Children of a Dead Earth: realistic space warfare game

[Spartwo]

I'm very glad someone has finally done this.

[Pixel of Life]

Interesting. The graphics are sort of meh at the moment, hopefully he'll improve them at some point. Does anyone know what game engine this is running on?

[STS Fallacy]

Definitely buying that one...

 

[tater]

The accuracy with respect to damage isgoing to depend a great deal on the damage model. Many games screw up at that point, and it's important to remember that in terms of "realistic" outcomes (the desired simulation should be to have realistic outcomes, right?) the game is only as realistic as the least realistic bit. In may cases a game would do better to have a board-game style damage table than "simulation" if the simulation is done wrong, honestly (think of a old board/tabletop/paper game like Harpoon as an example).

What's the mass of a KE weapon (missile) in this game? What do closing velocities possibly look like? Several km/s? More?

You can argue that the hull is a tin-can, and the missile, call it 100kg (incredibly light, IMO) hits at 5 km/s. There was some talk about it going on one side and out the other. I don;t disagree, but it's considerably more complex than that. The volume it passes through is not empty, it is filled with propellant. This means that there will be at the very least a shockwave moving through the tank, which means some of the energy is deposited along the way. Any sensible projectile in this situation might detonate some distance from the target, and pepper the craft with many, smaller fragments at the same velocity. It might not wreck the target outright, but I see almost any hit as a mission-kill on the target unless the propellant tanks are very compartmentalized (which is a mass-expense).

He mentions spaced armorm, which is light, but effective vs the sorts of threats that ISS faces... but the weapons hitting these warships are not flecks of paint at 5 km/s, they might be more massive penetrators at those kids of velocities. The spaced armor in this case means that more energy, not less, is delivered to the target. If you stop a round, that energy is left in the target, and this means heat. I'm honestly unsure what would happen... it's a non-trivial aspect of the game, and in fact much harder to get at than the orbital mechanics, etc.

Edited May 8, 2016 by tater

[weissel]

On 5/3/2016 at 2:51 AM, Accelerando said:

Sloped armor is still an assumption as you admit, and that is important, since the structure of a spacecraft's armor (or lack thereof) is certainly one of the defining features of a warship. Yes, this means that they have a vision of how things should be. Just because you agree or don't agree with it, or that it makes/doesn't make sense to you, does not mean it is not an assumption.

I ... disagree.  A "vision of how things should be" is starting with the 'this is how things should be' and fixing the universe to allow that, if necessary.  The WWII plane and sub  games had a vision of sorta-realistic simulators with many/lots tweaks towards playability.  They know the end goal fairly well from the start.
CoaDE is more of an experiment, or like maths: we start with a (fairly) small number of axioms (which at least seem to make sense), and see where they lead us.

 I also challenge you to show a spacecraft shape that has equal or better properties in interior size, mass of armor and thickness of armor at any given spot (say, frontal, 45°, 90°, 135°, 180°).

On 5/3/2016 at 2:51 AM, Accelerando said:

And CoaDE makes many other assumptions, such as:

1. Available technology (no, you cannot use photon torpedoes or shields)
2. Specific impulse - MPD thrusters and NTRs are the upper limit, limiting your Isp to some tens of thousands of m/s at most, and the game assumes a situation where it is economical to manufacture these en masse.
3. The theater of space combat. CoaDE's space combat model assumes that space combat occurs entirely between contesting fleets meeting around specific intercept points. No mention has been made of ground-based installations such as laser batteries, for instance (at least, not yet).
4. Weapons loadout. CoaDE's developer(s) have stated that lasers, kinetics, and missiles will all be useful to ships in different situations, and all combat focuses on large capital ships with a fairly even distribution of armaments of each type of weapon, plus drones. This certainly makes for a balanced and interesting game mechanic, but I'm not convinced that it is the model of space combat.
5. Situational assumptions. As mentioned before CoaDE assumes that it makes economic sense to build NTR thrusters en masse, and it also assumes that large heavily armed and armored capital ships make economic sense to build over any other type. It assumes that laser power generation and point-defense targeting software will not outclass kinetics and missiles. And so on.

1. Axiom, not assumption.  If the developer wanted to explore strategy and tactics of "infantry in space" or "WWII submarines in space" or "Star Trek", he'd have done that.
2. Which engines would you like to have added?
   What is their use case?
   Are they either able to produce so high thrust that they are usable in combat (as a total replacement) or their additional mass is in the end a mass win and their transfer speeds are at least as good? 
3. That ground weapons will not be there (or even are not there already) is an assumption on your part! 
4. Weapons loadout and vessel size: your choice!  Want an all-laser small gunboat? Go for it.  
   But do not complain if you find that your lasers (or projectiles) are not ruling the battlefield.  That is not a "more balanced game" decision (this is not a game!), that is simply how the numbers turned out to be, at least for the developer's vessel designs.
   And you may find that a fleet of small gunboats is not optimal:
   1. Start by cutting out all the stuff to enable humans to live and command the ship.  Add some computers instead and a remote-control system.
   2. Reduce the remass down to keep the same dV.
   3. Reduce the size and thrust of the engine to keep the same acceleration.
   4. Reduce the size and mass of the vessel and tanks and internal structures, so all that there's little empty space again.
   5. Note that you also reduced the armor mass without thinning the armor
   6. Oh, now we probably have a lot of extra dV again, if so jump to 2. and do the whole cutting again
   7. Now, this thing is much smaller.  Maybe we can carry it in a "capital ship".
      Which means we can remove all the dV needed for travelling, pre-combat maneuvering and getting back on course and home after combat.  Let's jump to 2. again with the new, lower dV.
   8. Maybe we don't even need the drone to 'come to rest' after combat --- either we discard it or we collect it with the mothership.  Needs thinking.
   9. Maybe we don't need that amount of armor, either ... less mass again.
   10. Or maybe we use an engine with more thrust again.  Humans squish easy at 10 or 30 gee, computers can be made resilient.
       
       But it is clear we have a harder to hit target that is much cheaper, lighter, smaller (also harder to hit) and as well armored as your gunboat.  And we can have a larger number of them. There is only a 'human crew on board can try damage control' feature loss, and it'd be doubtful said crew could patch much at all during a combat.
        
5. There are several problems with your assumptions:
   • There must be quite a cargo fleet.  
     • Only when you want to genoicide your enemy and do nothing else you need no ships to carry home loot or transport occupation troops or settlers.  You likely want to mine asteroids, too, instead of sending everything from the bottom of the gravity well.
     • You do not want most cargos to be slower than Hohmann transfers.  Some (like passengers) you want to be significantly faster.  Which means engines capable of that are going to be in much need.
     • Low thrust engines will not be able to reach Hohmann speeds on most distances, no matter what their ISP is.  There is a limit on combining low thrust engines for higher thrust (they do have mass, and their support system tends to be weight intensive).
   • You assume war vessels make economic sense[1].
     • They only do if you are out on a massive plundering and extortion campaign.  Otherwise they are a cost factor.
     • History has shown many times that effective weapons (or those believed to be) are produced.  In war, money is not much of an objection.
     • Even though strategic material may be rare in a war and will be replaced with more available materials where possible, that does not stop stuff being built.
   • You assume "that large heavily armed and armored capital ships make economic sense to build over any other type".
     • I have not seen any place where it is stated you cannot build lightly armed and armor-free small vessels, if you so desire.
       I am sure you will be able to build your nearly unarmed eggshell tin cans and see how they fare.
     • It seems however that the tests done by the developer indicate that big motherships and small drones make most sense.
   • Lasers and PD outclassing kinetics?
     • Again that seems to be the results of experiments done by the developer, not a "this is how it's going to be" executive decision.
     • I am sure you will be able to run these kinds of experiments yourself.

 

[1] putting all the money spent for development, building and maintenance of war vessels and the training and keeping of their personell into something which actually provides some return of investment without a war of conquering might be more economic ... 

 

On 5/3/2016 at 2:51 AM, Accelerando said:

The developers recently outlined another assumption, which is that space battles will use nuclear-thermal rockets exclusively over MPDs. This hinges on the assumption that slow burn times will be absolutely undesirable in a space war, which goes back to the economic assumption that NTRs are mass-manufacturable in CoaDE's universe, and that they are cheaper, lower-maintenance, and safer than slow-burning engines. I'm not saying that these assumptions are wrong, but I think the developers may be overlooking other viable possibilities in their zeal to say "Our model is objective".

They have run tests and concluded that the inability to choose your battleground, match speeds or evade kinetics is a very bad trait for combat ships.  And being slower than a Hohmann transfer is also not a good trait.  Thus this is not an "assumption", but the outcome of experiments.  However you assume, it seems, they did that on an untested gut feeling or something.

