Jump to content

KSP 2 Multiplayer Discussion Thread


Recommended Posts

This was an item I listed as part of a comment I made on the hopes and wishes thread, that I think would be an excellent feature in KSP 2:

 I think it would be really cool to see resource trading in multiplayer to act as an incentive to play the game more. It would also be a good mechanic to diversify interactions with other players. Maybe someone else has spec'd out into mining X resource on a celestial body which is basically the only place to get that resource, and you need some. Rather than go to the trouble of constructing a huge refinery there (or facing an attack of some sort if they have their refinery on the only good deposit), you could trade with that person using something that you have. Maybe you're the only guy on the (server?) with a high enough tech level to manufacture Wyvern engines. You contact them, and strike up a deal. Side note: making certain critical resources endemic to certain areas (especially small ones) would be a good way to encourage staking of territories.

I think competition would be just as exciting as cooperation in multiplayer contexts, and provided the physics engine is overhauled to allow multi-km/s collisions there could be... interesting interactions... if there ends up being resource or territory disputes. That unescorted guy hauling a thousand tons of valuable explodium sure doesn't have the dV margin to maneuver off his hohmann trajectory nor the acceleration to evade if fast movers start braking hard at him... who's to say we can't jet up, blow up his engines with some sepratron missiles, siphon off a hundred tons and then make a getaway? Maybe the guy who issued that space trucker's contract, who's now on his way to drop a class C rock on your base!

Link to comment
Share on other sites

Oh, and that gives me another idea: radar ranges! Imagine if, depending on the construction of your ship, you could only see other (unannounced, coasting and cold) players' ships if they get within several million meters of you. You could get better intel if you are in touch with one of your surface bases that has more powerful sensors, but then you have to rely on speed of light delay and that would bring its own challenges. Only detecting a would-be pirate or attacker at the very end of their coast phase would give you time to act, but not enough that it becomes a game of delta-v attrition. Alternately you could spend the extra mass and funds to mount a huge radio telescope to your ship to ensure that is never a problem. Either way, firing up an engine or heating up a radiator would be a surefire way to be seen by everyone in that system. Maybe a unique engine idea would be a cool-running solid projectile railgun-type thing, that allows you to give yourself gentle nudges while under the radar? If an enemy's "last known" trajectory line for your ship has you coasting around Tylo, the coldgun might just give your ship the few m/s nudge it needs to wildly alter your future trajectory via gravity assists and throw them off!

Link to comment
Share on other sites

On 1/15/2021 at 12:41 PM, notrealaccount said:

So this for the loskene question how the KSP 1 multiplayer mod works with time warp simply if you warp you go into diffrent time and you need to press tiny button to go back where everyone else are and i guess that devs are going to make that (they havent sayd or told anything about this) if you just warp and other player don't they just see your craft recorded so like if you warp they just see you if they follow you land on moon if you land or something

 

the thing about that is that if you timewarp x1000000 for like a minute and no one else does, ksp2 will have to keep in its memory 1000000 years of you floating around in space and simulate it for others

On 11/13/2020 at 2:01 AM, Multivac said:

On a separate note: The main thing I, personally, would want to see in KSP2 multiplayer is the ability for each player to run their own separate space agency, with their own funds, science, reputation, and crew roster. These different agencies could compete for achieving "firsts," compete for contracts from a common pool (and if a given agency fails a contract, not only could they be penalized, but their competitors could be given a shot at the contract instead!) or, on the other hand, could cooperate to achieve bigger objectives together, say with each player/agency contributing a module to a giant space station, or contributing something to a multi-stage interplanetary mission. If the larger mission succeeds, each player could then earn a reward commensurate with their contribution!

Another, related, idea might be giving players the ability to create contracts for each other! Say you launch a satellite, but can't quite put it in the right orbit before it runs out of delta-v; or say you land on the Mun, but your rocket tips over and your crew is stranded. Instead of sending a rescue mission yourself, you could create a contract for another agency to rescue your crew / refuel or reposition your satellite for you, and offer them funds, science, or reputation from your own budget in compensation! You would still benefit by getting your crew back (the whole Kerbonauts-join-their-rescuer's-crew mechanic would probably best be disabled for inter-player contracts) or by getting a satellite in the right orbit (potentially achieving the objective for your own contract, and thus getting your own reward), and your rescuer would not only get the reward you'd put up for them, but would also gain the satisfaction of knowing they helped another human being (as opposed to merely completing a computer-generated contract for a fictional agency run by a random number generator). There'd even be an incentive for players to specialize; one player might become the server's expert on planetary landings, for example, while another might run an orbital tug, becoming an expert in rendezvousing with other players' craft and refuelling them or efficiently putting them on new orbits.

Notably, this is something that multiplayer mods for KSP1 cannot currently do — currently, the only form of multiplayer available, even through modding, seems to be everyone sharing a common funds/science/reputation/Kerbonaut pool. I think everyone running their own separate agency would be a good way for a more competitive, yet still constructive, style of play to work in KSP2, and I think it'd add a lot to the game experience!

 

I 100% agree with this person's whole proposal

23 hours ago, Wubslin said:

Oh, and that gives me another idea: radar ranges! Imagine if, depending on the construction of your ship, you could only see other (unannounced, coasting and cold) players' ships if they get within several million meters of you. You could get better intel if you are in touch with one of your surface bases that has more powerful sensors, but then you have to rely on speed of light delay and that would bring its own challenges. Only detecting a would-be pirate or attacker at the very end of their coast phase would give you time to act, but not enough that it becomes a game of delta-v attrition. Alternately you could spend the extra mass and funds to mount a huge radio telescope to your ship to ensure that is never a problem. Either way, firing up an engine or heating up a radiator would be a surefire way to be seen by everyone in that system. Maybe a unique engine idea would be a cool-running solid projectile railgun-type thing, that allows you to give yourself gentle nudges while under the radar? If an enemy's "last known" trajectory line for your ship has you coasting around Tylo, the coldgun might just give your ship the few m/s nudge it needs to wildly alter your future trajectory via gravity assists and throw them off!

