Post by: Maverick on October 03, 2001, 04:30:00 PM
Eskimo,

If one FE sees collision and other doesn't only FAIR thing is for both to die or neither to die. Either the collision happened or it didn't. Totally unfair for one to suffer consequences and other to have no damage at all. That just ain't "real".  :p

Mav
Title: Collision Model Inquiry
Post by: hitech on October 03, 2001, 05:20:00 PM
Maveric: You simply are not looking at the net effect and the use of the term "fair".

It comes down to what is the least bad effect.

Neather die.
 This would cause collision only to happen in head ons only. Even then one guy could fly threw the other plane, the other guy breaks off to avoid the collision. The guy who flew threw the plane and acted totaly unrealistic gets the advantage because he gets to keep fireing all the way. In the end this would promote flying threw the other guy in a Head On because worst case is you would both die. Best case is you get more hits on other guy do to not breaking off.

In every other angle attack the 2 planes will never both collide. Bombers would suffer the most from this because people comming in from above would dive right threw the buff firing all the time and suffer no consiquences.

Both die.
This would just plane suck. Pickture how mad you will be when you are looking back , a con is 200 yards back and boom, you die from a collision. This is want would happen all the time.

As fare as FAIR goes. The current system is completly fair. It's simple if you hit the other guy you get damaged. If the other guy hits you HE gets damaged. The only thing you are considering not "Fair" realy has nothing to do with fairness, it has everything to do with beeing mad that you screwed up, and think the persone you hit should suffer do to your screw up.


As fare as the lag issue it is very rair that both people collide and are killed in the collison, most offten you end up with missing parts where lag has nothing to do with it.


HiTech
Title: Collision Model Inquiry
Post by: Karnak on October 03, 2001, 05:20:00 PM
Maverick,

I disagree.

The only fair thing to due is to award the pilot who avoided the collision and punish the pilot who did not.

If I avoid your P-51 in my Typhoon and you collide with my Typhoon, how is it fair if I die?  How is it fair if you don't?

 
Quote
From one of my really old posts on this subject:
If both aircraft were destroyed when a collision occurred people would use this to their advantage. If it takes the attacker 10 minutes to get to your base and you 30 seconds to takeoff and collide with him you can be back where you were 30 seconds later but it will take your enemy 10 minutes to regain what he lost from you colliding with him in a way he could not avoid. After all, on his FE you flew across his flight path 300 yards behind him, why should he have dodged?
On the other hand, if there has to be a collision on both FEs in order to for a collision to occur, then people will simply fly straight through their targets,firing all the way (hitting from 10 feet out is easy), without worrying about colliding. After all,I know that the B-17 I am diving on is, on my target's FE, 150 yards ahead of where it is on my FE. I know there won't be a collision, but I will definitely nail him with by guns as I fly through the image of his B-17.

[ 10-03-2001: Message edited by: Karnak ]
Title: Collision Model Inquiry
Post by: hblair on October 03, 2001, 05:23:00 PM
Sounds great HT.

Now, don't let these dewdz keep you from coding.

Back to work!   :D
Title: Collision Model Inquiry
Post by: eskimo2 on October 03, 2001, 05:55:00 PM
Mav,
So if I am putting along in my Tempest on autopilot, with a bad connect, and you make a nice bounce in a 202, but hold your fire till 100 or so where you know your guns can actually do something (you only have 50 rounds left because you have already made 5 kills on this sortie   :) ), and you smack into me just as you open fire (at 100), your OK with you dying in a collision because my FE read you as being much closer?  Is that really fair?
Are you are willing to die at any moment that you get within D=150 of any enemy plane because they might have a bad connect?  That's what will happen!  You will never know when you are about to collide and most people will break off fast closing attacks at 200 in fear of colliding with a ghost!

Until we all have ping times of 10, we all must accept that there is not 1 time related reality in this or any online sim... There are as many realities as there are players.

Fair, is when I make an effort on my perspective and FE to avoid a collision by 100 or even 5 feet, I do not collide.  And, when part of my plane, from my perspective and FE, hits another plane, from my perspective and FE, I collide.