And don't pull the "they are cheaper, lower-maintenance, and safer than slow-burning engines" bull.  War vessels of the wet navy have always have expensive, high power and occasionally less than safe engines, which certainly are not low maintenance, because they need the ability of high speed.  Same with the air force.  Just look at the Me 163 (where the ... 'temperamental' rocket engine and the highly corrosive and explosive fuel killed more Germans than Allies) or the Me 262 (where the jet engines loved to burn if you throttled up faster than slowly and where they had a life time of 25 hours before a total strip down (100 hours when not eschewing rare materials)).

But from what I read you surely will be able to buy these engines and see how you fare.

 

On 5/3/2016 at 2:51 AM, Accelerando said:

Broadly speaking, any experiment requires a design, which requires subjective decisions. An experiment must be pruned down to a certain chosen range of test data, which will be analyzed via specific methods chosen by the researchers. The researchers must ask a question, which defines a certain range of possible answers that can be derived from the experiment. This is not objective either; it is specifically chosen according to what the researchers consider important.

Claim: The heavier, the faster things fall.
(Thought) Experiment: 1 large cube and 1 small cube of the same solid material.  The large cube falls faster than the small cube (as per claim), so if you combine the large cube and the small cube, it will fall a) even faster, as the combined weight is higher and b) will fall slower, since the slower-falling small cube slows the large cube.
Result: The claim is false.

Which design does that experiment have, and where is the subjective decision?  Where was the experiment pruned down to a chosen range of test data?
I think your claim of "Broadly speaking ..." is, broadly speaking, quite wrong.

And the question asked here is, "given we use only that kind of technology, what could space combat look like?", and the "pruning" done here would be that technology that does not work/help --- like, say, truck wheels, aeroplane wings, jet engines, sails, cutlasses, catapults, photon torpedoes, handwavium engines, ...
 

Quote

Say we want to evaluate the viability of some space warship design; that is to say, can we build it? There are multiple ways to tackle that question, such as:

1. Is it viable in a purely physical sense - is it possible to build structures like these that function, and do so without instantly breaking down?
2. Is it viable economically, given a certain scarcity of resources and manufacturing infrastructure, a certain economic context to the space war?
3. Is it viable in an engineering sense - would this design be dangerous because it puts sensitive hardware right next to critical failure points, for instance, or would it be prone to failure for some other reason?

And so on. And then you have to decide what you're going to simulate and how granular the simulation will be for each thing. Will you need complex models, or will spherical cows suffice? None of these decisions are "objective", they are motivated by what the testers think is important.

The viability of any war vessel design would be: "Does the vessel perform it's function well enough in combat for it's TCO?", and nothing else!  (Yes, that depends on strategies and tactics employed.  Finding our which work is the point of that simulation.)

There is an objective way to decide: If increasing the complexity change the result in meaningful ways, you need to keep the more complex approach, if not, you don't.
Which is what happened to rail guns: they do not scale up and down easily, so a complex calculation is needed.

 

Quote

This isn't to say that nobody can ever come closer to describing reality, of course. Obviously, no technology would work if nobody could model reality to any degree of reliability. But the relative "objectivity" of science comes from repeated experiment and independent peer review, which is to say that "objectivity" derives from the synthesis of many different subjective analyses.

So basically what you are saying is that since we can not really experiment in space war, we cannot know.
But many many people have been doing experiments there.  Rail guns, rocketry, nuclear power, life support, etc etc etc.
Experiments.
And math.
And math to plan experiments.
And math to plan 'normal' work, like planning a space mission and designing the satellite.  Or designing a nuclear reactor.

And this simulation is an experiment using maths --- backed by experiments from lots of different sources, solid experiments, many experiments. Some of them using the whole history of electricity, which is quite objective by your standards.

And that may be one more reason to use only the technology we know that works and how it works, instead of mythical photon torpedoes.
 

Quote

Car crash simulators are somewhat of a strawman here because car crash simulators build upon a vast library of existing knowledge that is built up off of both decades of (subjective) theoretical work and review in physics and engineering, as well as previous experience with the operation and crashing of actual working cars and related machines. We know how fast cars generally will travel when they impact each other, and what range of speeds it's possible for cars to achieve, and we know how massive they are, for instance - because we have built millions of cars and we have an infrastructure built around cars. Space battles have none of that.

But we have tons of knowledge on satellite building, mission planning, rocket launching, life support, gee force tolerances.  We have extensive knowledge of armor and penetration of the same.  We have extensive knowledge of aiming.  Knowing the speeds on which converging or passing ships will meet and what speed retrograde vs. prograde orbits meet. We have knowledge of lasers and some of weaponized lasers, we have some knowledge of rail guns and so on.

We have destroyed satellites with ground launched rockets, we have destroyed satellites with guns on manned space stations.  And space does not have slippery roads, high winds, drunken drivers zig-zagging, hills and mountains, tunnels and passes and so on and so on --- space is a very simple and steady environment.
 

Quote

Again, this isn't to say that CoaDE is necessarily wrong in its vision - it just may be stepping over some interesting and viable possibilities. I'd like to see actual simulations of large monolithic lasers, or kinetics platforms, and other alternative designs entirely, even if the ship/fleet design model put forth by CoaDE seems more viable from the developer's own analysis. And I'd like to see space combat models that evolve from our present situation, where space access is expensive and very little infrastructure exists to support the kind of interplanetary economy you'd need to build the kind of warships CoaDE proposes (which is another subjective assumption).

Lasers are possible (up to 1 meter aperture, says the developer), though the developer says that they have not found much use above 10cm --- so you can prove them wrong if you want and my guess is that in that case larger apertures will be offered.  You can build kinetics platforms: just remove all other weapons and add no or a tiny engine.  There is much more that you can build to test out if your ideas work than you seem to think you are able to --- you seem to be thinking that the game is only allowing you very narrow range of vessels.

And if you like to see space combat models evolve from now to then, you may have to mod the simulator or build one on your own --- it's (as fas as I know) not what the designer is interested in right now.  You may need to build your own simulation or mod this one.  This is not a shortcoming of the simulation.

 

On 5/3/2016 at 6:05 PM, tater said:

You make assumptions about the tech (in terms of how you characterize the weapons, etc), then the effectiveness/tactics just have to fall out, with possibly unexpected consequences.

KE weapons should be able to explode into shrapnel, or disperse bearing-balls, for example. At a certain range this guarantees some hits at a much lower mass per impact, but also possibly avoids over-penetration vs a long-rod penetrator. 

Yes: you make 'assumptions' about the tech --- like assuming they behave as the experts say they do --- and then you see what happens.

What is the advantage of few fairly large shrapnel/ball bearing bombs?

• The bombs mean your munitions are explosive and will cause additional internal damage if hit.
• The bombs' explosive charge is likely to explode prematurely when touched by a laser, which means the ball bearings will spread much wider, causing many more of them to miss than usual, thus lessening the work load of PD (many more ball bearings are not going to hit than normally and can be ignored by PD)
• You accelerate not only the warhead (the ball bearings) but also the explosive and trigger, so you need more energy for the same damage at the target.
• A larger number of small ball bearing guns instead of a small number of larger ball bearing bomb guns means
  • more redundancy.  Taking damage or technical difficulties will knock out one or a few small guns, reducing the "broadside weight" only slightly.  If you loose even one larger bomb gun, the amount of damage you deal is reduced much more
  • Since you have a large(r) number of guns, you can decide where to concentrate your fire, splitting it between several targets (say incoming kinetics and enemy vessels) is much more flexible
  • Bombs do create a cone of ball bearings/shrapnel with a random distribution, guns can direct each shot, so you can pattern the shooting so that there is a clear pattern of ball bearings, i.e. your chance for one hit can be done with less ball bearings.
  • smaller guns can fire faster
  • likely more smaller guns are also lighter overall

The only advantage I can see is that you need less computing and tracking power since you have many fewer projectiles you fire off.  But then that should not be any problem with even today's computing power.

On 5/8/2016 at 3:25 PM, tater said:

The accuracy with respect to damage isgoing to depend a great deal on the damage model. Many games screw up at that point, and it's important to remember that in terms of "realistic" outcomes (the desired simulation should be to have realistic outcomes, right?) the game is only as realistic as the least realistic bit. In may cases a game would do better to have a board-game style damage table than "simulation" if the simulation is done wrong, honestly (think of a old board/tabletop/paper game like Harpoon as an example).

Huh?  The simulation tries to be as accurate as it can.  A "damage table" can only be as realistic as a damage table, and a damage table will always be more abstract than a bullet-passing-through-things simulation.  Not for nothing you say "board-game style damage table".  Game.  This is a simulation, not a game.  It's a tool to find out tactics and strategies --- and people like many of us here enjoy such simulations.