You're gonna join a ksp2 war server immediately, aren't you? You seem to have it all planned out.

Link to comment
Share on other sites

2 hours ago, ImANoob said:

You're gonna join a ksp2 war server immediately, aren't you? You seem to have it all planned out.

I just think that gameplay would be able to benefit massively from both cooperation and competition, whether that competition takes the form of a space race or violence. I think an ideal "large server with lots and lots of ships"-model of KSP 2 multiplayer should be an environment which is principally peaceful trading and development, but provides the opportunity for coalitions to form between players and the potential for those groups to have clashing interests. Piracy, territory disputes, trading, policing of orbital airspace and resources. I think that an ideal multiplayer KSP 2 would have the potential for all of these things.  Not that I wouldn't appreciate me a server or two dedicated solely to fighting in space &)

Link to comment
Share on other sites

On 1/20/2021 at 8:54 AM, ImANoob said:

You're gonna join a ksp2 war server immediately, aren't you? You seem to have it all planned out.

Hell yeah, dude! The nanosecond BDArmory is ported I'm starting a war server. Instead of generic shooter #147, we get to design our troops and units and the contest will be a test of "who can make the best close air support craft the cheapest" and not "who can dropshot you around a corner the fastest".

Link to comment
Share on other sites

I used to think multiplayer would transform games onto another level, but the reality is it almost universally wrecks them.

I have no interest in a multiplayer KSP. I can't see any need for it, or imagine how it would be a positive experience.

The time and money would be better spent on developing other novel game mechanics, imo.

Link to comment
Share on other sites

7 hours ago, Kernel Kraken said:

Hell yeah, dude! The nanosecond BDArmory is ported I'm starting a war server. Instead of generic shooter #147, we get to design our troops and units and the contest will be a test of "who can make the best close air support craft the cheapest" and not "who can dropshot you around a corner the fastest".

Close air support? You can play War Thunder or DCS World or something if you want that. The market is saturated with generic multiplayer dogfighting games. Do you know what there isn't a single one of? Games that allow multiplayer orbital combat. I NEED the ability to play a non-turn based version Children of a Dead Earth with another actual human and god help me I will do whatever it takes to twist and mutilate the game with 9,000 mods if I have to, just to see what it's like. We are so starved of the experience of multiplayer combat games featuring orbital mechanics. The closest game I can even think of that has done anything like this was the game Spacewar!, and that came out in 1962. I can understand and fully appreciate a KSP 2 where the main focus of the game is collaboration or friendly competition in space, but make no mistake. No matter how peace-oriented the game is, the moment my friends and I get it we're going to get into pro/retrograde orbits of Kerbin and try to smoke each other with thick clouds of TD-06 decouplers in a big jousting match just to see if we can, and there's no one who can tell us otherwise. It's going to be awesome.

4 hours ago, Bedazzled said:

I used to think multiplayer would transform games onto another level, but the reality is it almost universally wrecks them.

I have no interest in a multiplayer KSP. I can't see any need for it, or imagine how it would be a positive experience.

The time and money would be better spent on developing other novel game mechanics, imo.

Come on. You wouldn't even want to play with like, another friend in KSP? That couldn't make for a bad experience.

Edited by Wubslin
Was being a huge jerk
Link to comment
Share on other sites

14 minutes ago, Wubslin said:

No matter how peace-oriented the game is, the moment my friends and I get it we're going to get into pro/retrograde orbits of Kerbin and try to smoke each other with thick clouds of TD-06 decouplers in a big jousting match just to see if we can, and there's no one who can tell us otherwise. It's going to be awesome

I'm sure mods will support those that want to put weapons in space. However, since the game is rather realistic, the realistic nature of "space combat" would be rather boring, more akin to naval anti-sub combat rather than Star Wars, where finding the enemy and taking a position up to attack takes up almost the complete majority of your time. 

 

5 hours ago, Bedazzled said:

I have no interest in a multiplayer KSP. I can't see any need for it, or imagine how it would be a positive experience.

I don't see how this is an all or nothing thing.

Multiplayer support usually means you can rope your friends into the game to play with you. That alone should give you more incentive to approve it. The more people that buy the game means more funds go back to keep the game going, which means more features for you and all your multiplayer friends.  Even if you have no friends to rope in you can still enjoy a game getting features and updates from the game just being more popular because it supports fun multiplayer.

Even the simplest forms of multiplayer I can think of ("master control warp") still seems pretty fun and should be enjoyable for the average non-hardcore player. The "host" can be the veteran lugging up a bunch of engineers (other players) to build something in orbit and hi-jinks ensue. Or imagine trying to build a rocket at the same time in the builder with other people. Not only would it be more realistic since you have to work together, but it should be more fun at that.

I personally like the idea having an option of forcing IVA mode, which not only would improve a multiplayer experience with multiple people in the same flight, but also single player experience for those that really want a challenge. 

Finally the argument that it would take to much time to support probably isn't worth much, considering multiple mods have already been able to accomplish it in the current game with different degrees and different degrees of success. If its built in from day 1, its more a question of how far they went with it rather than how much time they spent adding it. 

Link to comment
Share on other sites

44 minutes ago, MKI said:

I'm sure mods will support those that want to put weapons in space. However, since the game is rather realistic, the realistic nature of "space combat" would be rather boring, more akin to naval anti-sub combat rather than Star Wars, where finding the enemy and taking a position up to attack takes up almost the complete majority of your time. 

