Pht, on 08 September 2013 - 10:04 AM, said:
Doesn't apply to battlemechs. We know that the are capable of aiming every weapon on them, independently of the mobility of the section they are mounted in.
And they evidently weren't very good (or even moderately good, for that matter) at doing so, considering how much of an impact the MechWarrior's own gunnery skill (or lack thereof) impacted the likelihood of whether one actually hit even a stationary target (which, incidentally, also demonstrates that the BattleMech is, by itself, largely incapable of aiming its weaponry without direct input from the MechWarrior).
Quote
The Jenner is firing four medium lasers at the Atlas. The target is four hexes away, which is in the medium range for the lasers, adding a Range Modifier of +2. The Jenner used Walking movement this turn, so the Attacker Movement Modifier is +1. The target did not move. The Base To-Hit Number is 4, so the Modified To-Hit Number is 7 (Base 4 + Range 2 + Attacker Movement 1 = 7).
(Total Warfare, pg. 108)
As a slight modification of the situation given in
Total Warfare: a stationary (no movement modifier) IS 3025-era (no special equipment modifiers) BattleMech with an average MechWarrior (Gunnery 4) using a Medium Laser at Medium Range (+2 modifier) attacking a stationary IS 3025-era target (no movement or special equipment modifiers) on open, level terrain (no terrain modifier) with
Standard Day conditions (no environmental modifier) needs a
2D6 result of 6 or better to hit a target.
In other words, a regular/average IS MechWarrior in
a stationary BattleMech has only a 72.3% likelihood of hitting a stationary BattleMech-sized (8-14 meters tall, ~4-7 meters wide) target with
a light-speed weapon at a range of 180 meters (~590 feet, or ~197 yards).
A slightly better IS Mechwarrior or regular/average Clan MechWarrior (Gunnery 3) under the same conditions would have an 83.4% likelihood of succeeding on the same shot, while a slightly below average IS MechWarrior or very low-grade (relatively speaking) Clan MechWarrior (Gunnery 5) would have a 58.4% likelihood of succeeding on the same shot.
And even then, identical weapons that are mounted physically next to each other (e.g. the paired Medium Lasers in each of the
Jenner's arms) stand a high chance of not striking the same general area of a target... assuming either strikes the target at all.
As such, either MechWarriors in general literally cannot aim to save their lives, or the BattleMechs' aiming adjustment systems are so sub-par that they'd arguably be better off with properly-harmonized fixed mounts anyway.
All that aside, the concept of gun harmonization can be made to apply to BattleMechs, which would have the dual effect of solving the "'high pinpoint alpha for all ranges' problem" once and for all
and providing a means to make the eventual implementation of the Targeting Computers relevant.
Pht, on 08 September 2013 - 10:04 AM, said:
Beyond which, the only way I can think of doing this is to do a realtime raytrace/cast from each weapons port; realtime traces/casts are prohibitvely expensive in computing power.
Doing it that way is no more "prohibitively expensive" than determining that any given direct-fire salvo hits the target (including the lasers, to determine where each "tick" of the beam lands).
Even if one had two full companies (for a total of 24 individual 'Mechs) of
Black Hawks running around firing all of their lasers simultaneously (12 lasers per
Black Hawk) and a ray was drawn for each takc of each laser beam (10 ticks per beam, IIRC), that's
a total of only 2880 rays for all 24 'Mechs that would have to be drawn to see where each "tick" lands.
Since there will be a need to see where and when each "tick" impacts some other object (be it 'Mech, terrain, building, etc) in the game world anyway, the computation is on some level already being executed.
Moreover,
research from Cornell University indicates that (as of 1997) "the native ray casting method is
O(N)" (where N is the number of objects or primitives in a scene) while "in the computational geometry literature, there is a method that has been proven to be O(log(N)) by DeBerg, but unfortunately its extreme storage requirements make it impractical."
Quote
O(N) describes an algorithm whose performance will grow linearly and in direct proportion to the size of the input data set. The example below also demonstrates how Big O favours the worst-case performance scenario; a matching string could be found during any iteration of the for loop and the function would return early, but Big O notation will always assume the upper limit where the algorithm will perform the maximum number of iterations.
(
source)
And, again, since the server has to compute the hit detection for each and every "tick" anyway, the what cost there is would already be paid.
So, please explain (in detail, and with supporting sources!) how employing fixed, harmonized gun mounts (
which are employed in games like Warthunder, among others) is "too computationally expensive" for MWO.
Quote
In Warthunder just as real planes are setup there is a convergence effect of weapons installed on the aircraft. This is set via the menu options prespawn viathe dropdown selection. All the guns will put the center of their bullet patterns at a point in space a set distance in front of the gunsight. Note that this is a static setting, and only "works" when sitting on the ground, with the plane held level. The general effect is that guns will pass through the gunsight center at close to the specified range if you fly straight/level. Dispersion is modeled in Warthunder however, and this makes guns fire in a "shotgun" pattern in cone pattern from the muzzle. Rifling was an important invention for gun accuracy when it was introduced, but it isn't perfect. In addition, gun recoil, especially on a flexible structure like an airplane wing, can cause subsequent rounds to vary in velocity direction. This can cause significant spreading of ammunition at further distances.
1