Quote

What's the mass of a KE weapon (missile) in this game? What do closing velocities possibly look like? Several km/s? More?

You can argue that the hull is a tin-can, and the missile, call it 100kg (incredibly light, IMO) hits at 5 km/s. There was some talk about it going on one side and out the other. I don;t disagree, but it's considerably more complex than that. The volume it passes through is not empty, it is filled with propellant. This means that there will be at the very least a shockwave moving through the tank, which means some of the energy is deposited along the way. Any sensible projectile in this situation might detonate some distance from the target, and pepper the craft with many, smaller fragments at the same velocity. It might not wreck the target outright, but I see almost any hit as a mission-kill on the target unless the propellant tanks are very compartmentalized (which is a mass-expense).

He mentions spaced armorm, which is light, but effective vs the sorts of threats that ISS faces... but the weapons hitting these warships are not flecks of paint at 5 km/s, they might be more massive penetrators at those kids of velocities. The spaced armor in this case means that more energy, not less, is delivered to the target. If you stop a round, that energy is left in the target, and this means heat. I'm honestly unsure what would happen... it's a non-trivial aspect of the game, and in fact much harder to get at than the orbital mechanics, etc.

I guess you can simply choose what sort of KE weapons you like.  I have seen 1g (0.001kg) projectiles at high speed in screen shots for guns.

And of course remass tanks will be compartmentalized, unless a designer wants the first shot to (mission) kill their vessel.  Yes, it's quite a bit of extra mass.  But the alternative is a one-hit-and-dead vessel, which tends to cause much more mass to hang around with no value than a few more tanks would.
Unless it's a small throw-away drone that would likely die from most hits anyway and thus the the extra mass would not help survivability enough for the cost.

Yes, for some things more massive penetrators may be better.  You can design your vessel and their weapons as you like and see what works.

The amount of energy if you stop a round ... is exactly as much as accelerating the round to that speed.   If you use a gun-like contraption, you need to put in even more energy, as conversion is not perfect and you get quite a lot of heat in your gun, too.   And the gun will heat with every shot, not only with every hit.

Edited May 12, 2016 by weissel
Cleanup

[tater]

2 hours ago, weissel said:

 

Huh?  The simulation tries to be as accurate as it can.  A "damage table" can only be as realistic as a damage table, and a damage table will always be more abstract than a bullet-passing-through-things simulation.  Not for nothing you say "board-game style damage table".  Game.  This is a simulation, not a game.  It's a tool to find out tactics and strategies --- and people like many of us here enjoy such simulations.

Simulation is only as accurate as the least accurate part. It doesn't do any good to track a 1 cm cross section projectile unless the fidelity of the damage model is at the same scale size. If every part internal to the ship is not modeled to 1 cm accuracy, you cannot measure the damage of a 1 cm projectile (or a 1 cm cross section laser). The "simulation" fails at that point, and you'd be better to characterize it otherwise if you want realistic outcomes. If you want to claim simulation, then you need to have ship models with the fidelity of the smallest particle that might hit them.

Take our 1 cm³ blob of steel. It hits the hull, and goes in one side. Its a fuel tank. Is the game (it's a game, sorry) doing the fluid dynamics of the contents of the tank with a hypersonic projectile moving through it? If not, why not? Have they at least been done for typical cases so the damage from that sort of projectile is characterized, or does the game just make simple assumptions--it's a hole, and it deposits X energy, and it leaks? It hits a more critical part... does it hit internal structures within? What does it hit? Did it hit any important piping, wiring harnesses, crew? Did it spall? Into how many fragment? If every single pipe is not modeled, how do you know what actually happened?

That's the problem with the silly claim of "simulation." 

A classic example is age of sail. Ball shot goes into the gun deck. No game I have ever seen tracks the ball, did it hit any metal going in, say a latch on a port lid, turning that into shrapnel? What about splinters, does the game model crossed planking, so that you get variant shard of oak many cm long flying? What about the ball ricocheting off a gun? The reality is that modeling a single round shot hitting a warship would likely be a daunting task to do properly at the scale size of the ball. Just that alone, forget simulating the rest of the ship at such fidelity.

Edited May 12, 2016 by tater

[weissel]

Tater, could you tell me:

What is a "realistic outcome" for space combat?
Space combat is not an age-of-sail combat.  We know what a "realistic outcome" could be in an age-of-sail combat.  Or car crashes.  Or Dreadnaught-class combat.  Or air combat in from WWI to today.  Could be.  There have been many cases where atypical, even "unbelievable" results have been documented.  Results that are not what you'd call a "realistic outcome", certainly not a common outcome.  Yet they happened, in the real world.  They were real outcomes.

So how do you know what is a "realistic outcome" for space combat?
Historic space battles?
Gut feelings or a "vision" how things should be?
Simulations (like this one) with all the many variations of components and their variations in placement?

 

Then you claim you need a layout accuracy of at least the same resolution as the projectile cross section for "realistic outcomes", or you need something like board-game damage tables for "realistic outcomes".  Why?
Your claim "The "simulation" fails at that point".  What exactly does fail?  Why does it fail?
Why would 0.5³m³ or 2³m³ voxels, which know what is inside them and thus what can (and cannot) be damaged by passing through it, be not enough for "realistic outcomes", not even 0.02³m³ voxels, but 0.01³m³ would?  (Ignoring secondary damage, which would need to be handled separately anyway.)

 

Why do you think a simulation needs to be simulating stuff that is irrelevant to the accuracy of the simulation?  Do car crash simulations have to include 3 m/s winds and turbulences caused by butterfly wings nearby?  Do grand strategy simulations handling units of corps size and above need to simulate the view of every point man of each squad and not only of the number of rounds on each soldier and magazibe, but of each single round, including all it's imperfections and the effects of each imperfection when used?

Do Boing 767 simulators (those used to train pilots for emergencies, not your computer games) need to simulate every blade in each turbine?
Does it simulate what parts of a bird striking the engine will come out well roasted and which will be raw minced meat?
Does it simulate which part of which turbine blade passes through the outer engine shell and pierces fuel tanks, fuel lines, wires and/or the passenger cabin?
Does it have a resolution of a couple cm cubed --- the size of a turbine fragment?
Since it does not do all that --- it must not be a simulation, calling is a simulation is (according to you) a "silly claim", right?  Maybe you should tell them that that all their expensive stuff is just a game ...

 

"If every single pipe is not modeled, how do you know what actually happened?" you say --- but how do you know what actually happened with a made-up damage table?  Even 2³m³ voxels will be way more accurate --- your hit table will have a projectile which passes through where there are only remass tanks, but kill a reactor, knock out life support and pierce the engine bell without scratching any remass tanks!  Maybe not every time.  But every now and then.
So much for a "realistic outcome" ....

 

Re your 'classic example': It does very obviously matter if a mast is hit and damaged or even broken or (important) rigging is parted --- or if the projectile misses slightly and only makes a hole in a sail or into thin air.  It does not matter when guy on the gun deck is incapacitated during a battle if that was by a wood splinter, a metal shrapnel, an upturned gun or by the projectile itself.  It does not matter during the battle where the guy is hit: hand, foot, arm, leg, torso, head, it does not matter if he dies immediately or later, becomes a cripple or returns to duty after the battle, the only question is: will the guy be able to perform his duty or not.
It does not matter who the guy is unless his replacement is significantly less efficient.  At some point the reduced man power will reduce combat ability (less people to handle sails, fire guns, repel boarders or swarm the enemy deck), but again --- it does not matter why exactly they cannot perform their duty, only that they cannot.  (And for a campaign it only matters if and roughly when they can perform their duty again.  And maybe for food and water how many die and thus need none any more.)
So please explain what a 'each single splinter, each single shrapnel, every single limb, each single finger' accuracy gives an age-of-sail simulation over a more abstract simulation.  What relevant differences does it cause to an age-of-sail battle, to an age-of-sail campaign?

[regex]

On 5/12/2016 at 11:07 PM, tater said:

If you want to claim simulation, then you need to have ship models with the fidelity of the smallest particle that might hit them.

Dayum...  And here I thought I was pedantic...

People have been calling games "combat sims" for a very long time; it's a genre.  Requiring the player to design the ship's wiring harness and plumbing in order to claim said game is a "combat sim" is ... jealous at best.

[curiousepic]

For those of us not too pedantic to be excited by the game/sim/software... It's been greenlit!

And! Quietly announced as a Steam news post:

Quote

By popular demand, the engine will now support in-engine mod support! Far future or hypothetical technologies can now be implemented as black boxes by modders, supported with an in-engine interface.