"Boring" is relative. I've spent almost 990 hours watching green space frogs just fall, for years at a time. Besides, if you do want to take a completely realistic approach to it it certainly will not be like attack submarine cat and mouse. In contradiction of some possible mechanics I mentioned earlier in the thread, there is no stealth in space. Everyone is going to be able to see where everyone else is at all times. Have you ever played Children of a Dead Earth? It makes for crazy gameplay. And considering that game would wait for you to conduct a (non-impulsive) burn before allowing the enemy AI to make a move, there could be many as-of-yet undiscovered elements to the art that will emerge as a result of having two players go against each other in real or accelerated time. I'm gonna repost a big huge essay I typed out on reddit about an ideal multiplayer space combat game I had envisioned:

 

I so badly want a version of CoaDE that's multiplayer and uses only chemical rockets in the Earth-Moon system circa an alternate universe 1970s-1980s. Both players could make maneuvers in turns (player 2 would thus able to initiate burns while player 1 was in the middle of one, and once player 2 was done initiating the burn player 1 would be able to cancel his if so desired etc.), and when an intercept is made the relevant engaging fleets would vote for whether or not the "live action" encounter occurs in real time, 0.1x time, 0.001x time, etc.

Imagine a game where player 1 has the goal of strategically killing a single target station in LEO starting from Lunar orbit, and player 2 commands a vast fleet of a thousand sentinel drones all in highly eccentric Earth orbits with their apse lines woven through the equator like bicycle spokes a few hundred km above the altitude of the target station.

Player 1 makes a lunar prograde burn just as the Earth fully crests the horizon, thus beginning a freefall trajectory which will take him close to a direct intercept of the equatorial station.

Almost immediately, player 2 detects the plume and locates where the strike on the station will occur relative to the inertial frame. Waiting for the appropriate sentinel drones to reach near-apoapsis, player 2 commands fifteen of the closest ones to the intercept point to make minor out-of-plane RCS squirts which tweak their inclination and longitude of ascending node in preparation for the coming attack. Each drone is, apart from engines and RCS, a nuclear weapon encased in a lead enclosure along with a rod of very lightweight and expensive metal alloy about the size of a baseball bat.

All the while, early into his freefall coast where delta-v is most important, player 1 has been ejecting solid reaction mass packets off electric railgun engines in order to silently, stealthily change his course enough that the sentinel drones might have to waste all their fuel trying to correct for the change in arrival time.

Back in Earth orbit, the drones are being fed constant telemetry from a huge constellation of sensors on the ground and in orbit as well as a massive Earth-based intelligence system. This intelligence network is now informing the software onboard the relevant fifteen drones that the enemy has likely been ejecting stealth reaction packets and to be aware of a possible large envelope of uncertainty in the final trajectory. The software begins to relax its grip on the delta-v conservation subroutine.

Sure enough, about sixty-five hours into Player 1's coast phase it becomes apparent on the ground that the attacking vessel had been hugging an outer boundary of the sensor network's margin of error for trajectory prediction. Player 2's drones receive the updated information on the attacker's path and immediately thirteen of the engaging-hopeful drones perform a delta-v slam burn, expending much of their fuel with less than a quarter of an orbit to go. Two of the drones stand down, having done the math that they would have run out of fuel.

Presently player 1 begins an immediate hard burn somewhat depressed from prograde, shifting his orbit to engage the station sooner but with the added risk of dangerously skirting the outer layers of the Earth's atmosphere when periapsis is later reached. The crew on board prepare their needler bundles and arm the anti-flash sprayers. Player 2's drones, now less than an hour out from intercept, predict the result of this burn as best as they can and immediately fire their engines in a hail mary to intercept the attacker. They gently orient and take aim.

The minutes count down as the attacker coasts quietly down towards the invisible sentinel wall and the path of the helpless station below.

Combat begins.

Mere moments before the target station is close enough for the attacker to fire the glorified hair-thin uncooked tungsten spaghetti bundles which are his needler projectiles, the sentinels all simultaneously come within range of the attacker. Automatically his ship sprays its silvery anti-flash powder all around it in less than one second.

In each of the sacrificial drones, a thermonuclear weapon has just detonated. The swollen metal cylinders are now proudly shining their hard X-rays into the lead casing of each drone, like thirteen little stars wrapped in cozy blankets. The interiors begin to heat, and emit their own X-rays. At a certain intensity the uber-classified "Q-jacket" layer surrounding the alloy rods energetically snaps open like a glass bottle and admits the X-rays within. The metallic matrix of the rods are instantly ionized, electrons beginning to excitedly jump up and down in even rhythm as a standing wave begins to set up. In less than fifty microseconds, the combined energy of several megatons of TNT is shining in tight beams directly at the glittery cloud of the attacking ship. Even the high-opacity anti-flash cloud is no match for the level of intensity which has been focused on the ship. The X-rays cascade and begin to lower in flavor, downgrading at first to brilliant white light and eventually to an angry infrared as they skip off the glitter.

Almost instantly, the ship is vaporized. The nighttime power grid of several countries underneath the altercation can be seen shutting down in response to the energy hitting the atmosphere. The expanding cloud of dust causes millions in damage to the target station, but everyone aboard survives unscathed to tell the tale.

PLAYER 2 HAS WON. PLEASE INSERT COIN TO PLAY AGAIN.

Link to comment
Share on other sites

54 minutes ago, Wubslin said:

"Boring" is relative. I've spent almost 990 hours watching green space frogs just fall, for years at a time. Besides, if you do want to take a completely realistic approach to it it certainly will not be like attack submarine cat and mouse. In contradiction of some possible mechanics I mentioned earlier in the thread, there is no stealth in space. Everyone is going to be able to see where everyone else is at all times. Have you ever played Children of a Dead Earth?