12 of my squad members live in Colorado Springs.  We regularly make 2 to 12 person conference calls and fly together (and use the phone as a radio).  When in formation, 3 people can have completely different perspectives and ranges on squad-mates in a formation.  We talk about this often.  I may read Jarbo at 125 behind me, but his FE may tell him that he is only 50 behind me.  We typically have a separation difference of 35 to 75.  And, we are not all that far from Texas.  
Now when Fariz in Azubakstan (Spelling... sorry), exchanges fire with Mitsu in Japan, imagine what kind of lag they may experience!
You know those impossible shots that some folks make?  The ones where you are getting hit and it looks as if the enemy plane is not even pointing at you?  That's because HIS FE shows a firing solution!  Should that shot not count because your FE didn't agree?
I bounced a guy (from Japan) once and was shooting him up at close range (about 100), then I started hearing pings.  He was shooting me up too.  From his FE, I overshot and was now in front!

My point is that weird things must happen in this game to compensate for net-lag!
One-man collisions stand out because they are one of the most apparent and non realistic results of compensations for dealing with net-lag.  Most of the other effects of net-lag do not stand out and we can almost believe that everything is realtime
   
As odd as it sounds, having sometimes only one plane collide is much more realistic and fair than enabling collisions with unseen ghosts.

IMO, mandatory dual collisions would ruin dogfighting in AH

eskimo

Sorry, when I started my responce, Mav was the last post on this thread...

 :)

[ 10-03-2001: Message edited by: eskimo2 ]
Title: Collision Model Inquiry
Post by: DanielMcIntyre on October 04, 2001, 01:48:00 AM
errr I have to disagree with you pro-1kill collision guys.

I was flying a Typhoon in MA today, HO'ed a Spitfire (capera), collided and died.   I asked if capera had seen the collision, capera replied he had a wheel imprint on his forehead.

So obviously, my FE and his FE both saw the collision but only I died from it?  This is what is unfair about collisions.  He got the kill.  I died?

I think the fairest way of modeling collisions would be if both FE's see the collision, then both planes should die.  No kill should be awarded to anyone unless enough damage was done prior to the collision to a plane to render it unflyable.

Just my opinion thou

  :D
Title: Collision Model Inquiry
Post by: fats on October 04, 2001, 05:28:00 AM
--- Apar1 ---
Your explanation makes sense, you react to what you see, but couldn't that situation lead to advantages in the rest of that fight?
--- end ---

Such is the nature of internet playing. Neither player sees the same thing. Unfortunately there is no way around it, with out making the game unplayable.

Most obvious case where lot of lag between two players gives advantage to someone is a tail-end-chase. It doesn't matter which one of the players is lagging, or if both are. It matters what their combined ping time is. A chasing B. A will always see B closer in his front view than what B sees A in his rear view. That means A is within guns range before B 'knows' it. This is the reason why you are adviced to break at ~900yds and not at 600yds when most people will be able to start scoring solid hits. Try it sometime with a mate. This is the 'obejcts are closer than what they appear in the mirror' of internet gaming. Well it applied to AW which had a mirrored rear view.

As an another note, Quake3 server has an option to synchronize all players with the highest ping player in the game. Everything you do is checked with the server first, then applied. This type of play would be fine with pings around 10ms, but then again the other method would provide just as reliable results in this scenario. How this applies to AH?

This type of setup would mean your controls would lag behind by the highest ping in the game. Someone from Planet Pluto with 500ms ping plays on the same arena with you, your aileron/elevator/rudder/throttle/flaps react to input you gave 0.5 seconds ago. That renders any accurate physics modelling of these WWII planes useless.


--- Apar1 ---
would you as player B have tried to turn your plane into the flightpath of a high E approaching enemy (in real life)? And should you as player B not suffer the consequence of that choice?
--- end ---

It's common ACM when bounced to turn into the attacker - real life and online gaming. The difference with these is that in online gaming you _cannot_ force the attacker to collide with you in any predictable manner due to varying ammounts of lag between the you two. So I still maintain that the 1-sided collision implementation AH has works like expected 99.9% times.

I agree it totally sucks to collide when someone f.ex. warps from one location to another in nanosecond and there is no way you  could have avoided it. But you must accept some bad things to be able to play with players all around the world most of the time in playable conditions.


// fats
Title: Collision Model Inquiry
Post by: fats on October 04, 2001, 05:29:00 AM
Oh and someone should make a FAQ about collisions and their modelling. These queries pop up every now and then in various games that model 1-sided collisions.