 

Edited May 13, 2016 by curiousepic

[tater]

1 hour ago, regex said:

Dayum...  And here I thought I was pedantic...

People have been calling games "combat sims" for a very long time; it's a genre.  Requiring the player to design the ship's wiring harness and plumbing in order to claim said game is a "combat sim" is ... jealous at best.

I was intentionally exaggerating

Look at the context. I suggested they would be better to use lookup tables for damage, and the response was that they were in fact "simulating" it, and it wasn't "a game."

Here's what I was replying to:

On May 12, 2016 at 9:17 PM, weissel said:

Huh?  The simulation tries to be as accurate as it can.  A "damage table" can only be as realistic as a damage table, and a damage table will always be more abstract than a bullet-passing-through-things simulation.  Not for nothing you say "board-game style damage table".  Game.  This is a simulation, not a game.  It's a tool to find out tactics and strategies --- and people like many of us here enjoy such simulations.

He's saying it's a simulation, but it's just lookup tables anyway, without actually understanding what would happen in a given volume area via simulation. I'm arguing it is in fact a game WRT damage unless they actually simulate damage to relevant scale sizes of structures within targets. (good luck with that)

Edited May 13, 2016 by tater

[tater]

16 hours ago, weissel said:

Tater, could you tell me:

What is a "realistic outcome" for space combat?
Space combat is not an age-of-sail combat.  We know what a "realistic outcome" could be in an age-of-sail combat.  Or car crashes.  Or Dreadnaught-class combat.  Or air combat in from WWI to today.  Could be.  There have been many cases where atypical, even "unbelievable" results have been documented.  Results that are not what you'd call a "realistic outcome", certainly not a common outcome.  Yet they happened, in the real world.  They were real outcomes.

So how do you know what is a "realistic outcome" for space combat?
Historic space battles?
Gut feelings or a "vision" how things should be?
Simulations (like this one) with all the many variations of components and their variations in placement?

 

Then you claim you need a layout accuracy of at least the same resolution as the projectile cross section for "realistic outcomes", or you need something like board-game damage tables for "realistic outcomes".  Why?
Your claim "The "simulation" fails at that point".  What exactly does fail?  Why does it fail?
Why would 0.5³m³ or 2³m³ voxels, which know what is inside them and thus what can (and cannot) be damaged by passing through it, be not enough for "realistic outcomes", not even 0.02³m³ voxels, but 0.01³m³ would?  (Ignoring secondary damage, which would need to be handled separately anyway.)

I'm not the one claiming simulation, you are. Since we don't know what realistic outcomes would be, then we can both agree that for this game it's simply made up, with no data.

If you want to claim simulation, then you need to simulate. What happens when a non-microscopic particle hits a LH tank at a closing velocity of 10 km/s? Do you know, because I don't know. To know as best we can, we'd simulate it. That means code that is likely not in this game, including the CFD code to model the projectile through the tank.

If we both admit that this has not been done, then any claims of "simulation" regarding damage are just a different kind of lookup table based on what "seems reasonable," we'd just hope the real world is not counter to our intuition.

Realistic outcomes requires that we have some idea what happens when different things hit a target characterized in the game (and how often they can hit, etc). We need to know what the chances are of a mission kill on the target with certain types of hits. If the game is predicated on ships having many holes punched in them with little effect (the video makes it look that way), then I want to see why a few hundred kg at 10km/s does't deposit more energy/damage in the target than my gut tells me it should. Whats the RL difference between putting a 10g chunk through the hull every m² at 10km/s and hitting it in one place with 400kg? Again, I don't know.

Quote

Why do you think a simulation needs to be simulating stuff that is irrelevant to the accuracy of the simulation?  Do car crash simulations have to include 3 m/s winds and turbulences caused by butterfly wings nearby?  Do grand strategy simulations handling units of corps size and above need to simulate the view of every point man of each squad and not only of the number of rounds on each soldier and magazibe, but of each single round, including all it's imperfections and the effects of each imperfection when used?

You did;t read, or understand what I said. I said that in fact we only need to simulate what is relevant. What happens when particles of 1gm (shrapnel), 10g, 100g, 100kg, and so forth hit a target at various velocities absolutely matters to the simulation of damage. You either abstract it as I suggested, which was poo-pooed because this is a "simulation," or you actually simulate it. My point is that simulating it is not the same as having it hit multiple abstracted volumes where damage is applied (you're just abstracting, but using smaller hit boxes, it's still a made-up abstraction).

 

Quote

"If every single pipe is not modeled, how do you know what actually happened?" you say --- but how do you know what actually happened with a made-up damage table?  Even 2³m³ voxels will be way more accurate --- your hit table will have a projectile which passes through where there are only remass tanks, but kill a reactor, knock out life support and pierce the engine bell without scratching any remass tanks!  Maybe not every time.  But every now and then.
So much for a "realistic outcome" ....

Re your 'classic example': It does very obviously matter if a mast is hit and damaged or even broken or (important) rigging is parted --- or if the projectile misses slightly and only makes a hole in a sail or into thin air.  It does not matter when guy on the gun deck is incapacitated during a battle if that was by a wood splinter, a metal shrapnel, an upturned gun or by the projectile itself.  It does not matter during the battle where the guy is hit: hand, foot, arm, leg, torso, head, it does not matter if he dies immediately or later, becomes a cripple or returns to duty after the battle, the only question is: will the guy be able to perform his duty or not.
It does not matter who the guy is unless his replacement is significantly less efficient.  At some point the reduced man power will reduce combat ability (less people to handle sails, fire guns, repel boarders or swarm the enemy deck), but again --- it does not matter why exactly they cannot perform their duty, only that they cannot.  (And for a campaign it only matters if and roughly when they can perform their duty again.  And maybe for food and water how many die and thus need none any more.)
So please explain what a 'each single splinter, each single shrapnel, every single limb, each single finger' accuracy gives an age-of-sail simulation over a more abstract simulation.  What relevant differences does it cause to an age-of-sail battle, to an age-of-sail campaign?

Damage is damage, and the same fidelity is required. If the ship has crew, they do jobs, so it matters when there are fewer crew once it's past a certain level. It also matters from a "victory" level, winning the battle with a loss of all hands is not desirable.

I'm not the one claiming this is a simulation, I'm FINE with abstraction. But don;t claim that damage is "simulated" accurately if you in fact have no idea if the outcomes map even a little to reality.

Edited May 13, 2016 by tater

[weissel]

Quote

I'm not the one claiming simulation, you are. Since we don't know what realistic outcomes would be, then we can both agree that for this game it's simply made up, with no data.

Since you do not have the slightest idea of the damage handling in the program, you simply made up, with no data, some imaginary problems.  For all we know the ship's interior is modeled to sub-millimeter precision!  (Everything seems to be created procedurally.)

Satellites and their builders have been coping with space debris (11 km/s average impact speed) and micrometeorites (20 km/s average, but up to 70 km/s possible), people have thought long and hard about space combat, flew guns into space and had them fire and hit satellites.

Why should the developer --- with the clearly stated goal to find out about real space combat --- work so hard on modelling real weapons and basically ignore their main function, dealing damage?

 

Quote

If you want to claim simulation, then you need to simulate. What happens when a non-microscopic particle hits a LH tank at a closing velocity of 10 km/s? Do you know, because I don't know. To know as best we can, we'd simulate it.

"If you want to claim simulation, then you need to simulate."  Well, your damage table is nothing but a (very bad) simulation of what damage occurs.

Google would tell you in a minute: nothing happens as that particle disintegrates harmlessly in the Whipple shield, or failing that on the ship's armor or skin or, failing all that, on the tank skin.[1]  See above for debris and micrometeorite speeds.  Mir in her 15 years in space was hit quite a lot, especially her large, fragile solar panels, which cannot be armored (unlike radiators).  Yet all that did not impair the function of Mir.

Would it be accurate to say you simply made up that damage simulation was needed here with no data being consulted?

I'll hazard a guess: a damage table made by you for such an impact would have the particle rip through the tank and another and maybe a third one, rupturing them instantly and causing a huge explosion as the LH contacts warmer parts of the craft and boils off to gas, at the same time the affected spacecraft parts get deep-frozen and brittle, causing them to crack under the sudden pressure rise and the vessel breaks into 2 parts. 

 

[1] a 100µm flour particle (they go from 1 - 100µm, you can see from ~ 40µm, 100µm has more than 15 times the volume, so it is very much non-microscopic) weights below 50µg and has less than 0.0025 Joule of kinetic energy at 10km/s.  You could boil (w/o temperature increase) less than 5.4µg hydrogen with that energy.  Here's a 6.4mm aluminum sphere impacting on aluminum at 7km/s (3600 times as much energy), which also demonstrates a simple Whipple shield.