I have heard of the game and looked into it a while back.

Yea its true boring is relative. However just because it isn't boring for some, doesn't mean its a popular idea (sorry for the double negative). I consider realistic space combat to be rather niche. For every Children of Dead Earth, there are tons of "fake space" combat games that sell more games, and more people consider "fun". I only brought up the submarine cat and mouse example as its something relatable that most people will have a reference with which isn't exactly that popular either. There's always a fan base for such games so don't get the idea that its a bad idea, only that its a niche one compared to the more obvious fake physics pew pew in space games.

To get KSP to have enjoyable realistic space combat just seems beyond the scope of the current game. I think we can agree that KSP probably wont ever include stock "weapon" parts, so implementing game play elements to support as such is questionable at best. Something as simple as what happens when you intercept something to "ram" them at orbital speed? Probably nothing, because the physics wont be able to track these target collision models to such precision as doing as such would be too intensive to general game play, let alone provide enough information for a human pilot them into such a fashion for collision. 

Sure you can slap a short range laser on a Shuttle and blast satellites to bits and work around that physics engine problem, but beyond that "laser part" your just playing KSP to intercept for "docking". Which sounds perfectly fine for a mod, but yea KSP isn't the right platform, let alone the right game engine to make the kind of game your describing. 

 

Just for reference I'm a World of Warships player, where you drive around super large and slow machines to blast players miles away using rounds that take 10+ seconds to land and then wait another 20 seconds to reloadand a full 25% of the match time to turn around. "Slow combat" is the name of the game, when you can only fire a handful a times a match. Or you drive a destroyer into actual engine smoke screening your suicide torps that take over a minute to reload, and you'll die essentially instantly if spotted. Its pretty tactical due to its speed of execution, and I personally like it but I know its not a popular game or genre. 

 

So yea I understand the want for a game that is what you want, but KSP isn't that game mainly due to the work that is required to get even the bare features you want. 

Link to comment
Share on other sites

16 hours ago, Bedazzled said:

I have no interest in a multiplayer KSP. I can't see any need for it, or imagine how it would be a positive experience.

The time and money would be better spent on developing other novel game mechanics, imo.

You should try a Factorio run in coop with good friends, multiplayer doesn't automatically mean "competitive FPS shooter with a toxic community".

 

Link to comment
Share on other sites

11 hours ago, MKI said:

I have heard of the game and looked into it a while back.

Yea its true boring is relative. However just because it isn't boring for some, doesn't mean its a popular idea (sorry for the double negative). I consider realistic space combat to be rather niche. For every Children of Dead Earth, there are tons of "fake space" combat games that sell more games, and more people consider "fun". I only brought up the submarine cat and mouse example as its something relatable that most people will have a reference with which isn't exactly that popular either. There's always a fan base for such games so don't get the idea that its a bad idea, only that its a niche one compared to the more obvious fake physics pew pew in space games.

To get KSP to have enjoyable realistic space combat just seems beyond the scope of the current game. I think we can agree that KSP probably wont ever include stock "weapon" parts, so implementing game play elements to support as such is questionable at best. Something as simple as what happens when you intercept something to "ram" them at orbital speed? Probably nothing, because the physics wont be able to track these target collision models to such precision as doing as such would be too intensive to general game play, let alone provide enough information for a human pilot them into such a fashion for collision. 

Sure you can slap a short range laser on a Shuttle and blast satellites to bits and work around that physics engine problem, but beyond that "laser part" your just playing KSP to intercept for "docking". Which sounds perfectly fine for a mod, but yea KSP isn't the right platform, let alone the right game engine to make the kind of game your describing. 

 

Just for reference I'm a World of Warships player, where you drive around super large and slow machines to blast players miles away using rounds that take 10+ seconds to land and then wait another 20 seconds to reloadand a full 25% of the match time to turn around. "Slow combat" is the name of the game, when you can only fire a handful a times a match. Or you drive a destroyer into actual engine smoke screening your suicide torps that take over a minute to reload, and you'll die essentially instantly if spotted. Its pretty tactical due to its speed of execution, and I personally like it but I know its not a popular game or genre. 

 

So yea I understand the want for a game that is what you want, but KSP isn't that game mainly due to the work that is required to get even the bare features you want. 

Combat aside, if multiplayer KSP 2 doesn't have a revamped engine that allows for orbital collisions it would be nothing short of a complete travesty - if only because it meant orbital hijinks wouldn't be possible.

Also, saying "Y can't exist because X already does and it's popular, and therefore Y would make way less money" is not a valid argument. It's obvious that KSP has a vibrant, active community across the forums, reddit and youtube. These people constitute a market who are almost certain to find the idea at least somewhat appealing, and as it stands there are ZERO games that cater to this area.

To use stupid business lingo, it's "blue ocean". If intercept games decides to implement any provisions for space combat in their game, they will have to compete with exactly no one, and no matter how crusty the implementation is it will be the best one around because it will be the only one around .

Final note, world of warships player who hasn't ever even touched CoADE, I wouldn't be so quick to put down other genres of game just because you're not into them. What if I was a huge Battlefield fan who told you that no one would ever want to play a singleplayer game where you fly semi-realistic spaceships around?

Different genres appeal to different people. And most of us here love spaceflight. Orbital combat would require entire other dimensions of orbital spaceflight mastery, and would be a competitive challenge to boot. I think a lot of people on here sour to the idea because it would somehow be made arcade-y or else impinge on normal gameplay. It wouldn't.

