Advanced search

[ New messages · Forum rules · Members ]
Page 1 of 11
Forum » SpaceEngine » Gameplay Discussions » Autonomous Empires: The multiplayer soloution. (How to have meaningful multiplayer in an infinite universe)
Autonomous Empires: The multiplayer soloution.
PacificMaelstromDate: Sunday, 26.10.2014, 01:14 | Message # 1
Group: Newbies
United States
Messages: 4
Status: Offline
I know that one issue with multiplayer in a game like this is that individual players will quickly disperse and never see each other again. Action will only occur around well known places like earth, but even that will be limited. In the end, the universe will seem mostly empty and players will get bored. But I believe there is a way in which this problem can be overcome, and a game can be produced that is nothing like any other space game ever before.

I call it the Autonomous Empire feature. The inspiration comes from the basic premise of the Fermi paradox. We expect that intelligent civilizations older than our own should have colonized earth by now (if they were out there). Ive heard it said that even at subluminal speeds, an alien civilization could occupy an entire galaxy in only tens of millions of years. But if warp travel is possible (as in SE) then the process can occur much faster. It was mentioned elsewhere that players will have robots working for them (to build ships and stuff) but why not expand that concept? It could look like this:

-Players start as an individual with a single ship and a few robots.

-They gather materials and build more robots and more ships.

-They put the new robots on the new ships and send them to the next star with instructions to:
- -Gather more resources
- -Build more robots
- -Build more ships
- -Send out those ships to repeat the process at additional stars.

-Players would control what kind of ships and facilities their robots build. They could focus on expanding quickly or fortifying each system extensively. They would decide how many robots would stay at each system to build facilities, and how many would be sent off.

-They would also control what kinds of ships are being built (warships, cargo, transport etc)

-Their personal avatar could stay in the capital and issue commands and/or tour the newly established robotic facilities and take personal control of any individual unit.

-The more systems under their control, the more new systems they can colonize, and the more machines and resources they can obtain. Expansion is virtually EXPONENTIAL, with the player's sphere of influence increasing rapidly. For example, 1 system becomes 2, 2 becomes 4, 4 becomes 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192.... In a relatively short time, players will amass control of a galacticly significant number of star systems.

-Before long, players will come into contact with other, similarly expanding empires of other players. The robots will alert the player of the situation. The two sides can choose to make peace and form an alliance, or they can attack eachother
- -Players would direct their robots to amass at the battlefront and instruct their facilities to produce warmachines
- -Conflicts over contested systems would be decided by:
- -The size of the conflicting forces
- -The type of units present
- -An element of randomness
- -If a player takes direct control of managing the strategy of his/her units, odds of victory could be increased (or decreased depending on player skill)

-Most conflicts would happen on "paper" only. The locations of Fleets and their properties would be recorded and the fates of systems would be often decided autonomously based on the above variables, and without actually simulating any real conflict unless a player was present. Instead, the forces of each side would be reduced according to the outcome. and the control of the system would pass to the victor.

-When a player goes to a system under conflict, then the game takes the basic data: e.g.
- -System 291y54182373hd
- Controlled by player X
- Facilities: 25, located at x locations on x planets
- Industrial output: 4.563, making warships and drones
- Defense forces: Large ships: 495, Smaller ships: 2355, drones....
- Fleets: 12315, 13562, 13563, 12346, 13535, 13563, 11246... are with in the player set radius of 23 lightyears and have been summoned to provide reinforcements.
- Attacking: Player Y's AI forces
- Fleets: 334, 352, 326, 236, 236, 27743, 2678....
When an actual player arrives, the game actually renders the units and shows the AI units fighting the battle. The player can then take control of individual units and direct the action if he/she wants to.

-As the MMO progresses, the maximum warp speeds of units will be gradually increased, but a limit is necessary: perhaps 0.05ly/sec (23 IRL DAYS to cross the galaxy) to start out. This limits the speed at which empires expand. As "technology progresses" this cap will be gradually raised to 1 ly/sec (27 IRL hours to cross the galaxy) and beyond.

-New players can choose to join the empire of a friend, or start their own from scratch out in the uninhabited fringes of the universe.

-Player empires will eventually come to encompass entire galaxies. Defeated players will be pushed outwards to start a new empire in a galaxy far far away.

-Limited speed of travel without instantaneous communication would make for a unique strategic experience. For example, in a large empire it may take days for news of an attack to reach the player by courier ship. In that time, the player's AI would be following it's previous instructions on how to deal with an attack.

-If the actual player's avatar is killed in battle, their entire empire would fall, or (if they have other players in their empire) the mantle of emperor would pass to their chosen successor and they could re-join as a new player.