Quote

That means code that is likely not in this game, including the CFD code to model the projectile through the tank.

If we both admit that this has not been done, then any claims of "simulation" regarding damage are just a different kind of lookup table based on what "seems reasonable," we'd just hope the real world is not counter to our intuition.

Whoa!
How do you get from a non-event you deem near-catastrophic to what is "likely not" in the code of the simulation and asking me to "admit" that that code absolutely is not there?
And how comes you think applying maths and physics and knowledge is identical to making up a damage table from whole cloth 'based on what "seems reasonable"'?

Quote

Realistic outcomes requires that we have some idea what happens when different things hit a target characterized in the game (and how often they can hit, etc).

Realistic outcomes do not depend on anyone's knowledge.  People falling from planes without a working parachute from 1000, 3000, 5000 or more metres height --- you'd probably say that the "realistic outcome" is "100% dead" --- but people have survived that, no matter what your knowledge or gut feelings are.  Not only that, quite a number of them made a full recovery, and some had no more than a sprained leg, or some gashes and a few broken bones to begin with.  That happened more than once.  And it is a realistic outcome, because it's a real outcome, even if it does not happen every time.

 

Quote

We need to know what the chances are of a mission kill on the target with certain types of hits.

We do not know (that's why we simulate damage to a greater accuracy instead of rolling 1D10), and we cannot know: What constitutes a mission kill is the inability to perform the mission.  History teaches us that there are very many possible missions, not all of them require fighting, or fighting capabilities, some don't even require anything except the enemy believing such capability is there.  I point you to the merchant submarine Deutschland, which --- crewed by civilians and totally unarmed --- transported critical materiel through the English blockade twice.  Or the "Milchkühe" (Typ XIV) submarines, unarmed (except for AA defense), refueling and reprovisioning other submarines in 1941/42/early 43.  I also point you to the concept of a "Fleet In Being".

Quote

If the game is predicated on ships having many holes punched in them with little effect (the video makes it look that way), then I want to see why a few hundred kg at 10km/s does't deposit more energy/damage in the target than my gut tells me it should. Whats the RL difference between putting a 10g chunk through the hull every m² at 10km/s and hitting it in one place with 400kg? Again, I don't know

Nothing stops you throwing massive projectiles, as far as I can tell.[5]
But what your guts tell you is not necessarily valid --- your guts would tell you to increase speed to overtake that vessel in front of you in the orbit!

The RL difference is likely that you'll overpenetrate badly with the 400 kg projectile, which --- depending a lot on armor and angle and impact points etc. might make the 10 g projectiles much better damage dealers.  (And that is all assuming you hit with your single shot.  And so on.)

[5] you may be limited by stuff like railgun mass and aiming speed, capacitor mass, and so on, and it may turn out to have distinct drawbacks in combat ... but that is physics, not the simulation.

Quote

My point is that simulating it is not the same as having it hit multiple abstracted volumes where damage is applied (you're just abstracting, but using smaller hit boxes, it's still a made-up abstraction).

So ... not having any however formed hit boxes (and not having a care for computing speed), how would you simulate a complex ship?
How do you find if a projectile's path passes through something?  And if so, through what it passes?  And what the damage is?
I am all ears ...

 

Quote

Damage is damage, and the same fidelity is required.

Yes, slightly stubbing your toe, having your hand smashed and having your head removed are all damage, and we must track exactly where your head flew and if it is further damaged by impact with walls or ground --- and exactly how much your toe will bother you in 4 weeks time.

Quote

 

If the ship has crew, they do jobs, so it matters when there are fewer crew once it's past a certain level. It also matters from a "victory" level, winning the battle with a loss of all hands is not desirable.

 

"If the ship has crew, they do jobs"?  Are you 100% sure this is true in every circumstance?  What again was the job of Gagarin during his space flight, except being a Guinea pig?  The controls were locked away behind a combination lock and everything was automated.

Alas, you cannot win a battle with the loss of all hands.  (You can sink each other, but then both lose.  Which may or may not be a good thing for one side strategically, e.g. in WWI and WWII the UK, with it's much larger fleet, would have gladly lost a ship for a ship of the same type from the German Navy.  The German Navy obviously would not have.)

Finally: You are pointing out that "Damage is damage, and the same fidelity is required." is crap: it is enough to track if there are enough able bodies to do the jobs needed, and if not, how much the capability degrades.

 

Quote

I'm not the one claiming this is a simulation, I'm FINE with abstraction. But don;t claim that damage is "simulated" accurately if you in fact have no idea if the outcomes map even a little to reality.

You are "FINE" with abstraction when the abstraction is extremely abstract and a pure invention made up from whole cloth.
You seem to have terrible problems when the abstraction is more fine grained.  I do not understand why.  Why is a totally fake (and necessarily unrealistic) damage table fine, but a clearly more realistic approach wrong?
And how would you know how damage is simulated or how accurate the simulation is?  Have you studied the code?  Tried the program?  Asked the developer?

 

On 5/3/2016 at 3:19 PM, curiousepic said:

 CoaDE's "vision" is a very specific domain.  But within that domain, the assumptions are few.

I believe the domain is very specific because there are so few other contenders, not because of something inherent in the domain.

[tater]

5 hours ago, weissel said:

Since you do not have the slightest idea of the damage handling in the program, you simply made up, with no data, some imaginary problems.  For all we know the ship's interior is modeled to sub-millimeter precision!  (Everything seems to be created procedurally.)

Satellites and their builders have been coping with space debris (11 km/s average impact speed) and micrometeorites (20 km/s average, but up to 70 km/s possible), people have thought long and hard about space combat, flew guns into space and had them fire and hit satellites.

Why should the developer --- with the clearly stated goal to find out about real space combat --- work so hard on modelling real weapons and basically ignore their main function, dealing damage?

If he's not in destructive testing or computational fluid dynamics in RL as a job (with things like rail guns), he probably isn't simulating it at all. He's making some assumptions, and creating a damage model. Which is JUST FINE. Just don't treat it like SDI people can use his game to decide if their satellite is properly armored, lol.

 

Quote

"If you want to claim simulation, then you need to simulate."  Well, your damage table is nothing but a (very bad) simulation of what damage occurs.

Google would tell you in a minute: nothing happens as that particle disintegrates harmlessly in the Whipple shield, or failing that on the ship's armor or skin or, failing all that, on the tank skin.[1]  See above for debris and micrometeorite speeds.  Mir in her 15 years in space was hit quite a lot, especially her large, fragile solar panels, which cannot be armored (unlike radiators).  Yet all that did not impair the function of Mir.

Would it be accurate to say you simply made up that damage simulation was needed here with no data being consulted?

I said I have no idea, what are you going on about. Mir was not hit by macroscopic debris often (if at all). So what happens when a 1 kg, or 10kg object hits spaced armor (which is all a "Whipple shield" is)? It will not render macroscopic particles (say bearing balls made of tungsten) "harmless" at 11+ km/s, it will mitigate them, possibly (this can likely be simulated decently well as we have penetration data for regular arms, albeit at lower velocities.

Note also that if typical spaced armor defeats X gram impacts at typical velocities, all sides will simply make sure their dispersed particles are large enough to do actual damage. (this reduces to-hit chances, as there is a cloud of larger particles, that are smaller in number over a given volume).

 

Quote

I'll hazard a guess: a damage table made by you for such an impact would have the particle rip through the tank and another and maybe a third one, rupturing them instantly and causing a huge explosion as the LH contacts warmer parts of the craft and boils off to gas, at the same time the affected spacecraft parts get deep-frozen and brittle, causing them to crack under the sudden pressure rise and the vessel breaks into 2 parts. 

 

[1] a 100µm flour particle (they go from 1 - 100µm, you can see from ~ 40µm, 100µm has more than 15 times the volume, so it is very much non-microscopic) weights below 50µg and has less than 0.0025 Joule of kinetic energy at 10km/s.  You could boil (w/o temperature increase) less than 5.4µg hydrogen with that energy.  Here's a 6.4mm aluminum sphere impacting on aluminum at 7km/s (3600 times as much energy), which also demonstrates a simple Whipple shield.

Whoa!
How do you get from a non-event you deem near-catastrophic to what is "likely not" in the code of the simulation and asking me to "admit" that that code absolutely is not there?
And how comes you think applying maths and physics and knowledge is identical to making up a damage table from whole cloth 'based on what "seems reasonable"'?

Microscopic damage is not the issue. We can certainly do the math in terms of energy left in the target (heat), but without some formal information on the physics of hypervelocity impacts, any damage model in this game is just that, a game's damage model. 

Again, I'm fine with a game having a game damage model, and the designer can make some assumptions about what he thinks might happen. It't not a simulation, however, unless it can be checked against real data and verified. Claiming otherwise is absurd.