Link to comment
Share on other sites

16 hours ago, Wubslin said:

Close air support? You can play War Thunder or DCS World or something if you want that. The market is saturated with generic multiplayer dogfighting games. Do you know what there isn't a single one of? Games that allow multiplayer orbital combat. I NEED the ability to play a non-turn based version Children of a Dead Earth with another actual human and god help me I will do whatever it takes to twist and mutilate the game with 9,000 mods if I have to, just to see what it's like. We are so starved of the experience of multiplayer combat games featuring orbital mechanics. The closest game I can even think of that has done anything like this was the game Spacewar!, and that came out in 1962. I can understand and fully appreciate a KSP 2 where the main focus of the game is collaboration or friendly competition in space, but make no mistake. No matter how peace-oriented the game is, the moment my friends and I get it we're going to get into pro/retrograde orbits of Kerbin and try to smoke each other with thick clouds of TD-06 decouplers in a big jousting match just to see if we can, and there's no one who can tell us otherwise. It's going to be awesome.

 

1: War thunder doesn't let me design my own double-body AN-225 Gunship.

2: I would kill for CODE-esqe combat in real time. The decoupler idea is really sick. 

Link to comment
Share on other sites

 

5 hours ago, Wubslin said:

Combat aside, if multiplayer KSP 2 doesn't have a revamped engine that allows for orbital collisions it would be nothing short of a complete travesty - if only because it meant orbital hijinks wouldn't be possible.

I don't understand how multiplayer in KSP "would be a complete travesty without a revamped engine", nor do I understand how that would even be considered to be put on the table. That's essentially building 2 games if it requires 2 different engines. Why not just improve one? :confused: Heck, idk what approach they are even taking. Jumping to "complete travesty" seems like a pretty big jump and a "quick put down". 

 

5 hours ago, Wubslin said:

Y can't exist because X already does and it's popular, and therefore Y would make way less money

I didn't say can't exist. I only point out the current situation where there are way more games oriented to a casual player... because there are more casual players, which in turns means more money out there. Yea KSP's community is big and vibrant, but relative to what?

 

5 hours ago, Wubslin said:

Final note, world of warships player who hasn't ever even touched CoADE, I wouldn't be so quick to put down other genres of game just because you're not into them. What if I was a huge Battlefield fan who told you that no one would ever want to play a singleplayer game where you fly semi-realistic spaceships around?

Er are you saying I'm making a straw man argument? I'm not. I more wanted to try to relate our common experience with "niche" games to the fact they are niche haha. I understand you love an orbital combat concept, as I like a super tactical and slow boat shooter. Sorry you feel like I'm "putting down your game", Its not that the game is bad, its just the fact if you have to ask about it, you probably already know odds are most people don't know about it right, and thus it is "niche". Niche isn't even bad, the game can be amazing, but it can be agreed that it isn't mainstream for whatever reason.

 

5 hours ago, Wubslin said:

To use stupid business lingo, it's "blue ocean". If intercept games decides to implement any provisions for space combat in their game, they will have to compete with exactly no one, and no matter how crusty the implementation is it will be the best one around because it will be the only one around .

As someone who works in software, just because there is an "open market" doesn't mean your game/software instantly becomes a success and sells because there is no competition. There's no competition for a reason. You can always say "that reasons sux", but to get anyone to invest into what your essentially calling a "void of a genre" is a big financial risk. At this point KSP 2 has been pushed back years pivoting to support a whole new genre is definitely no longer in the cards. Again, this doesn't mean I disagree with the game concept as "being a bad idea", but rather you have to convince those with the $ its a great idea, so they fund such a thing.

 

5 hours ago, Wubslin said:

Different genres appeal to different people. And most of us here love spaceflight. Orbital combat would require entire other dimensions of orbital spaceflight mastery, and would be a competitive challenge to boot. I think a lot of people on here sour to the idea because it would somehow be made arcade-y or else impinge on normal gameplay. It wouldn't.

I don't sour on the idea itself, I don't think anyone is "attacking" or "putting down" the idea that realistic orbital combat is a bad idea. I more want to point out the fact focusing on that concept probably wont ever happen for KSP due to other reasons related to the games development and overall target audience. Unfortunately, I do not believe "orbital space combat players" will or are a target audience for KSP 2. 

I originally stated that "putting weapons in space" in stock already is pretty much a blocker, since KSP wants to be used as a STEM teaching tool, and making it about war shouldn't be part of that message. Obviously mods can change this (which is totally what they are for), but mods have limits. 

Those limits are the engine isn't designed to support such complex orbital game play at high velocities, so focusing on such is a pipe dream. The KSP team wont move to another engine just to support a few scenarios, such as super fast and accurate collision detection. Its too busy trying to get all the other core game-play concepts going.

Its not that having a better more accurate engine "impinges on gameplay". Rather I'm not sure if practically possible to have it both ways where the game runs fine, and allows you to blast your enemies at orbital velocities with enough precision. This has nothing to do with "wants", and more to do with what the game just runs on. If KSP was built using its own advance engine that can calculate all that heavy duty math without melting my computer and letting me have the same game play elements from the first game and all the new stuff they announced, then yea maybe, but thats a heck of a lot more work.

 

So don't think I'm "putting you down", I'm more just trying to make the argument that an "orbital space combat game" just wont be KSP 2 for a number of core reasons. This doesn't mean the idea is inherently bad, its just that doesn't seem to line up enough with KSP 2's core vision and what they are building. Sure mods can always plug in and maybe get you close? But for a stock experience, I can't see that happening.