-Just like a "real" advanced civilization, players would have the option to build megascale engineering objects like Death-stars and Dyson spheres to advance the cause of their empire.

Why do I think this is such a cool idea?
- We would actually be able to interact with the universe on a significant scale.
- Players would rise through the Kardashev scale with the ultimate goal of conquering as much of the universe as possible
- The universe would not seem empty, as empires would clash and because players could eventually control billions of stars at a time.
- The developers would allow the players to expand at rates dictated by computing power available to keep track of who controls what star systems.
- This would utilize the full power of Space Engine, and actually make every star and planet meaningful
- Players would have a combined first person and Real time strategy experience.
- Players could form alliances between their forces and carry out galactic and intergalactic diplomacy on a level that EVE can only dream of.
- This would make every other space game look like kids playing with toys.

I know it would be a lot of work, but I think the various units and functionality can be sufficiently optimized to make it possible, at least on the galactic scale for starters. I think this would be something more realistic, and more closely approximate what a real civilization would do, as opposed to the usual sci-fi baloney we get in other games. Even games like Elite Dangerous do not allow such immense ability for players to literally conquer and re-shape the universe!!! cool

Edited by PacificMaelstrom - Sunday, 26.10.2014, 01:23
HarbingerDawnDate: Sunday, 26.10.2014, 13:45 | Message # 2
Cosmic Curator
Group: Administrators
United States
Messages: 8709
Status: Offline
This is impossible. Activity in billions of systems would have to be processed simultaneously. No system could handle it.

On a personal level, I dislike the concept of the Kardashev scale, mostly because I think the scale itself is not a useful measure of the progress of a civilization. Also, if so many different levels of technology would need to be represented, this would require an exponential increase in the amount of modeling and texturing work, not to mention it would destroy any sense of balance. If someone attains Type III status and succeeds in occupying the entire Milky Way, then decides to crush anyone who shows up, that's game over. Nobody else could play ever. Which you seem to have addressed in the following quote:

Quote PacificMaelstrom ()
New players can choose to join the empire of a friend, or start their own from scratch out in the uninhabited fringes of the universe.

Which renders everything you wrote meaningless. If everybody is in a distant galaxy from each other, unless they're on the same team, then even with your "autonomous empires" no opposing players would ever encounter each other.

All forum users, please read this!
My SE mods and addons
Phenom II X6 1090T 3.2 GHz, 16 GB DDR3 RAM, GTX 970 3584 MB VRAM

Edited by HarbingerDawn - Sunday, 26.10.2014, 13:55
PacificMaelstromDate: Sunday, 26.10.2014, 20:08 | Message # 3
Group: Newbies
United States
Messages: 4
Status: Offline
Impossible? Maybe, but I don't think so.

It is not all that difficult for modern computers to keep track of billions of objects as long as each is sufficiently simple. Most of the activity could be managed by a simple log, with no computer resources devoted to actually simulating all that is going on. Instead, the information would only be translated into game objects when a player is present.

All of this could take place server-side, with one server keeping track of the star map and who owns what. All of this information could be extensively simplified. (depending on size of universe occupied by players, it would be initially be as little as a few terabytes of data)
- -AI units would be grouped in "fleets" which would travel in straight lines between systems
- -Stars would store information on ownership, fleets and facilities present, production levels etc...
- -All these units exist only as numbers until the player arrives, at which point they are procedurally generated according to the system logs.
- -Player computers could even be utilized to split the work of computing the log data in a "cloud" like fashion.
- -Overtime, as computing power increases, the available universe would increase

As for new players, maybe I didn't explain that very well, but the point is that the distance at which new players start would be dynamically calibrated to prevent them from encountering any overwhelming forces until they are ready. For example, Although he milky way and surrounding galaxies may be type III, farther out, new players would still be type II, but they would be sufficiently far away that the type III cannot reach them until they are able to develop to type III themselves. New players wouldn't be out in the middle of nowhere, they would be only far enough away to allow them time to build up before they encounter others.

Finally, we aren't really talking different technology levels, only different amounts of machines. A new player would have the same types of ships as advanced players, the advanced player would just have more of them. Even mega-scale objects could be constructed by new players... it would just take an unfeasible amount of time.

And last of all, the scale could be adjusted to reflect computing power. I said players "could" own billions of stars, but the concept would work just as well on a smaller level. If the game gets 1,000,000 players and they are confined to the milky way, that's an average of 400000 stars or less per player.

