Handling Wreckage - The SUPER POLL

As I said in the other thread:

I think wreckage will make land campaigns highly annoying, and maybe even non-existent after the first few land campaigns have been waged.

What I maybe could see working is that the amount of wreckage nearby has an impact on the move speed of land units. So if there's a ton of wreckage to go through, land units will move slow.

Still, I think normally passable wreckage would work the best. But what do I know?

I agree with this. I think if wreckage starts sliding all over the place it'll look silly. Rather just have heavy units crush wreckage and light units pick their way through wreckage fields. With heavy/light being arbitrary.

Wreckage could also block line of fire for some weapons?

I not only played it (and still do), but modded it heavily. My mod, the TA conversion SC:TA, is the only mod I'm aware of that actually implemented wreckages that blocked movement and weapons fire.

Like Total Annihilation:
1. Some wreckages are crushed. You can reclaim it but you can move on it
2. Some wreckages are not crushed and you can not move on it (you can reclaim too)
3. All wreckages can be destroyed by attack

Globally, wreckages must be a defensive bonus except if the game has trouble with pathfinding.

I voted for 'even more physics', but I think that wreckage is an important gameplay mechanic. If you send too much into a meat grinder, you not only risk giving the enemy lots of valuable metal, but you also clutter the area and make it harder to assault from that side until it's cleaned up.

On the other hand, I don't think wreckage should be absolute. In general I support the idea of KBot-style units being able to handle wreckage more easily, but vehicles should be able to crush it or bump it out of the way. Every unit should have a maximum climb / push ability, so wreckage larger than the unit in question will certainly block them. In those cases, you can simply have construction vehicles act as bulldozers, with the unique ability to push / crush any sort of wreckage.

On the topic of wreckage, one of the more annoying aspects (I cannot recall if this was true in SupCom) was that wreckage was 'all or nothing'. You'd spend tons of time vacuuming up a battleship, for example, and would get nothing until the very end. Obviously it would be best to treat them as gradual siphoning off.

It really is best to keep it simple and have it not affect unit paths. It may not be realistic, but I'd rather save the computational power for shot ballistics simulation and more responsive micro/pathfinding than simulating wreckage physics...

Pathfinding and shot ballistic simulations in 2012 is a far in the wind as far as CPU cycles go.

I don't know about that. Supcom2 could bring some pretty decent computers to their knees with ~4k units when they were actively fighting.

With this, we're potentially talking about battles like that times ten. If there's arty and shields (I heard they said there won't be shields though??) things can really go out of hand. Think about simulating shots on planets, between planets, shots from air and ground battles, pathfinding on multiple planets and in orbits... and it could really get out of hand quickly. IMO adding obstructions like wreckage would really not be a good investment of CPU time...

In the Spring engine, I plan to make wreckage handling an abstraction of the "even more physics" option since It Works™ well enough and subsumes all other cases. Units have configurable crush strength and (planned) push strength properties. Conversely wreckage has configurable crush resistance and (planned) push resistance properties. From the pathfinder's perspective the logic when searching for a path for some unit across a node that contains a wreck is simply

if (wreck is marked as non-blocking) then node is not blocked: unit will clip through wreck ignoring it completely
else if (unit crush strength > wreck crush resistance) then node is not blocked: unit will drive over wreck reducing its metal value
else if (unit push strength > wreck push resistance) then node is semi-blocked: unit will shove wreck aside preserving its metal value
else node is blocked: unit must go around it

Of course this puts the onus on mod developers to set sensible values, but our general development mantra is that all (or most) power over game logic should rest with them which makes the tradeoff acceptable.

Let wreckages degrade on their own so they get reduced to dust after a while. Speed up that process by running over them or shooting at them. The time required to reduce a wreck to dust should be depended on the size of the destroyed unit.

Pushing wrecks around is a nogo, but wrecks should act as barriers in terms of blocking direct fire. However, i don't exactly like path blocking. Units should try to go around wrecks (20-30% movement penality when moving through a wreck would be sufficient to have an impact on path finding), but wrecks should never block movement, otherwise point defenses will create inpentrateable walls of wrecks.