 

Who knows, if this game is popular enough, they might make a "Kerbal Space Command" sequel that is orbital space combat. In the mean time enjoy what ya got coming :)

 

 

Edited by MKI
Link to comment
Share on other sites

1 hour ago, MKI said:

I don't understand how multiplayer in KSP "would be a complete travesty without a revamped engine", nor do I understand how that would even be considered to be put on the table. That's essentially building 2 games if it requires 2 different engines. Why not just improve one? :confused: Heck, idk what approach they are even taking. Jumping to "complete travesty" seems like a pretty big jump and a "quick put down". 

I don't mean completely overhauled, I just mean modified so that when ships are close enough that they load in for each other they can actually collide with each other at orbital speeds instead of powerpointing through each other in two frames. It's not that hard a concept to think up a solution for. And for your other points, again the idea of competitive actions or combat would be completely optional. I'd be happy enough with a KSP 2 where orbital collisions are possible. I'd be *overjoyed* if the game also was moddable enough that someone with my kind of mindset could overhaul the game into a fighting only game. But at the same time, I don't think some innate stock capability for combat or other offensive action would immediately transform the entire game into some kind of charlie foxtrot. By the way, you're misinterpreting what I'm saying. Engine changes wouldn't make KSP 2 into a war game, gameplay feature implementation would. And as I've said before, my ideal KSP is one where the primary player interactions are trading and flying together, but where violence and griefing are possible. Like if billy bob and his four friends claimed ownership of Tylo orbital space, and were demanding a flat 200 credit fee for using the moon to gravity brake into orbit of Jool, maybe it would make for interesting gameplay if they had the authority to laser you with a few probodobodynes as they swung by when you refused to pay up. It would be up to everyone else in the system to decide whether that is fair, and then up to those people to take action about it, diplomatic or otherwise. I know the game as it's going to come out will be nothing like this at all, but I can still dream about it.

And for the record, KSP is not a STEM teaching tool. It's a video game that gets used as a STEM teaching tool.

Edited by Wubslin
Link to comment
Share on other sites

3 hours ago, Wubslin said:

I just mean modified so that when ships are close enough that they load in for each other they can actually collide with each other at orbital speeds instead of powerpointing through each other in two frames.

Ah see that's actually a HUGE change!

The issue comes down to how computers compute collisions. To keep things simple, computers don't and wont calculate collisions at orbital speeds because it takes up too much computation time and memory with the system it currently has (which is simple collision physics provided by Unity). It just checks every so often if things are inside of each other and calls it a day. This is how Mario did it, and this is basically how Unity does it. It works until you start going so fast your calculations don't run fast enough and you just magically "fly thru" your target. You could just calculate more, but that still stops working once speed increases, so its unpractical, and expensive. 

A "cheaper" method if you wanted to build your own custom collision detection on-top of KSP's current engine, to work at orbital velocities means adding onto the existing physics collision logic where KSP would have to "estimate" if you are going to his your target at orbital velocities. This would be done by calculating both trajectories and seeing if they "collide" with increasingly large and more accurate calculations as the target comes into current physics range, at which point the game not only has to load all the parts and physic, but it also now has to extrapolate the target from what was originally a single dot in 3D space, to a 3D dimensional collision model to estimate against and continue processing this "collision" estimate until impact, again increasing with decreasing distance. This would only decrease the overall computation costs by a specific factor the further you are away (as your calculations can be more and more off), which will actually decrease as you get closer, so the game would run worse and worse as you get closer to your target, as the computer needs to keep checking from its original estimates to become accurate enough to overcome that "gap" of flying thru the target. Obviously if the estimations find your sufficiently not going to hit your target they would stop, and the game would run normally, however this assumes they also didn't change their orbit/velocities, in which case the entire complex process starts again.

Since I haven't played CoADE I can't say exactly how they do it, but here's my guess about how they do it. All orbital craft are actually just points in space and half the collision detection is just "faked", where it looks like something is being calculated from orbital velocities, when really its either localized (like Kerbal), or estimated from a point in space and calling it "homing" as it hits your target. (maybe with some RNG sprinkled in, so things look less "straight") IF you can see stuff "hitting my ship" in the game, its probably faked, as if it was being calculated you actually wouldn't see anything because it moves to fast. (try to imagine seeing a bullet, and now increase that so it moves at 4 miles a second)

The "homing" aspect would require a whole new element to the game specifically to calculate orbital collisions, and thus a whole new problem space to calculate object interactions in this specific scenario case. (IE more work) If its important to the game it seems worth the effort to build this sort of "estimated-fake-homing", but for stock KSP it isn't since orbital collisions are already so insanely rare it basically never happens in real life, which has no calculation limits. (The universe must be running on a few 3090's ;P ) So spending time on this specific functionally means less for other features, and the stock game doesn't really have a use-case for such a feature either.

 

Mind you, these problems aren't just something "KSP can fix", its a computation problem that underpins the problem itself that essentially ends up in trying to implement complex orbital equations into what is essentially calculus, which computers already struggle with due to their limited floating point math (0.1 + 0.2 != 0.3 most of the time LOL). This is the same reason why your orbital number is flipping around all the time in the few m/s digits when in orbit according to mechjeb/KE. You might think computers are fast, and your right! But this problem is actually so complex in terms of the size of calculations, even a super computer would be challenged for a non trivial case, let alone provide you with a 60fps enjoyable experience. (I realized a super computer could work here if the engine was setup to correctly fake it haha) 

KSP already does a number of "hacks" to keep your experience enjoyable, some of which are very noticeable, like the render distance limitation, and other's less so, like rendering Planet as an estimated distance and using scaling tricks, rather than actually calculating their full distances. Or straight up does not consider parts of actual physics, like n-body calculations to keep things simple. 