Edited by PacificMaelstrom - Sunday, 26.10.2014, 20:15
SpaceEngineerDate: Sunday, 26.10.2014, 21:04 | Message # 4
Author of Space Engine
Group: Administrators
Russian Federation
Messages: 4783
Status: Offline
Quote PacificMaelstrom ()

All of this could take place server-side, with one server keeping track of the star map and who owns what. All of this information could be extensively simplified. (depending on size of universe occupied by players, it would be initially be as little as a few terabytes of data)

As HarbingerDawn said, this is impossible. You need a supercomputer to compute a single galaxy even at star level. The only way is using some sort of "probability maps", i.e. the player empire is a 3D cloud of a "unit density per cubic light year". If a player clicks on a star, then the procedural engine will generate the actual system with actual units in it. Control over the empire will look like playing with a fog cloud, the player changes its shape and density. Control in a star system level is simply impossible! Not to mention control on a individual unit level. Of course, the player can take control over any unit, or take a closer look at a specific system, but all things should handled by the engine on a probability scale.

Interaction between two empires will look like a collision of dust clouds. They may pass through each other, but some systems may explode in conflict. Another issue is realism. It is hard to believe that fleets can capture an entire system within days of real time. The time of construction of a single drone must be many days, to say nothing of things such as Dyson spheres (which are unrealistic and have no purpose even in your scenario). Establishing a colony and building infrastructure will require months of real time.

The interface would be very complex and hard to play. 99% of the gameplay will be manipulating the 3D cloud in the galaxy map, and working in huge overloaded menus. Planets with detailed 3D landscapes and first-person mode controlling of the ships will be on the last places for players. Your scenario is better for a game with schematic graphics and a non real time scale. Like an intergalactic multiplayer "Spore" with advanced control over fleets based on a "probability density map" interface. It is a cool concept of course, but I think not for SE.


PacificMaelstromDate: Monday, 27.10.2014, 02:52 | Message # 5
Group: Newbies
United States
Messages: 4
Status: Offline
Thanks for considering, I guess the idea will have to wait for the computers of the future sad Todays supercomputer is tomorrows smartphone tongue

Your "cloud" description is a good idea. I knew there had to be a way to make it simpler like that. Actual units would only have to be "remembered" when they were interacted with or something. I was imagining something with the same seamless "scaling" that space engine has, where you can wage RTS galactic scale war, and move in to actually see the small scale conflict. Maybe you could point me in the direction of some resources that might help me build a small scale prototype. I have a little experience with unity.

(Small, with a radius less than the orbit of mercury for example) Dyson spheres could have use as a means of converting the output of a star into power for warp drives or death stars, but I see that power generation will not be dealt with in the game, so never mind. As for construction times, It seems plausible that in an established industrial system, advanced assembly techniques could churn out thousands of probes/robots/ships simultaneously, so although one probe might take a few days to build, new probes would be produced on the hour. The conflict over a system could take several days or even weeks, but that isn't really a problem, because, 1) there would be action going on. 2)lots of systems would be simultaneously being contested, so an advancing (endgame) empire could gain many systems every day. I don't think it would take all that long for Warp-capable ships to go through a system and blow everything up.

Your point about game-play experience is pretty valid though. It would certainly take away from the enjoyment of individual systems, and reduce everything to statistics. (today we lost 10000 star systems... oops)

On the other hand, the "real time scale" could mean that players would never actually get very far, meaning that the computational load would be significantly reduced, because each player would only have a star or three, and a smaller number of units. Players need not have the ability to conquer the galaxy... conquering a neighboring star system would be enough.

If nothing else, I think some sort of player-directed AI units would be awesome. It would be cool to send out von-Newman probes and stuff like that, and be able to build up large forces without (unrealistically) doing the all the work myself. Instead of implementing some sort of "background AI civilization", it would be awesome (and probably easier for you and more immersive for the player) if all the AI was actually player directed. AI need not be robots... colonists would work just as well.

Edited by PacificMaelstrom - Monday, 27.10.2014, 03:34
CesrateDate: Monday, 03.11.2014, 16:41 | Message # 6
Group: Users
Messages: 68
Status: Offline
I don't think autonomous exponential expansion is a good idea. The technical problem is obvious as discussed above - not possible with current computation resources, and, the strongest power would be those with most of the time just hanging online or the oldest inactive players. The exponential expansion may also result in a world which each (several) player(s) conquering a galaxy respectively then remain less interactive with others, which is similar to things happening in a star group - even better, as intergalatic travels are more difficult. Also, the game could be too "macro", then some of the players may not be satisfied. SpaceEngine could generate many interesting planets, but in "game version autonomous empire" those are no longer important.
However this idea could be a feasible concept among 2030. Hope to see how could that game be.
Forum » SpaceEngine » Gameplay Discussions » Autonomous Empires: The multiplayer soloution. (How to have meaningful multiplayer in an infinite universe)
Page 1 of 11