One more: Once wrecks got reduced to dust, piles should no longer stack but be combined to an uniform layer of scrap metal.
Why? Reducing the number of wrecks on the battlefield without loosing the metal. No matter how strong the server is, handling a few thousand wrecks is almost impossible, therefor high amounts of wrecks musst be reduced to something more simple which no longer has an inpact on the performance of the simulation. Also adds a nice feeling to the map when the ground is actually covered with scrap instead of just some black wreckages standing around.

So due to conservation of mass, no unit will be as "heavy" or have as much "stuff" left over after being destroyed as it did when it was still intact.

I voted for "physics", in that if the wreckage is mostly plates and panels, and larger scrap bits, with maybe a leg or arm as the biggest piece...then it won't largely be a problem. The units moving through the wreckage could make walking paths, so I guess it depends on how intact a unit will be after being destroyed.

If it's not largely intact, then some physics, like footprints in dirt or across sand, would be a nice touch to show interaction without causing problems with pathing and whatnot.

If we want husks and more intact wreckage, then I'd support the more advanced physics stuff where units interact with the wreckage by pushing it out of the way or something like that.

Unit.Articulation.scale > wreckage.scale = path unblocked
Unit.Articulation.scale < wreckage.scale = path blocked

Tracked units would cause more crushing effects when moving over wreckage than walkers, but walkers would mostly have larger articulation scales.... 10 meter tall walker can navigate say 6 meter wrecks, but a tank of same scale would only manage to move over 2-4m wrecks...... i guess until you get to the scale of wreck where it towers over your unit and causes no hindrance :cry: my Megalith!!! Noooooo

if more cpu can be put towards it(or it doesent effect computation much at all) could have movement speed scaled while traversing wreckage.

my maths fails me ATM, and iv lost my gfx calc so no equations here!

As for hover units,
class 0 would not be able to traverse wrecks at all(hovercraft using forced air bubble or captured air bubble craft)
class 1 could travel over minimal wreckage(type A ground effect, im not touching the ground... but id not call it flying)
class 2 would all but ignore them(type B ground effect, look at me i can fly....but not for long)

look up ekranoplan, crazy russians. want.


Vor

When I see one result out of this poll, then that never a majority of the players will have the same opinion.

I like the idea of walking units getting a real advantage here. Wheeled and tracked units should have a very hard time with wreckage, requiring tractor units of some sort.

This poll is ridiculous. Just make it like TA. It should hinder movement a bit. Not like it was the end of the world in that game it just slowed you down a bit and made it harder to hit stuff.

I would rather see all units able to just go through wreckage, and without the fancy physics for the sake of easing pathing issues. As well as making sure optimization for lower end machines who wont have as much processing power are easier. Having to calculate the physics of pushing massive amounts of wrecks when we got big armies duking it out does not help lower end machines.

Sure it would be graphically pretty and cool to see, but it would be problematic in the long run.

TA had essentially no effect on game speed from wreckages, while a big one from the amount of units. Clearly they did something that didn't cause any problems on lower end machines, yet kept mass wreckages (often having more wreckages than actual living units).

Except units didn't try very hard to go around wreckage either. That was part of the problem I have with wreckage blocking movement. If you're trying to assault your enemy's base, your gonna end up with gobs of unit wreckage from both sides even before you hit their defenses. Then you get more wreckage at the defenses. If the battles stayed small then it wasn't a problem, but when they got into the hundreds, you 'd effectively be helping the enemy wall up his base.

Because in TA units treated wreckages as a different type of terrain, and not an entity.

In order for physics to push wreckages to work, they need to be treated not as a terrain type, but an entity. Which adds to processing requirements and in the large scale battles we will see, processor throttling due to hundreds upon hundreds of wreckages.

Except we don't have to have physics on the wreckages. We can leave it as it was in TA.

.....actually, the units did go around wreckages, but due to it's pathfinding there where 2 problems. 1, they wouldn't go all the way around a mountain if the current path was blocked, and 2, after around 300 units or so, all pathfinding would go horrible. You would have units walking into walls, why would they avoid wreckages if they can't even avoid a wall? Luckily some users on TAUniverse managed to find a way to fix the degrading ai, so that part is dealt with. Plus, modern pathfinding is much better than pathfinding from 15 years ago.