Planetary Orbits?

I want the game to focused on the player versus player interaction, I do not want to be bothered by in-game mechanics that are not a result from my or my opponents actions. The only thing a mechanic like that would do is "Now you have 3 minutes of day and extra energy income. Now you have 3 minutes of night and reduce energy income". It is not something I want to deal with when I am playing against another player, and since he would be subjected to exactly the same effects, what would it add to the gameplay?

Higher lever players would optimize their build so they have the exact number of solar collectors they need before moving on to a more dependable energy source, making the whole mechanic unnecessarily complex because it does not add to the PvP interaction.

The map shown in the .gif would be very very cool, and also very very possible with in game editors and still makes the point of true to life orbital mechanics unnecessary. Of course I guess i can't say for 100% that will be in the editor, but I think it's fair to say that adjusting the celestial objects orbits would be a rudimentary procedure for us to be able to make custom maps. And with the thought of custom maps, simple orbital dynamics means the average player could easily make cool systems to fight in, without needing a degree in mathematics.

Technically that would also include the map terrain. Oceans and mountains are game mechanics that aren't caused by another player.

The one reason I think orbits would help the game is that empty space is really just not that interesting. I think it was a good decision to not have any fighting in space because a battle on the surface of a planet will always be more interesting than a fight in an empty void.

But once you say that fights will only happen around planets then the solar system just becomes a glorified menu where you zoom in and out of the maps you are fighting on. Having people deal with the timing of orbits and day / night cycles will be an added layer of complexity that people could safely ignore if they wanted to (there will be some long trips across space, or some ill timed night attacks), but a clever player could line up orbits for shorter trips or take advantage of the time of day.

But more importantly it will give people a reason to look at the entire solar system, other than just zooming into the next map.

Just on the maths, there's no point in real orbital mechanics when simpler mechanics will give the same effects on gameplay. Simple 2-body keplerian motion will suffice to provide a dynamic solar system while patched conics can easily provide the basic orbital transfers units, asteroids and planetary bodies will require (and since it's a segmented method, can probably be more easily adapted to handling the different object classes that are moving). Why add more complicated maths and numerical simulation when simpler analytical methods provided the same outcome gameplay wise.

I think orbits with a small period (at least half an hour ranging up to 5 or six hours) for moons and planets would add much to the game. It would create natural points where it would be easier or harder to reach or attack your enemy. For example an orbiting moon would only be able to bombard your enemies base while in view of the enemy base. Another example would be increasing or decreasing the time an asteroid would take to reach your planet and cause awesome.

The important thing is that the orbit mechanic should be made to work in such a way as to not significantly disadvantage newbies who don't yet understand the finer points of PA gameplay. It should provide however an advantage to veterans that should both understand how the mechanic affects gameplay and be able to take advantage of it.



Edit: Spelling, a few clarifications.

There's a reason why the "Fox only. No items. Final Destination." attitude is mercilessly mocked.

Patched conics should be used.

While n-body is possible, it's impractical. Firstly, with patched conics, you can calculate an entire orbit in one go, and you'll know what's going to happen almost immediately. this makes saving a planet's orbital position much simpler, but also, it means we can do a variety of other things; for example, when looking at a recording like you could in SupCom, it'll need to keep track of and calculate much less than if using an n-body solver. You can infinitely slow it down or speed it up with no performance or accuracy change, and in general it's just a lot quicker and more efficient than n-body.

With n-body, you need to pretty much calculate the force of every single object with a considerable mass to every other object. That's slow, and you need to do it every frame. It may be alright when playing a small game, but imagine if playing one of those large battles that was mentioned by a game dev; 40 or more players, all in a massive playing field, each of them sending out dozens at a time. You need to calculate an awful lot, there.

Edit:

Personally, if we really just get a bunch of different dimensional playing fields with things like planetary collision just being a parameter instead of anything physics based, I will have felt lied to. It really needs to look like the trailer made it out, which is actual round planets.

Edit Again: I just read people saying that they should have no orbits calculated whatsoever and be stationary in space... Yeah, no. It's not even very hard to implement this. If that happens to this game, I will be extraordinarily disappointed.

No. No orbits.

Its sounds cool, but then so do flying cars, implementation and some foresight makes you realize the idea is too flawed.

