*DAMN R6 Forum

My theory is, that in Ghost Recon (and most other 3D Shooters), the player's accuracy is determined by a certain algorithm. This algorithm calculates into which circle the shot will hit with a certain probability (see the graphs). You have to imagine this like an ice cone. Your gun is at the narrow end, and the far end's width is determined by various factors, like the running speed, if the player is wounded, the type of the gun you have, and others.

Secondly, there is a random element in the "chance to hit" - a factor that ensures that a shot can also hit outside the shooting cone. This explains the fact that sometimes, even if you are standing still, your shots hit areas which aren't within your crosshairs.

Please note that this is only a theory, and I can only suspect that GhR works this way.

*DAMN Elandrion

(http://dynamic.gamespy.com/~damnr6/elandrion/graph.jpg)

My theory on accuracy is that most players are pretty slow thus I hit them faster....making me more accurate!

Go Chiefs!, Beat Denver!

Heres a question i have wondered on for awhile, and have yet to get a solid answer on..

In GhR, does every bullet have a ballistic path programmed so the computer at all times knows where each bullet is

or...like elandrion said...

Is just so that if you have your reticule on a guy, the computer will determine if you "hit" and play some graphical effects like blood.

Does this make any sense?  :)

     First, let's start with a premise: that full ballistic simulation would lead to a much more realistic and accurately-simulated game. This is arguably true.

     However, no game that I know of (save possibly Combat Mission, a turn-based war simulation) is made this way. Why is that? Surprisingly, the answer is not "Red Storm sucks". It's because of the burden on already-scarce resources that full ballistic simulation would impose. To simulate a single bullet from firing to impact is not particularly difficult, mathematically. However, it does require a fair bit of processing time to calculate the path. If the game simulates wind and other atmospheric variables (which, as you know, GhR doesn't), the simulated bullet becomes that much more realistic in that its flight is not guaranteed to be a perfect ballistic arc. But the cost of that extra simulation is more processing time. Still, not a big deal for one bullet. Modern processors could handle the load easily. The problem comes when you consider full ballistic simulation for many bullets at once (suppression fire with an MG3, anyone?). Now you're talking about a significant processing load, and that comes at the expense of the normal processing the CPU is doing for the game. Now, consider the added problem of full ballistic simulation in a multiplayer game. Suddenly, each client isn't merely calculating the trajectories of its own bullets, but is also watching to ensure that nothing gets in the way of the trajectory that wasn't there when the bullet was fired, which means prompt network traffic becomes a real problem.

     The way FPS games handle hit detection is generally using a method similar to "roll some dice to determine what direction the round is going to take out of the barrel; now check to see if there's anything on that line. There is? OK, the round hit that." There isn't any "flight time". All rounds hit instantaneously. Naturally, very few games have engagement distances which are sufficiently large for people to notice that bullets go too fast, so this shortcut works well in most cases.

beautifully said, Loth  8)

*DAMN Elandrion

Are you sure about that, Loth.  Cause on maps like red square I know i've seen muzzle flashes a good second before the bullet actually hit me.

Must have been lag Splinter

Maybe lag, maybe not... I seem to remember RS haveing a "Muzzle Velocity" tag in its weapon data files. Time to target is one thing that wouldn't be too hard to calculate, so I wouldn't be surprised if it was worked in.

     I know that in Ghost Recon (and so probably in RS as well) the muzzle velocity figure is used to calculate ballistic energy for purposes of damage calculation. It's not used at all to actually control speed of the round. I actually tested that once. I was toying around with an idea for a paintball mod, and as part of the research, I made some gun or other's rounds have a muzzle velocity of something like 1 ft/sec (m/sec, maybe?). Now, that should not only have made the rounds take a very long time to hit their target, but if there'd been real simulation, they should have hit the ground not more than a dozen feet from the barrel. Neither effect happened. The rounds hit just as promptly as always.

     Any delay between firing and impact is purely attributable to network lag. Unfortunately.

     Something just occurred to me: the muzzle velocity of a 5.56mm round is something on the order of 900-1000 m/s. The M-16 has a muzzle velocity of 945 m/s, so using that as an example, a shot fired from one side of Red Square along the east-west axis to the other side (assuming a clear line of fire, of course) would reach the other side in less than a second.

     So, Splinter, since the limit of visibility in that map is considerably less than from one side to the other, it's safe to say that if the bullets were being accurately simulated, you wouldn't have seen a second's delay between firing and impact. Thus, it's just good old lag.