 

3 hours ago, Wubslin said:

And for the record, KSP is not a STEM teaching tool. It's a video game that gets used as a STEM teaching tool.

Yes semantically you are correct, but the underlying point still holds so I'm not sure what the point is here.

 

If these sorts of problems interest you I suggest checking out more math stuff related to calculating orbits. I've only dabbled a little into it myself, but I can firmly say that  the reason why in real life everything is almost estimated is because real life is even more complex than anything in the game, and even though computers are running at the speed of light, such a speed is actually hilariously slow compared to how fast and complex real orbits work when it comes to heavy calculations like these.

Edited by MKI
Link to comment
Share on other sites

44 minutes ago, MKI said:

Ah see that's actually a HUGE change!

The issue comes down to how computers compute collisions. To keep things simple, computers don't and wont calculate collisions at orbital speeds because it takes up too much computation time and memory with the system it currently has (which is simple collision physics provided by Unity). It just checks every so often if things are inside of each other and calls it a day. This is how Mario did it, and this is basically how Unity does it. It works until you start going so fast your calculations don't run fast enough and you just magically "fly thru" your target. You could just calculate more, but that still stops working once speed increases, so its unpractical, and expensive. 

A "cheaper" method if you wanted to build your own custom collision detection on-top of KSP's current engine, to work at orbital velocities means adding onto the existing physics collision logic where KSP would have to "estimate" if you are going to his your target at orbital velocities. This would be done by calculating both trajectories and seeing if they "collide" with increasingly large and more accurate calculations as the target comes into current physics range, at which point the game not only has to load all the parts and physic, but it also now has to extrapolate the target from what was originally a single dot in 3D space, to a 3D dimensional collision model to estimate against and continue processing this "collision" estimate until impact, again increasing with decreasing distance. This would only decrease the overall computation costs by a specific factor the further you are away (as your calculations can be more and more off), which will actually decrease as you get closer, so the game would run worse and worse as you get closer to your target, as the computer needs to keep checking from its original estimates to become accurate enough to overcome that "gap" of flying thru the target. Obviously if the estimations find your sufficiently not going to hit your target they would stop, and the game would run normally, however this assumes they also didn't change their orbit/velocities, in which case the entire complex process starts again.

Since I haven't played CoADE I can't say exactly how they do it, but here's my guess about how they do it. All orbital craft are actually just points in space and half the collision detection is just "faked", where it looks like something is being calculated from orbital velocities, when really its either localized (like Kerbal), or estimated from a point in space and calling it "homing" as it hits your target. (maybe with some RNG sprinkled in, so things look less "straight") IF you can see stuff "hitting my ship" in the game, its probably faked, as if it was being calculated you actually wouldn't see anything because it moves to fast. (try to imagine seeing a bullet, and now increase that so it moves at 4 miles a second)

The "homing" aspect would require a whole new element to the game specifically to calculate orbital collisions, and thus a whole new problem space to calculate object interactions in this specific scenario case. (IE more work) If its important to the game it seems worth the effort to build this sort of "estimated-fake-homing", but for stock KSP it isn't since orbital collisions are already so insanely rare it basically never happens in real life, which has no calculation limits. (The universe must be running on a few 3090's ;P ) So spending time on this specific functionally means less for other features, and the stock game doesn't really have a use-case for such a feature either.

 

Mind you, these problems aren't just something "KSP can fix", its a computation problem that underpins the problem itself that essentially ends up in trying to implement complex orbital equations into what is essentially calculus, which computers already struggle with due to their limited floating point math (0.1 + 0.2 != 0.3 most of the time LOL). This is the same reason why your orbital number is flipping around all the time in the few m/s digits when in orbit according to mechjeb/KE. You might think computers are fast, and your right! But this problem is actually so complex in terms of the size of calculations, even a super computer would be challenged for a non trivial case, let alone provide you with a 60fps enjoyable experience.

KSP already does a number of "hacks" to keep your experience enjoyable, some of which are very noticeable, like the render distance limitation, and other's less so, like rendering Planet as an estimated distance and using scaling tricks, rather than actually calculating their full distances. Or straight up does not consider parts of actual physics, like n-body calculations to keep things simple. 

 

Yes semantically you are correct, but the underlying point still holds so I'm not sure what the point is here.

 

If these sorts of problems interest you I suggest checking out more math stuff related to calculating orbits. I've only dabbled a little into it myself, but I can firmly say that  the reason why in real life everything is almost estimated is because real life is even more complex than anything in the game, and even though computers are running at the speed of light, such a speed is actually hilariously slow compared to how fast and complex real orbits work when it comes to heavy calculations like these.

Look, there are easier ways to compute collisions. Consider a craft, ship A, in equatorial 70 km by 70 km orbit of Kerbin, and another craft, ship B, in a polar orbit of the same dimensions. The craft in the polar orbit makes a burn to line up its intercept to "0.0 km", and the craft in the equatorial orbit does nothing.  When the two craft come within a certain distance of each other (a distance which need not be physics range but a distance which is proportional to relative speed), the game halts its normal procedure for craft motion. At this given distance, the crafts are immediately "frozen". Ship A's motion relative to ship B is drawn in a straight line out to some distance, and Ship B's motion relative to A is drawn out to some distance as well. The outline of each ship is extended out in these directions as a boundary which represents where collision is able to occur, and this "shadow" is treated as an instantaneous boundary of collision. For a brief moment of close approach, both ships could be considered as thin, solid extrusions reaching out some distance which is far greater than the dimension of the "collision bubble" and in existence simultaneously. If the slender extrusions intersect each other, the game will count this as a collision at whatever the relative speed of the two craft is. Any kind of destruction or other damage physics can be secondary to this process and applied after the fact. The span of time in which the two ships would have crossed their respective collision bubbles using traditional motion should be so vanishingly small that no player or automatic control system should be able to affect its outcome - to allow the system to remain a decent approximation of a collision. This system might not kick in until relative speeds exceed a few hundred meters per second.

g9Zjgxw.png

This is just one way of coming up with a mechanism for orbital collision. The way it's done doesn't have to be super crazy, I would just like it to be there.

