Player Features and Ideas Discussion

This idea was probably suggested a few times in some modification. What i miss in eve is a little more individualism in game, mainly in large battles, it doesent really matter what equip you take, you just hope you won't get targeted and alphaed.

How about some targeting system disruption or something like that. If there are more people than X targeting 1 ship, signals are disrupting each other, and every next attacker will have it's targeting disrupted by xy% thus taking targeting longer, or targeting is less accurate and more missing is occuring or something simmillar.

For example 5 ships attacking one produces 5x 100 dmg, but whne 6th ship enages, targeting gets a little disrupted, so 5% more missing occures on all ships, so together they will get 6x 95 dmg = 570 total dmg. Thus making additional ship a little less effective. 7th will for example rise missing to 7% generating total 7x 93 dmg = 651 total etc. XYnth attacker will bring so little dmg addition, that it won't be worth it and will target someone else instead. Exact numbers are offcourse for extensive calculation and testing (also other mechanics like blocking targeting by friendlies, attacking capitals etc)

This will open space for new tactics, skills, will bring up little more individualism (as you won't be fighting vs 100 at once) and will make active tanking maybe more wiable if the numbers are right.

Then everyone in the fleet targets five friendlies and all of a sudden it's even harder to beat a blob.



As for the LOS argument, that makes structure shoots literally impossible.

thats why i wrotethat calculations and testing (also friendly target blocking) are necessary. Im talking about general idea, not exact thing.

I hear this all the time: "doing X would be too hard, we shouldn't even try." Line of sight is a very simple variant of ray tracing. If the positions of all ships are fed into a specialist process(or), the "X shoots Y hits Z" answers might all be ready by the time the full set of "X shoots Y" commands are processed. I'm sure nVidia has some raytracing or collision detection functions on their GPUs.

A simple raytracing approach could be to give every ship on grid a different "colour", then use a ray tracer and trace all the "rays" of turret fire to see what "colour" that weapon turns out to be. That colour indicates which target the weapon actually hit. So you prepare the "X shoots Y hits Z" table (it's a table approximately N^2 in size, containing elements which are big enough to enumerate every shootable object on grid), then as the "X shoots Y" instructions come through you feed out the "X hits Z" to the ship proxy processes. Or the ship proxies can query the table directly.

A more complex solution would include determining where on the target the directed weapon actually impacted, with this information being sent back to the client to direct the rendering of weapons fire. No longer will you have every shot impacting a single vertex on the hull.

And on top of this all we have time dilation which means the simulation can afford to spend a little more time doing the ray tracing/collision detection/whatever, and that process might even be disabled for very large fights.

Edit: note that raytracing is often optimised to trace thousands of rays emitted from a specific 2D plane, being the viewing plane or "window", so I don't know whether or not resetting the workspace to trace arbitrary rays from many points in the space will benefit from raytracing. I think the technology exists to do Line of Sight calculations, and thus I don't believe that "it's too hard" is a good excuse for not trying. Once CCP can determine, "Line of Sight using the best possible technology adds 100 cycles to a process that is 1000 cycles long, therefore we will not implement it at this time," then we'll have a technical rejection for the technical solution to the technical problem ;)

Of course if it turns out that "Line of Sight using the best possible technology adds 20 cycles to a process that is 1000 cycles long" then it's worth discussing the inclusion of LoS with the option of disabling it when time dilation reaches its technical limits.

I'm on the fence about LOS. Being a small fleet pvper it would be nice because enemy blobs might not be as efficient at damage projection ie they hit each other. But if you think about it a little even 1k is a huge distance for a 425mm round so if we added it would it really make any difference other then the occasional mishit.

Edit:
_____________
Using this logic you can say we already have LOS the ships computer calculates a trajectory that will bypass all objects by lets say 2m to hit its target ;).


Furthermore I feel like it will just be exploited against small fleets (3-5 people) vs the masses. It already takes some skill to maneuver around blobs and pick off the retards letting them hide behind each other at 0 (hitting approach) seems like it will be a skillless trick that will thwart us.

What's to be explained? You target the structure, you shoot it. The 30km diameter bubble on a POS means that it's super easy to hit. Even an Avatar isn't fat enough to block your view unless you're parked inside its hangar bay.

Of course there's also the option of taking away the bubble and giving POSes real hangars. You know, like CCP was thinking about doing with the upcoming POS revamp

TLDR;

The problem is hard not on technical but on the computational level and it's not going to change unless you have power to bend laws of mathemathics.

Yes, the raytracing-like can be optimized but you forgot that you have to perform it for every ship taking part in the battle. And it all have to be performed on the server (I could hack my client so it always sends the message I hit my target). You yourself observed it takes N squared calculations and they have to be carried every time a shot is made (let's optimize not to do it every second).

Can you sort it out by throwing more hardware on it? Sure! But it scales badly. Even if you manage to get, say, 1000 people fight going on (1 milion calculations every time), 3000 people fight will bring the server to its knees (10 milion) and last time I checked, CCP would like to have battles scale way beyond that.