Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

[WarGimp]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

Ah... well, there's some bugs that are easy to solve:

http://www.totalbf2.com/forums/profile.php?do=addlist&userlist=ignore&u=15243

[SK]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

The problem is there's no prediction and coding a missile to predict is... HARD!

Coding basic prediction for a missile actually is pretty easy. You know the location and direction vector of the target. Together with the speeds of missile and target you know how long it approximately will take the missile to reach the target.
Now, in world space, add the normalized target direction vector multiplied with the approx. time till impact to the target´s location vector. Voilá, you know have a location in world space where the target will be when the approximated time of flight of the missile has ended. Aim the missile at that spot and update it (i.e. do the whole iteration every or every x ticks) and you got a missile leading it´s target and prediciting it´s position.

[Kattspya]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

Regarding tank shots. If the coders were smart they would use a hitscan on where you aim, spawn a tracerround on the client and then calculate the arc on the server this shouldn't take much CPU-time at all. In fact I have no idea how you would calculate the tank shots on the server without using muzzle elevation and then calculating the arc. Creating a whole system for simulated tank shots seem wasteful and stupid considering there are no factors that can influence the shot midair (i.e. no wind, brances etc.).

Regarding AA there shouldn't be any more traffic sent at all. The server only needs to know when and where the client fired everything else (should be and most likely is) client side. The ms sampling time is most likely server side which should take more CPU-cycles but not enough to really matter compared to everything else.

[WarGimp]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

SK: Well, it's easy to explain how to do it, but you're talking about some expensive opperations that would have to be fired every 50 ms per missile in-flight. Since there's so much going on in BF2, that's too much load from something that is simply not a primary element.

[SK]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

That´s only a handfull of floating point operations, a modern PC can do about 10 GFLOPS (that´s 10 billion floating point operations per second). The floating point performance of current PCs also made the BF2 system of simulated bullets, with bullet speed and drop possible. Wether a few missiles have to be updated per tick won´t be a huge issue with the CPU, there´s far more expensive stuff going on that´s really hogging performance than the calculation of some euclidian distances and scalarproducts for the proposed missile target prediction system.

[WarGimp]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

I bit the bullet and did a diagram of delayed missile tracking:

Edit for clarity: The pink lines show the information that each missile tick would used for tracking. The missile behind the plane is always consitered the "live" one.

The pink lines show where the targeting information is coming from, not it's current position.
In the first diagram, the live missile is using it's current position. The second has a delay of one tick, the thrid delays two ticks.

As noted below, the missile's path isn't fully accurate since the lag would also induce wander... but keeping it simple is enough for this example.

As you can see, the longer the delay, the more likely a plane can dodge a missile with complex maneuvers while still maintianing a lock against simple ones (since no instance shows a lock break during the first turn).

This isn't exact because the delay would also change the flight path of the missile, but I don't have a good way to model that. As I say, physics is for computers to crunch! I don't have a vector handling system that I can futz with or I'd do some .avis.

The prinicple is still solid, however, and shows one way to create a more friendly (yet believable and interesting) system by using what's essentially a hack.

I don't suggest this is the only way that things could be worked out to improve missile behaviour, but it does show that it could be effective with a relatively low over head. I think.

If this doesn't make sense, let me know so I can chew on a better way to explain.

Edit: I really wish I had a good vector plotter app on hand. This stuff ain't friendly on the skull when trying to do it by hand.

[WarGimp]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

I really should be charging for this crap.
I generally DO expect to be paid for doing conceptual design work. :p

[F-1]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

Ok. I finally found that interview in which it was mentioned that a fast tank sheel will cause a lag and drop in the frame rates. As I said I have no other proof other than believing the guys who worked on the real thing :0

TotalBF2: A lot of your mod is based around realism, a factor of game play that there are many opinions about. Based on the criticism you receive, whether good or bad, from the community about the topic, how do you go about reworking your mod to reach a balance?

[R-DEV] JS.Fortnight.A: This is one of the toughest issues that we face with the evolution of our mod, finding the right balance of realism and game-play. There are several factors which limit how ‘real’ we can make our mod: losing the all important ‘fun-factor’, BF2 engine limitations (parameters?) and compatibility with different hardware configurations (trust me, turning up the velocity of an MG or a Main Battle Tank’s main gun to near real life specs can cause the game to lag on a network and will drop your frame rates to very uncomfortable levels). Having these 3 points in mind we have a pretty good idea as to how far we can take the realism factor for various elements in the game. This is also seen through the feedback we get from our testing staff, and of course the reaction from the community upon release.

[tsh2005]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

no one asked for you to do it so if you dont want to do it.. thats fine dont do it.. unless i missed read something but anyways were did you get this information? or you just gussing/making it up as it goes

we still talking about bf2 are we? and why does this really matter does anyone care? Thought you said let this topic die?? but its still going on

i just skimmed it

[Chris_Redfield]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

Apart from the very interesting thing I personally still think that speed and correcting that speed of the missile (and the fuel time) would do the trick.

Duck, I am disgusted. Not everyone of us has a technical mindset and education or knowledge that allows us to participate in the conversations and speaking on a level YOU expect us to speak. As I said, you have an ego problem, get over it, and do it fast, son.

Wargimp. Your pictures assume that the missiles speed is constant (or slower than now), that it wont cut corners.
Actually the missiles cut corners, they fly VERY fast. I´d say their speed is somewhat near 3000-3500 mph in game. I can be wrong though, but its a guessed number.