The problem I have always seen is firstly that orbits create overbalance or sevre reduction in map variation.

secondly I see traveling between planets, especially in the cool "Using gravitational wells" in the conept video as being something you would need a pretty complex physics engine that is beyond the scope of this games budget.

I would much prefer to see a form of "Invisible railroads/train tracks" between planets. Its simple and allows for routes to destinations of certain strategic planets to pass by other planets allowing for the possibility of combat scenarios in the form of orbital defences. Basically that if you want to make a run for high resource planet c then you have to pass through planet B's orbit and the enemy there can take pot shots with certain weapons.

Also I think this is an excellent way to squeeze basic space combat in. Eg; planet a and B both have a small orbital fleet, they both send them to each others planet but the "rail tracks" of space flight pass by each other, as the fleets approach each other broadsides are unleashed etc.

Wat.

Good thing we sent astronauts to the moon with a computer that could calculate trajectories, which incidentally had less computing power than an iPod Nano. They didn't need a physics engine to do this.

There's no "physics" involved. It's math, and your computer can solve a bajillion math problems a second.

http://en.wikipedia.org/wiki/Patched_Co ... roximation

Given that the 'gravity' of worlds really doesn't need to exist beyond the surface, and all the orbital mechanics would be simple and known to the system, calculating a trajectory would have no impact on performance whatsoever.

I don't think anyone seriously doubts we have the compute hardware to make it happen. The problem lies with the users' wetware.

A good game is intuitive, easily comprehensible at the beginning, but complicated and deep enough to reward skillful play once you learn the ropes.
Anyone can make a game complex, it's the other stuff that takes real insight and effort.

Go is probably the quintessential example, ridiculously simple, but devilishly deep.

"facepalm"

Right firstly all planets in a map would have a starting point of there orbit, meaning that with tech progression you can reach a level of "interplanetary tech" in a pretty calculated time, position of the planets at that point means either every map must be symmetrical with equal opportunities for balance. Thus if your really looking at orbits that are not completely stupid map variation goes down the plug hole.
Imagine if Sins of a Solar Empire forced map creation to be correct according to the theories of gravity?

Seriously though

There is so much wrong in this one sentence it staggers my mind. In fact somewhere a coder is crying.

and you talk about trajectory like your firing a gun, Space travel revolves around using the gravity of planets to slingshot to desired locations, and it looks cool. Throwing custom solar system designs and the idea that your computer automatically calculates a "bajillion" calculations and solves everything and then doesn't look stupid is insane.

its just...... ARGGGH!

God dammit and I bet if this went to a poll people would want it too.

I don't get what your saying at all Rab777. Physics simulation on computers has been around for years, it was in TA.. A 2-D version of PA's potential orbital trajectories simulation is out for smartphones called Angry Birds Space. Ever played that?

I think that orbits would make the game a little more strategic, and a whole lot more more fun.
If the planets didn't move, it would just be boring as everything would be in a predictable location all the time. If the planets had complex orbits that reflected physics in real life to a high degree, it would also be boring as there would too many things to think about when travelling between planets (ie alignment burns, high or low powered transfer, etc)
If simple orbits (keplerian motion and patched conics) were implemented along with a short orbital period (2-5 hours), it would prove to be a fine balance between the two. Players would have to think about when it is best to move across planets, if they where in danger from attack, and if there were resources nearby. If planets didn't move, one player may always be closer to a resource, permanently giving them an unfair luck-of-the-draw advantage. If orbits are too complex, players may accidentally throw themselves out of the system when moving between planets.

Just my two cents.

Its not that the technology doesn't exist.

I highly doubt a correct orbital physics simulation on the 40 man scale were talking about would be
A-Within budget considering priorities
B-Fun
C-Look good

Actually I think this is one topic that Uber really need to jump in on to explain firstly what they had in mind originally and if it was static planets then explain to everyone what that would mean if implemented because I'm sure it hasn't been thought through fully by a lot of people.

I mean if they were to add it as the 4 stretch goal for example I think quite a few people might actually remove pledges as they assumed it would be in automatically.

This is something I'll fight for until proved otherwise. I'm convinced its a very bad move.