Regarding spaced armor, this has been well understood for some time. In a world where all the enemy warships are so armored, the penetrators change. Now hit the shield in the image you linked to with a tungsten rod of the same diameter, but several times the length (and energy goes as v² so the velocity matters). Such a long-rod penetrator will only sacrifice the leading cross-section of the rod on the spaced armor. You could disperse these easily by having them wrapped around the missile like a fasces (roman), then spinning the missile slightly (on its long axis) right before dispersal. You now have a cloud of long rod penetrators moving along the velocity vector of the missile. The radial velocity given, and the time to impact establishes the density of rods per cross sectional area that might cross the target. The central missile mode is free to maneuver as well, if that's useful. Or, the impact can just be a 10 kg blob of dense material (steel, whatever), and that Whipple shield in the image is just gone compared to the hit from under 3 grams of Al.

It's important to remember what the spaced armor NASA looks at is designed to deal with, vs people intentionally trying to defeat it.

Quote

Realistic outcomes do not depend on anyone's knowledge.  People falling from planes without a working parachute from 1000, 3000, 5000 or more metres height --- you'd probably say that the "realistic outcome" is "100% dead" --- but people have survived that, no matter what your knowledge or gut feelings are.  Not only that, quite a number of them made a full recovery, and some had no more than a sprained leg, or some gashes and a few broken bones to begin with.  That happened more than once.  And it is a realistic outcome, because it's a real outcome, even if it does not happen every time.

A realistic outcome in a game for this example would be that in a statistical number of samples within the game, you'd expect this to be replicated (the rare case of a non-fatality). A physics model would need to include rather a lot of detail to have this be possible without it happening much of the time. With a table you could require a simple random % roll that matches the RL % of such failures that result in living--if it's 1 in 10,000, then you set the chances to that, and boom, realistic outcomes. It's wrong-headed to think that an incomplete computer simulation is somehow automatically better than paper "simulations" using random numbers and lookup tables. Military (naval, usually) war games could produce reasonable outcomes in this manner---because they calibrated their % tables based upon RL data. They knew how often their battleship gunners hit targets at different speeds and ranges, for example, and they knew that certain events (critical hits) could occur with some % (RN Battlecruisers hit in the magazines, for example). Over large numbers of simulated battles, they could have "realistic outcomes" in their war-games.

 

I'm not saying, and haven't said that such a system is better, or preferable, I've said that with an incomplete "simulation" it might not in fact be any worse. The proof would be real life data.

Quote

You are "FINE" with abstraction when the abstraction is extremely abstract and a pure invention made up from whole cloth.
You seem to have terrible problems when the abstraction is more fine grained.  I do not understand why.  Why is a totally fake (and necessarily unrealistic) damage table fine, but a clearly more realistic approach wrong?
And how would you know how damage is simulated or how accurate the simulation is?  Have you studied the code?  Tried the program?  Asked the developer?

You seem to like putting words in people's mouths. I've made no such claim. I'm fine with arbitrarily fine-grained damage models, my only claim is that they are not likely simulating anything, as we have little data on how real spacecraft systems fare when clobbered by hypervelocity objects. Better examples might be the targets of a-sat tests, or spacecraft collisions. Of course we only know they were utterly destroyed, not the mechanics of their destruction.

I have seen ships in that game (vids) covered with holes (glowing) and they are still mission capable---but I have no idea what size projectiles they were hit with, and at what velocity. I cannot recall if it was up the thread, but someone talked about through and through hits wasting energy... I'm unsure if that would be entirely correct.

Anyway, it's only a simulation to the extent it models real life to some level of fidelity (within the parameters set as givens within the game). KSP is a sort of zeroth order orbital mechanics simulator, for example. Throw in principia, RSS/RO and it's slightly better at simulation. Still a game, but you'll get sort of realistic results (the right craft will be capable of what they are actually capable of to some margin).

CoaDE is a game, and it looks like it might be really fun, just don't make assumptions about accuracy of simulation if you have no idea (and you and I both have no idea).

 

Edited May 18, 2016 by tater

[WinkAllKerb'']

not sure if dune or star wars or star trek or stargate or ... :3

Edited May 18, 2016 by WinkAllKerb''

[regex]

We have a Greenlight on Steam, folks!

[weissel]

On 5/18/2016 at 8:41 PM, tater said:

If he's not in destructive testing or computational fluid dynamics in RL as a job (with things like rail guns), he probably isn't simulating it at all. He's making some assumptions, and creating a damage model. Which is JUST FINE. Just don't treat it like SDI people can use his game to decide if their satellite is properly armored, lol.

Unless your day job is simulation or designing and writing computer programs, you probably don't know what the statistical coincidences are between "current held job" and "likelihood to 'simulate' [as per your definition of 'simulate'].  You of course need to do statistics as your day job to understand statistical coincidences.  Or maybe that "job -> likelihood to do/not do things" relation is just prejudice.

Also: Does a "simulation" in your meaning of the word only simulate atoms, are individual electrons, neutrons and protons needed or do you insist on every single quark to be in the simulation?

Also: every simulation makes assumptions.  That's called the "model" of the thing you want to simulate.

Also: unless you are one of those SDI people, you would not even know if and how they armor their satellites.

On 5/18/2016 at 8:41 PM, tater said:

Mir was not hit by macroscopic debris often (if at all).

Merriam-Webster's reliable book defines "macroscopic" as "1 : observable by the naked eye".  So do other dictionaries.  So, yes, Mir was hit by a lot of macroscopic objects, be they debris or (faster!) micrometeoroids.

On 5/18/2016 at 8:41 PM, tater said:

So what happens when a 1 kg, or 10kg object hits spaced armor (which is all a "Whipple shield" is)? It will not render macroscopic particles (say bearing balls made of tungsten) "harmless" at 11+ km/s, it will mitigate them, possibly (this can likely be simulated decently well as we have penetration data for regular arms, albeit at lower velocities.

Note also that if typical spaced armor defeats X gram impacts at typical velocities, all sides will simply make sure their dispersed particles are large enough to do actual damage. (this reduces to-hit chances, as there is a cloud of larger particles, that are smaller in number over a given volume).

Whipple Shields are similar, but not identical to spaced armor, see stuffed Whipple Shields.  What happens with a 1, 10, 100kg object hits spaced armor depends on the speed difference, the armor itself, the angle of impact, the form of the projectile, etc.  If the armor is designed to cope with that specific set of parameters, nothing much will happen.

Please do research the shielding of the ISS, the data is public and available, before speculating wildly! (Design criteria was a 1 cm aluminum sphere at 70 km/s (that being the top speed of micrometeorids)). Yes, a shielding designed to cope with tungsten bearing balls (or the ISS shields) will make them harmless.  Small hole in the bumper(s), some scorch on the catcher, that's all.

Worst for Whipple Shields: slow impacts.

Note also that in sea warfare the guns and armor of a WWI or WWII destroyer would have had little impact on a battleship or even cruiser, but the same would not have been true in the reverse.  According to your theory all destroyers would have been upgraded to guns able to pass through battleship armor.  Same with land combat: fortified positions and forts do care little about the weapons of an infantry man or even many tanks, yet infantry does rely on rifles and MGs.  And again the same in air combat, say between armed transport helicopters, bombers and fighters.
 

On 5/18/2016 at 8:41 PM, tater said:

without some formal information on the physics of hypervelocity impacts, any damage model in this game is just that, a game's damage model. 

Any game, by definition, will have a game's damage model.

Any simulation --- be it some sort of "physics" in some jump&run game, KSP, flight simulators (for training real life pilots), weather forecast --- is using a model to approximate a real or fictitious system.  For example weather simulation on a national scale will ignore the fact that urban canyons can funnel wind and affect weather in the canyons a lot.  Since weather forcatsing do not model everything, I guess they are not using a simulation but a game, according to you.

 

But that is beside the point.  The point is that you seem to be either 'roll a D10 on a damage table for realistic outcomes" (the developer has said that the way things turn out in his simulation are surprising him again and again, so, without simulation, how would you know what table offered that?) or a simulation to the last molecule, if not to the last quark!  Any damage model that is in between (i.e. while not perfect it will not cause the engine to be shot off with a shot scratching the bow) is not good enough for you.

 

Quote

It't not a simulation, however, unless it can be checked against real data and verified.

"What happens when a non-microscopic particle hits a LH tank at a closing velocity of 10 km/s? Do you know, because I don't know. To know as best we can, we'd simulate it. That means code that is likely not in this game, including the CFD code to model the projectile through the tank."

So ... you call it simulating even if you cannot check it against "real data" (because if we have that, we'd simply use that!) nor verify it against "real data".  Could you please at least be consistent in your own use of terminology?
 