// fats
Title: Collision Model Inquiry
Post by: hitech on October 04, 2001, 07:28:00 AM
Zygote did he die, or at least take damage, with what you discribe both planes should have had damage.

HiTEch
Title: Collision Model Inquiry
Post by: Creamo on October 04, 2001, 07:35:00 AM
Morning HT.

 I'll bribe ya to take Zygoat face value and release 1.80.

Thats a bet.
Title: Collision Model Inquiry
Post by: Apar1 on October 04, 2001, 08:51:00 AM
Thx again fats, S!

P.S. U flying under the handle fats in AH?
Title: Collision Model Inquiry
Post by: SKurj on October 04, 2001, 09:08:00 AM
He did go down HT just after Zyg did.  I was in the arena at the time.


SKurj
Title: Collision Model Inquiry
Post by: Maverick on October 04, 2001, 01:22:00 PM
First off.

Collisions are not a big deal to me inspite of what anyone inferred from my previous post. I was posting my opinion on the topic.

My opinion is based on, GASP, reality. In real life when 2 aircraft collide they BOTH get damage. Doesn't matter who was at fault, was looking in the right direction or gaming the game. They both suffer consequences of the collision. That is the premise for my posts on this situation. If one FE "sees" a collision and damages or kills a plane both participants should suffer. It either happened or it didn't.

Dropping kills in a collision or awarding the kill to both is "fair". In any event there were cases of ramming of aircraft in combat. Quite a few were actually intentional, particularly over Japan.

Mav
Title: Collision Model Inquiry
Post by: Regurge on October 04, 2001, 02:37:00 PM
A while back I was dogfighting a 109 1 on 1, neither of us had taken damage. As I watch him pass by on a perpendicular course about 200yds off my 6, he sheds a wing and goes down. I asked why and he said collision.

Turns out he was in Sweden, and with my 56k connect we had some serious lag between the two of us. So while he saw a collision on his FE, I never saw him closer than 200yds on mine. If both parties always die in collisions, my plane would have just fell apart for no apparent (to me) reason.

The way it is now, if you fly into an ememy plane on your FE you die. Avoid him and you live. The same rule applies to the other guy too.
Title: Collision Model Inquiry
Post by: SKurj on October 04, 2001, 02:43:00 PM
and it has nothin to do with which way you are looking.


SKurj
Title: Collision Model Inquiry
Post by: Eaglecz on October 04, 2001, 03:23:00 PM
I had few HO when we both didnt fire and 1 of us die and the second one had dmg radiator or engine .. I saw this situation few times. Im sure that he didnt fire, because we spoke about it later.
Title: Collision Model Inquiry
Post by: fats on October 04, 2001, 03:32:00 PM
Apar1:

No I don't currently play AH - or any other on-line flightsim for that matter. There aren't any that I know of around for Linux. I used to play though under fats, AW on Amiga, then AW on PC in DOS, CK beta, WB and then AH.


// fats - holding his breath until a Linux port is made
Title: Collision Model Inquiry
Post by: DanielMcIntyre on October 04, 2001, 04:41:00 PM
You sure Skurj?  Okay, I got the impression he was still flying.  Maybe he died.  Had a similiar collision today we both died as well.

Of course I did'nt get the kill thou  :) but who gives a flying kangaroo ehh.


Never mind HT, all sorted out!!! 1.08 please


  :D
Title: Collision Model Inquiry
Post by: AKSWulfe on October 04, 2001, 04:47:00 PM
Quote
Originally posted by Maverick:
My opinion is based on, GASP, reality. In real life when 2 aircraft collide they BOTH get damage.

Real life/reality runs in real time. Virtual battlefields, such as AH, can run up to 1 second behind.

For example: My ping is 190ms to HTC. Someone from Russia with a really shotty ping is at 300ms. That's 490ms between us, the latency of a 14.4k modem.

Now, if this ran in real time and both parties saw the same thing, making game netcode would be a lot easier. We wouldn't need smoothing code either...
-SW

[ 10-04-2001: Message edited by: SWulfe ]
Title: Re: Collision Model Inquiry
Post by: Gr8pape on December 29, 2011, 04:06:21 PM
See Rule #10