People here seem to have different assumptions about the rest of the game. For example someone said that a short orbit would be 3-5 hours long, to me that is forever. I feel we have another meeting of the SP crowd who want to play massive games over many days and the MP crowd who want balanced games of less than an hour.

We're not going to be able to cater to everyone with a rigid system. People don't even know yet what will work, we can theorise but we should definitely try other systems to see if they work. Also few people will play purely one game type and there are many ways to play both MP and SP games. For example if you are able to set some things to not be affected by other things then maps could be made with a set of static planets.

This patched conic approximation system sounds nice from a assimilationist perspective because you get most of real orbital mechanics. In practise it sounds very complex to use. There are a lot of things for the player to take into account.

• Launch Windows There will always be small launch windows to take advantage of but there would be too many to communicate to the player. Would there be some UI for upcoming launches? There are probably too many.
• Multiple Paths If you want to go from one planet to the other there are likely to be multiple paths. Some might be more dangerous but take less time. Again how do you communicate all these and let people choose one?

Basically if the system moves at a noticeable speed then paths between planets will be continually created and destroyed. Players don't have much control over this, I doubt it would be intuitive and I don't see how to convey the information to them.

Rockets could have ridiculously powerful engines and just brute-force their way between planets. Their comparatively high speed would make the effect of gravity negligible and it is easy for players to understand that close planets implies less travel time.

I'm sorry, but you are wrong. Firstly, patched conics can calculate a trajectory instantly. That's not a bajillion calculations each frame. It's calculated only rarely, like when an SOI switch occurs.

And, really, I can personally say as someone who knows a decent amount about space travel that firing a gun and planetary gravity are really the same thing, and can be calculated through similar means. In fact, I believe a bullet could be calculated using patched conics as well as a spacecraft.

I mean, c'mon. It's not intensive at all to calculate it. NASA did it on bloody computers that were thousands of times weaker than our phones. And, also, to those who complain that patched conics are too inaccurate, NASA used it to get to the bloody moon as using a n-body solver would be too slow. Good enough for NASA, good enough for me.

You keep bringing up NASA, but the budget of NASA around the time of the moon landings was 21,376 million a year in today's money. Uber will probably only have ~1.5 million (0.007%). NASA could spend money solving problems Uber just can't.

Also, I want as much calculation power as possible to be reserved for the simulation of MORE UNITS and MORE PROJECTILES. Not for planetary simulations.

Yes, but we have technology many time cheaper than NASA does but many times nmore powerful. And we know all the algorithms they used.

And it doesn't take and bloody calculation power anyways, it's just one calculation. That's all that's needed.

You do realise you wont even notice your cpu's is busy calculate those paths? My calculator can even do it...and iam pretty sure my calculator is ALOT slower than my quadcore pc.

I think the real problem lies with the fact that a fully simulated system may produce unexpected results. Map makers would have to be really carefull not to make a system that destroys itself in time. So pre set paths for all planets and asteroids would work best. They would stick to this path until a player manipulates their path with those city sized engines.

You've obviously never written code.

You seem to overestimate the processing power of your pc and underestimate the complexity of moving planets.

Usage of moving planets is also not as easy as you make it out to be:
- You have to do it every simulation tick (~30 times per second).
- You have to do more calculations when doing things like launching missiles/units to other planets.
- You have to alter your projectile physics system to take the movement of planets into account. There are several ways to do this, I'll grant you that, but all of them result in a more complex physics system than you'd have if planets (and everything on them) is stationary.

When programming a rendering system, people generally use occlusion culling. This is a method used to see which polygons actually have to be rendered on screen. Polygons that will not show up will be dropped before being sent to the GPU.
If objects are always stationary, you can use a quite fast, grid-based occlusion culling algorithm. If object are moving/rotation you have to use a more complex (and thus slower) algorithm.

All of these things will impact the game's CPU usage, so I'll take more units over moving planets, thank you.


In short, there's a lot of things that will need to be altered if planets were to move. The calculation of the actual trajectory/position of the planet is the easiest of them all.

I would tend to agree, basic simulation would probably be less a problem than making use of it in terms of predictability and user interaction. Some pre set orbits would be easier on the interaction side and still leave room for bit of timing and planning for transfers or attacks.