On 5/18/2016 at 8:41 PM, tater said:

Regarding spaced armor, this has been well understood for some time. In a world where all the enemy warships are so armored, the penetrators change.

Really?  How would you know?  Did you run a 'simulation', ah, a 'game' in your head? (It could not be a 'real' simulation by your standards, or could it? --- no real data!)  Or did you just have some made up  assumptions how the world should work, AKA a model for a simulation?

Fact is: In history there have always been more and less armored variants.  There's heavy and light infantry, heavy and light cavalry,  there's the wood&sail navy with all their different ship classes, from dispatch boats over frigates to ships of the line, there's the WWI / WWII navy, with both destroyers and battleships, there's the Japanese Zeke/Zero and the Buffalo Brewster ...

Now, obviously, it would be of a distinct advantage to have, say, destroyers regularly withstand attacks from cruisers and battleships.  Now, tell me, why did they not build destroyers that way in the late 19xx to mid 20xx?  And in the light of that: would it be advantageous to have all space warships armored up the same amount?  And why not?  So would you agree that "a world where all the enemy warships are so armored" is pretty unlikely, unless the enemy just has one light class of vessel?
 

Quote

Or, the impact can just be a 10 kg blob of dense material (steel, whatever), and that Whipple shield in the image is just gone compared to the hit from under 3 grams of Al.

It's important to remember what the spaced armor NASA looks at is designed to deal with, vs people intentionally trying to defeat it.

We were talking about "What happens when a non-microscopic particle hits a LH tank at a closing velocity of 10 km/s? Do you know, because I don't know. To know as best we can, we'd simulate it. That means code that is likely not in this game, including the CFD code to model the projectile through the tank."  Technically, yes, 10 kg blobs are 'non-microscopic', but at least in my world, you'd not call them "particle". 
So please stop moving goal posts by astronomical units after being shown that, no, for particles you don't need CFD modelling as they never reach the tank.

And remember:  "It's important to remember what the spaced armor NASA looks at is designed to deal with" (which would be said particles) "vs people intentionally trying to defeat it."  So why do you use particle Whipple Shields versus 10 kg blobs?

 

On 5/18/2016 at 8:41 PM, tater said:

A realistic outcome in a game for this example would be that in a statistical number of samples within the game, you'd expect this to be replicated (the rare case of a non-fatality). A physics model would need to include rather a lot of detail to have this be possible without it happening much of the time. With a table you could require a simple random % roll that matches the RL % of such failures that result in living--if it's 1 in 10,000, then you set the chances to that, and boom, realistic outcomes. It's wrong-headed to think that an incomplete computer simulation is somehow automatically better than paper "simulations" using random numbers and lookup tables. Military (naval, usually) war games could produce reasonable outcomes in this manner---because they calibrated their % tables based upon RL data. They knew how often their battleship gunners hit targets at different speeds and ranges, for example, and they knew that certain events (critical hits) could occur with some % (RN Battlecruisers hit in the magazines, for example). Over large numbers of simulated battles, they could have "realistic outcomes" in their war-games.

 

I'm not saying, and haven't said that such a system is better, or preferable, I've said that with an incomplete "simulation" it might not in fact be any worse. The proof would be real life data.

You may notice, if you reread what you wrote, a very heavy reliance of "RL %" and "calibrated their % tables based upon RL data" and  "They knew how often their battleship gunners hit targets at different speeds and ranges, for example, and they knew that certain events (critical hits) could occur with some % (RN Battlecruisers hit in the magazines, for example)."
Now, if you could point me to a reasonable number of space battles so we can get reasonably accurate RL data to calibrate such a look up table ... especially considering that people are going to build their own ships, unlike most naval table top games, we'd need lots and lots of accurate RL data ...
... and come to think of it, Taffy 3.  Can these naval table top games cope with that bit of RL data and repeat it?

 

A realistic outcome for this example (surviving 1000+ meter falls, sometimes almost unhurt, with no parachute), using the technology of military war games and assuming 1:10,000 chances, would have a table of:

Fall damage, 1000+ meters. Roll 1D100:

0 - 99	you die!
100	Roll again on Table "Fall damage, 1000+ meters - B"

Fall damage, 1000+ meters - B. Roll 1D100:

0 - 99	you die!
100	You live! Roll on Table "random damage" to find out how much you hurt.

and of course a damage table ...

Now, can you name me naval table top games that model a chance of non-explosion of a direct, critical magazine hit & shell explosion?  And if so, is the chance of non-explosion of no more than 1:10,000 modeled, or is it just a case of "the magazine explodes, let's see if this tears the ship apart or not"?  If no, could it be that they want 'realistic' examples, not examples that mirror the whole range of possible (and sometimes really happened in the real world) outcomes?

 

On 5/18/2016 at 8:41 PM, tater said:

I'm fine with arbitrarily fine-grained damage models, my only claim is that they are not likely simulating anything, as we have little data on how real spacecraft systems fare when clobbered by hypervelocity objects.

Ah, yes, your "if you don't have real world outcomes, you cannot simulate".  As you put it: "What happens when a non-microscopic particle hits a LH tank at a closing velocity of 10 km/s? Do you know, because I don't know. To know as best we can, we'd simulate it." And actually we have quite a good idea how real spacecraft systems fare when clobbered by hypervelocity objects --- it's not as if there were not thousands of satellites up there and there is a certain ISS up there, never mind all the space stations that have been up and have come down again.

You may have a point regarding huge hypervelocity objects, so ... are you up to shooting some targets into space and get RL examples?  Or would a simulation suffice?

 

Some of the statements you made in this thread: 

1. "The accuracy with respect to damage isgoing to depend a great deal on the damage model. Many games screw up at that point, and it's important to remember that in terms of "realistic" outcomes (the desired simulation should be to have realistic outcomes, right?) the game is only as realistic as the least realistic bit. In may cases a game would do better to have a board-game style damage table than "simulation" if the simulation is done wrong, honestly (think of a old board/tabletop/paper game like Harpoon as an example)."
2. "Simulation is only as accurate as the least accurate part. It doesn't do any good to track a 1 cm cross section projectile unless the fidelity of the damage model is at the same scale size. If every part internal to the ship is not modeled to 1 cm accuracy, you cannot measure the damage of a 1 cm projectile (or a 1 cm cross section laser). The "simulation" fails at that point, and you'd be better to characterize it otherwise if you want realistic outcomes. If you want to claim simulation, then you need to have ship models with the fidelity of the smallest particle that might hit them."
3. "Look at the context. I suggested they would be better to use lookup tables for damage"
4. "He's saying it's a simulation, but it's just lookup tables anyway, without actually understanding what would happen in a given volume area via simulation. I'm arguing it is in fact a game WRT damage unless they actually simulate damage to relevant scale sizes of structures within targets. (good luck with that)"
5. "If you want to claim simulation, then you need to simulate. [...] That means code that is likely not in this game, including the CFD code to model the projectile through the tank.
   If we both admit that this has not been done, then any claims of "simulation" regarding damage are just a different kind of lookup table based on what "seems reasonable," we'd just hope the real world is not counter to our intuition."

But you say now "I'm fine with arbitrarily fine-grained damage models".  That must be the reason why you are so insistent on "a board-game style damage table" being better, that an arbitrarily fine grained damage model is not good, it must be really fine grained: "It doesn't do any good to track a 1 cm cross section projectile unless the fidelity of the damage model is at the same scale size".  And CoaDE uses "just lookup tables anyway", implying it is no better than your "board-game style damage table". 

And you admit that even your "board-game style damage table" would be pure guesswork "[...] then any claims of "simulation" regarding damage are just a different kind of lookup table based on what "seems reasonable,"" based on "our intuition", which we "just hope the real world is not counter to".

Quoting the developer (who likely has seen the code he produced and likely knows where tables are used and at which granularity stuff happens):

Quote

As you know, it started with the question: “What would space warfare actually be like?” A lot of sources, mostly hard science fiction literature, have tried to answer that question at one point or another. Other sources, like the Atomic Rockets website, does a great job of detailing the ideas at a high level.

For me, though, I wanted a simulation, one that was actually based on real equations. This is because in my experience, whenever you develop system this complex, it tends to surprise you, and will often overturn your assumptions. And trust me, it has surprised me continuously throughout the project.