And when I say "homing systems", I don't mean anything so crazy that it accounts for orbital mechanics or anything. I just mean the kind of short range control theory stuff that normal homing missiles are capable of. And that math isn't really that complicated. The video below is a really good example of what I mean. All the orbital stuff should be up to the player's own manual flying.

 

Link to comment
Share on other sites

22 minutes ago, Wubslin said:

For a brief moment of close approach, both ships could be considered as thin, solid extrusions reaching out some distance which is far greater than the dimension of the "collision bubble" and in existence simultaneously. If the slender extrusions intersect each other, the game will count this as a collision at whatever the relative speed of the two craft is

The cylinder approach doesn't consider if the space craft arrive at the same point at different times. So if you cross paths with a ship nearby you explode magically due to the "collision cylinder" crossing paths within the physics distance. At that point you can't use a cylinder and must calculate where you are within the cylinder relative to time, which brings us right back to where we started, where your calculating multiple points along an orbital path thru time in relation to whatever else is in the physics space, and you can miss if your points are to far apart relative to how often you check and how fast your moving.

You could say to "fake it" and kick this collision detection in at a certain point, but when and where? That would add another performance hit to do all these extra calculations, and assume you don't move after. So you either continue the checks as you make final maneuvers, or you end up back with the cylinder checks "missing" and you flying thru your target again. 

 

Edited by MKI
added missing part
Link to comment
Share on other sites

1 minute ago, MKI said:

The cylinder approach doesn't consider if the space craft arrive at the same point at different times. So if you cross paths with a ship nearby you explode magically due to the "collision cylinder" crossing paths within the physics distance. At that point you can't use a cylinder and must calculate where you are within the cylinder relative to time, which brings us right back to where we started, where your calculating multiple points along an orbital path thru time in relation to whatever else is in the physics space, and you can miss if your points are to far apart relative to how often you check and how fast your moving.

 

Yes, but for a sufficiently small collision bubble this problem is minimized. The distance at which instantaneous hitscan rules kick in need be only the distance between the normal spacecraft "clicks" in motion. And that distance is proportional to relative speed.

Link to comment
Share on other sites

9 minutes ago, Wubslin said:

Yes, but for a sufficiently small collision bubble this problem is minimized. The distance at which instantaneous hitscan rules kick in need be only the distance between the normal spacecraft "clicks" in motion. And that distance is proportional to relative speed.

This sounds like a solution that would require a whole new level of collision detection. Which goes back to what I said about this having little value outside of ramming other stuff, intentionally or unintentionally and would negatively impact the main performance use-case of physics orbital calculations, which is docking. 

At a certain point this isn't even worth it as at slower speeds where you can actually see what your doing and existing physics checks works just fine. So it would be of use only at higher speeds which wouldn't mean its a "small collision bubble" at all due to it being only practical at high speeds, which brings us back to the original problem. 

 

Edited by MKI
Link to comment
Share on other sites

2 minutes ago, MKI said:

This sounds like a really solution that would require a whole new level of collision detection. Which goes back to what I said about this having little value outside of ramming other stuff, intentionally or unintentionally and would negatively impact the main performance use-case of physics orbital calculations, which is docking. 

At a certain point this isn't even worth it as at slower speeds where you can actually see what your doing and existing physics checks works just fine. So it would be of use only at higher speeds which wouldn't mean its a "small collision bubble" at all due to it being only practical at high speeds, which brings us back to the original problem. 

 

Scott Manley did a video way back about orbital collisions. He was trying for a perfect retrograde collision, and noted that the physics stepping was on the order of 100 meters for double Kerbin orbital speeds. That isn't big, in my opinion. The collision shadows wouldn't be likely to hit anything but each other. And besides, that's just one way of computing collisions. Why not just slow physics down to make the steps closer together when craft get near each other at high speeds? There are lots of ways to solve this problem, I don't see why you wouldn't want Intercept Games devoting at least a little effort towards the problem of high speed collisions in a multiplayer space game.

 

Link to comment
Share on other sites

1 minute ago, Wubslin said:

Why not just slow physics down to make the steps closer together when craft get near each other at high speeds? There are lots of ways to solve this problem, I don't see why you wouldn't want Intercept Games devoting at least a little effort towards the problem of high speed collisions in a multiplayer space game.

I never said I didn't want this. Its a feature, who wouldn't want more features? My only goal is to point out how it isn't "an easy fix" as you originally described. However, unlike other features this also does increase computation costs so it isn't even "free" to solve. Its a very fine balance for a very specific feature, which makes it hard to implement. 

Do it wrong and everyone's game performance suffers. Do it right and you get more in orbit explosions, but doing it right is hard. Doing hard stuff with pitfalls just increases risk to everyone, so it has to be worth the risk. In a perfect world everyone gets what they want, but this isn't a perfect world. 

 

 

Link to comment
Share on other sites

Should we address the problem of the fact that people are just going to be able to launch ICBMs at our bases and so therefore if our VAB gets hit we can't strike back unless we have enough funds to repair. Are we going to need a governmental system established by players or a bot government or can we declare independence from the government in the game and then have relations or something, honestly this kinda freaked me out when I thought of the complications of political space entities?!?!?

 

Fortiter Ad Astra

Edited by BigStar Aerospace
Link to comment
Share on other sites

×
×
  • Create New...