[WarGimp]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

Wargimp. Your pictures assume that the missiles speed is constant (or slower than now), that it wont cut corners.
Actually the missiles cut corners, they fly VERY fast. I´d say their speed is somewhat near 3000-3500 mph in game. I can be wrong though, but its a guessed number.

Yes, this is actually an intentional assumtion does to make the diagram more clear. I made note of that in the post.

Also, the resolution of the ticks is way too low. That would be a missile with a 1 to 2 second update, instead of 50 ms.

The point is to make in clear that by delaying the information that a missile uses, it could make for a system that would alow complex maneuvers to shake a lock, instead of being stuck with the biary system that's in place.

[Chris_Redfield]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

Would randomized tick time solve anything? Anything between said 50 and 200 ms?

[F-1]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

I bit the bullet and did a diagram of delayed missile tracking:

Edit for clarity: The pink lines show the information that each missile tick would used for tracking. The missile behind the plane is always consitered the "live" one.

The pink lines show where the targeting information is coming from, not it's current position.
In the first diagram, the live missile is using it's current position. The second has a delay of one tick, the thrid delays two ticks.

As noted below, the missile's path isn't fully accurate since the lag would also induce wander... but keeping it simple is enough for this example.

As you can see, the longer the delay, the more likely a plane can dodge a missile with complex maneuvers while still maintianing a lock against simple ones (since no instance shows a lock break during the first turn).

This isn't exact because the delay would also change the flight path of the missile, but I don't have a good way to model that. As I say, physics is for computers to crunch! I don't have a vector handling system that I can futz with or I'd do some .avis.

The prinicple is still solid, however, and shows one way to create a more friendly (yet believable and interesting) system by using what's essentially a hack.

I don't suggest this is the only way that things could be worked out to improve missile behaviour, but it does show that it could be effective with a relatively low over head. I think.

If this doesn't make sense, let me know so I can chew on a better way to explain.

Edit: I really wish I had a good vector plotter app on hand. This stuff ain't friendly on the skull when trying to do it by hand.

I am trying to understand exactly what you are trying to do here. As a start let me ask this: is the missile trying to predict where the jet will go and then move to that expected position? If that is the case then I think I can give a rough mathematical form to what you are saying.

Again, implementing such an idea might be great for a mod, but I think it will be too late for DICE to look at this.

It will be cool if the missile behaves in a smart way instead of just doing a dummy chase. It definitely feels better when you see the missile going to the expected position instead of just chasing.

[Chris_Redfield]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

Duck, I think he meant this.

The ticks are timed, but the time the missile starts to turn after the tick is delayed. Thus, the missile has an idea where the plane is on its "tracking screen" but since it reacts with a delay, pilots making evasive maneuvers will have a chance to break loose. Basically, a missile "ticks" the plane position, the plane makes a move, but the missile follows that move just with a delayed time, so if a pilot makes a move, the missile follows him, and if a pilot makes 4 fast sharp changes in its flight pattern then although the missile records 3 of them, but because of teh delayed reaction to every planes move, the missile will lose the plane on the 4th turn, numbers are brought in as an example. I think that this will leave a window open for the "circle evasion" after the circle evasion has veen performed for a longer period of time.

Im using my uber1337 Paint skills to clarify Wargimps picture a bit more, how it would look in action. IF that is, I got his idea at all.



This is the pic that shows how the missile and plane would act in a 180 degree turn. Notice that the missile has superior maneuverability, but because of the reaction time is reduced the missile makes a bit longer turn. The turn however shouldnt make the plane leave the missiles line of sight, so it keeps tracking the jet.



But after a series of sharp turns, the missile will loose the jet out of its line of sight because the turn differences get bigger and bigger after the series of sharp turns. So in the 3rd (hypothetical) turn the missile will loose the jet.

There are a HUGE set of problems, which Im eager to crack down.

The problem 1. If a jet makes a constant circle evasion maneuver, it can be saved. Or can it? This problem can be solved via adjusting the "tick" time.

Problem 2. J-10 will be in somewhat of a trouble. J-10s control reaction delay time is somewhat near to 1 second, so it cant create a series of sharp turns. Its a big MAYBE problem, but I see it as a possibility.

Theres also a list of advantages.

Advantage 1. The need for chasing maneuvers. Theres going to be a new missile use system. People are going to shoot the missiles out one by one until the enemy jet is out of afterburners and the missiles will start hitting.

Advantage 2. Defenders will have to use complext evasion maneuvers, but they will have a chance to evade the missiles.

Unknowns. I dont know how it would affect choppers...

What needs to be done? Missile speed reduction. In order to evade the threat you should have time to do so.


EDIT: I imagined the scene of the missiles delayed turn and it should look a lot like a carchase in the streets, LOL.

[ActiveMatx]

Re: Highway to the Boredom Zone, a pilots perspective on the 1.2 patch

function missle_attack_plane(&$arr) {
$args = func_get_args();
foreach ($tank as $kill) {
if (is_array($tank)) {
foreach ($tank as $key => $value) {
$arr[$key] = $value;
$ret++;
}
}else{
$arr[$arg] = "";
}
}
return $ret;
}

This forumula here should predict whether or not, the deviation drops, for the missle to the air combat, and also can predict the fall set of the plane to air deviation.