[emphasis not in original]
Since a damage table is based on assumptions, even when real-life data is available --- how likely is what happened to the Hood?  If you rerun the scenario over and over, what would be the most common outcomes? --- how do you think it'd work when your assumptions are overturned at every corner?  You can properly calibrate to real life only if you have a large enough sample size.  And wet navy ships' capabilities are harder to collect, since between classes of even the same type there are important differences.  Shells that hit an enemy given these guns and that crew quality --- sure, shells are fired in hundreds.  Shells that hit are a bit more rare, but still ... The chance that a shell is on a path that will lead it to the magazine, actually breach the same (or for British WWI ships, penetrate turret or barbette of the main guns and cause a cordite explosion) and destroy the ship ... how often did that happen?

 

[tater]

You keep putting words in my mouth. I said that a lookup table might be every bit as accurate as a poor "simulation," where "simulation" is here defined as a computational damage model making simple assumptions at each step (say following a particle as is passes through a target).

A particle can be any size, I'm using the term for ease of discussion, nothing more. We're dealing with space, and astronomical velocities, so anything smaller than masses that matter for solar systems is a "particle." "Macroscopic" to me would relate to the general picture of significant events being modeled, ignoring the low-order stuff that doesn't matter (paint chips, etc). We can pick another indeterminate word if you like, for whatever particle size has non-trivial damage, if it was entirely natural we might use micrometeorite/meteorite/asteroid, but we don't have vague terms past small, medium, and large. Note that I'd call anything heavier than Helium a "metal," as well (old astronomy habits die hard).

My entire point in this discussion is to suggest that unless this can actually replicate known data accurately, we have no idea how accurate a simulation it is, and it might in fact be no better than a lookup table with "intuitive" damage filling the slots. It might be far better, but we won't really know.

If the damage done by micrometeorites and orbital debris is well-characterized with spaced armor, then that is a good anchor point for a model... if the model replicates what is known to happen (with a similar armor level in game to ISS applied vs similar particles). Then we know that the model works for small particles of a given type (X mm Al ball at Y km/s). What happens for larger particles, and how shape matters is another issue. Impact angle is at least somewhat straightforward as there is decent data on sloped armor penetration (it is slightly more complex than just a longer path length through armor because of angle, but well characterized). So that's a plus.

Let's forget every particle that is defeated by the armor. Those deposit energy to the hull to be radiated away (or not), and nothing more. That's easy to characterize. Where it becomes non-trivial is where the penetrators actually do their job and defeat the armor. Now is when the tricky damage modeling comes in. Above someplace we talked about through and through hits (in one side, out the other). That's is a clear case where what the particle(s) hit, and what happens when that occurs, matters (again, the particle might be a telephone pole chunk of metal for all I care). If the interior is not actually modeled, then it's functionally a lookup table, right? X cm particle enters crew compartment at Y km/s that is not modeled fully down to an X cm scale. The game has a check for crew compartment damage, and applies it. That's not substantially different from a miniatures game, it's just likely finer grained (the ship might have many thousands of hit areas instead of hundreds or fewer).

In short:

The extent to which this damage model mimics reality is the quality of the simulation. Since we have little real data, it is by definition speculative past the range for which we have data. I have no problem with this, while you seem to be insisting that it is a great simulation without evidence. I am making no positive claims whatsoever, as I have said a few times, it might be a great simulation, it might be no better than a lookup table in a board game mapped to correspond to whatever real data we have on one end, and it might in fact be worse than that table. Short of real data, there is no way to tell.

Edited May 25, 2016 by tater

[KSK]

I don't really have a feel for how realistic this game is as a spacecraft combat simulator, although I do lean towards tater's point of view. However, as a space warfare simulation it seems rather contrived in that it's predicated on your enemy meekly allowing you to get all this wonderful hardware into space in the first place. Which sounds optimistic to me, to put it mildly.

To quote a probably apocryphal phrase: amateurs talk tactics, professionals talk logistics. 

A large mothership with drones may be a wonderful fighting machine on orbit but if I can whack it with something fast, heavy or nuclear before it ever leaves the atmosphere, then I'm not going to care.

I strongly suspect that orbital mechanical duels between spacecraft are only going to be an extremely small part of a space war, no matter how scientifically accurate.

[curiousepic]

12 minutes ago, KSK said:

I don't really have a feel for how realistic this game is as a spacecraft combat simulator, although I do lean towards tater's point of view. However, as a space warfare simulation it seems rather contrived in that it's predicated on your enemy meekly allowing you to get all this wonderful hardware into space in the first place. Which sounds optimistic to me, to put it mildly.

To quote a probably apocryphal phrase: amateurs talk tactics, professionals talk logistics. 

A large mothership with drones may be a wonderful fighting machine on orbit but if I can whack it with something fast, heavy or nuclear before it ever leaves the atmosphere, then I'm not going to care.

From the game's name, I'm presuming the justification/setup is simply that warfare only breaks out after there is a significant human presence and infrastructure in space, spurred by Earth being made unliveable in some way.

Edited May 27, 2016 by curiousepic

[KSK]

That would certainly make sense and it's a perfectly reasonable (if slightly depressing) premise for a game. However (in my opinion) it does make any claims to to the game discovering how space warfare would work, a bit hollow. 

Discover how space warfare works once you set things up specifically such that space warfare involves fleet combat.

Admittedly I have a bias against that premise in any case, so I'm definitely not a neutral commentator here.

 

[Accelerando]

If you can get any large amount of infrastructure in space, then it can be used to develop weapons, if not used as weaponry itself, so while I would personally predict a future of space warfare revolving mostly around ASAT weapons and related tech for the next (few) centur(ies), I wouldn't rule out large-scale space warfare altogether if we assume that there is ever going to be any sort of large-scale space development. I'm somewhat optimistic myself, for reasons that I've elaborated on before, although I'm not holding my breath for it.

However, any future history calling itself "accurate" to some degree needs to at least take into account material considerations, and how does one propose to build large-scale space infrastructure up enough to create heavy-duty space warships if Earth itself, likely the site of your entire starting industrial base and supporting population for a long time to come, is in jeopardy? This premise, which judging from its name CoaDE presumably shares with many other SF stories, always mystifies me in its cognitive dissonance:

"Earth is f****d and everyone is dying; let's harness the ruined global industrial base to crank out rockets to build giant spaceships with lots of guns so we can shoot them at each other."
"Sir, all the factories are burning and the entire population is rioting, starving, or evacuating their homes in the face of warfare and disaster--"
"GIANT SPACESHIPS WITH LOTS OF GUNS"

I guess we'll see how that works out, anyway.

To follow along with tater's neat conclusion to this mess of a discussion, CoaDE is a "simulation" in the sense that SimCity or Dwarf Fortress is a simulation -- it takes a set of extremely abstracted assumptions and interacts them with one another to produce interesting results that may square with reality to some degree, but that doesn't make it a military-grade simulation. This is why I always chuckle when hard SF creators purport to be objective in some way when in reality they're writing down a list of (perhaps reasonable, but hardly objective) assumptions and dressing it up with pretty polygons to look like a spaceship.

Edited May 29, 2016 by Accelerando

[Diemosthenes]

Children of a 'Dead' Earth, not Children of a 'Dying' Earth. 

In the CoaDE world, earth is useless in terms of production and population and all resource management is derived from accessible materials in our solar system by the separate factions who have ownership over different areas. I can't tell you the exact timeline, but humans have already spread throughout the system where they can make it work, and the game takes place as humans finally decide to f***up the newest thing they got and begin warfare in space.

Edited May 29, 2016 by Diemosthenes

[Accelerando]

2 hours ago, Diemosthenes said:

Children of a 'Dead' Earth, not Children of a 'Dying' Earth. 

In the CoaDE world, earth is useless in terms of production and population and all resource management is derived from accessible materials in our solar system by the separate factions who have ownership over different areas. I can't tell you the exact timeline, but humans have already spread throughout the system where they can make it work, and the game takes place as humans finally decide to f***up the newest thing they got and begin warfare in space.

And so... how did they get to this point? By bootstrapping off of a dying Earth? I'll be happy to take back my reservations about the setting if there's some detail about how they actually develop their spaceborne infrastructure before Earth is all gone and ruined.

At any rate, the "Earth is destroyed but humanity is safe because we went to space instead" trope is pretty played out, but whatever.

Edited May 30, 2016 by Accelerando

[FungusForge]

I'm buying this game whenever it gets released, because despite being a game second, it looks like great fun.

[regex]

On 5/30/2016 at 6:22 PM, Accelerando said:

At any rate, the "Earth is destroyed but humanity is safe because we went to space instead" trope is pretty played out, but whatever.

I take it more as a reason for humanity to be in space in the first place.  It's probably the most realistic reason for us to create colonies elsewhere in the system barring some random MacGuffin; ruining Earth is something humans are very good at.

I'm looking forward to this game because the author's claims all appear to be true at this point.  Even if the actual space warfare looks nothing like it actually will in the year 2525 it'll still be the most scientifically accurate space warfare simulator ever made.