Consider a WOW-like MMORPG game. The only difference between it and WOW is that there is no limit on the number of players (Suppose that there is no limit on hardware performance and network bandwidth).

For such a game, developers can still open multiple servers for players to play.

My question is:

Assuming that the game can run steadily over a long period of time, will it eventually lead to the concentration of all players on the same server?

Whether the answer is yes or no, is there a corresponding economic, sociological and ecological model behind it?

(This may be an open question, I am not sure if I should ask questions here, but there seems to be no better place to ask. I hope that I can get some inspiration through discussion)



4 Answers 4


New answer

Due to the comments, I will assume that network latency is not a problem, and forget about the server side.

At the end it depends on whatever or not the developers facilitate (or enforce) player migration.

There are few reasons for the developers to want to keep multiple servers despite not having hardware or network limitations.

And yes, there are reason why people might want to stay on a server or move to a different one.

I am throwing a broad net here. Brevity eludes me.

On content and the level gap

There are features that affect player migration. One of them is whatever or not there is support to copy or move characters across servers. Being unable to move to a new server without bringing your character with you means that for some people migrating to another server has a psychological cost (you can't use your character). For other this is not a problem...

IN fact, for some people, to start over with friends (not bringing their old character) is a plus! That is, given a conventional MMORPG, having a large level gap between the characters means there is little content they can enjoy together.

Which brings me to the next feature: the ability to reset your character. If I cannot reset my character to start over with a friend, we might as well decide to start over in a different server.

By the way, another motivation to change to another server – aside from friends who are starting over – is because there is no new meaningful content for the player. There are no more quests, no more areas to explore, no more relevant items to collect. Perhaps some new stuff was added, but it is the kind of stuff that only new players would enjoy. A good example are quests that cannot be repeated getting an update. Until developers add new quests and additional content, at which point we might want to go back.

On social structures

(A) player (Player 1) can initially play in Server A. He play for one year and then quit the game. After two years later, he return to the game due to his friend (Player 2). At this time, he create a new account on Server B, because Player 2 is in Server B.

If the Player 2 is new, in theory Player 2 has less to lose in changing servers that Player 1. And I do not mean characters, but social structures.

MMORPGs have a social difficult curve. Go back to Bartle taxonomy of players. It makes sense to have features intended in attracting and keeping players of different kinds, and given that some players are there for the social experience, some features are going to be social... like what? contact lists, friends, groups, squads, teams, guilds, clans, alliances, etc... building and maintaining these social structures takes time and effort.

It could be that the old player is unable to enjoy the same content as the new player... yet, the old player could have access to a guild or clan that is not available in the other server. The old player does not only not want to lose the social structure... the new player would benefit from joining it.

Well, of course, if after a year the old player was not kicked out.

Social structures has advantages. It is not only finding people to play with. It is also access to know-how, and people to trade with.

A larger social structure is better... However, game developers we want to keep balance. Because of this we will introduce features that scale with social structures... as a result, we put pressure on those structures to keep players active (and we, game developers, want active players), which in turns means inactive players get kicked out of these social structures.

In fact, a social structure that cannot stay afloat, will likely dismantle.

In fact, if players leaving a social structure will make more likely that other players leave it. This is similar to how a few soldiers breaking an infantry line can precipitate a retreat. Or, you know, a financial crisis.

Player identity and sense of belonging

People like to belong to a team. Yet, people like to compete. Refer back to Bartle's work. Because of that, it makes sense to have teams that compete. These can emerge naturally, or artificially.

For example, we can have clans complete against each other as a game feature. Similarly, we can have people of different servers complete by unofficial means.

At clan level, we probably want to join one with people with which we have some commonality. Players can develop a sense of identity with the clan. Language is very important. So is time zone (who is online when). And so are interests and values. The same thing can happen at server level.

In fact, you will find that clan/guilds – once the game has a large enough player base – will advertise outside of the game. They want you to join their server so you can join their clan/guild.


Going beyond a conventional WoW-like MMORPG. We also need to consider player generated content. It can be a motivation to visit other servers.

Also, if the game has faction wars and limited area... a new server is "free" real state.

Another reason to go to a different server is to get away from "them". People are people, and there is spam, trolling, bullying, harassment, etc… That sense of identity people have, could motivate to attack others. If moderation tools are not enough, a close second to consider before leaving the game is to go to another server. In particular if you can move there with a group of friends in agreement.

Yes, you can have something resembling fascism in the server. The trolls are ruining the game. The new players are ruining the game. “They” are ruining the game. Not us. Let us make them want to leave the game.

Thus, we need good moderators. We need them around the clock. We need them to handle a huge number of people who speak different languages and come from different cultures. They have to deal with people who might bring their real conflicts into the game, and even their nationalistic views.

Does it not make sense to partition people, so that each moderator is responsible to handle a fraction of them? At a given time, because there ought to be rotation. It can make sense to move people to a different logical server.

For reference, a moderation technique that began in forums is making it so that only trolls and spammers can see trolls and spammers. This means that the quality of the forum is low for them, while it does not affect the rest. The result is that people are more likely to just leave the forum, while banning them would result in them creating a new account. I used to know this as being "Hellbound", however I think that terminology has fell out of favor.

Something similar can be done in games with cheaters. So that if you are a cheater you get to play with all the cheaters. Apex Legends announced they were going to do this, not sure if they did.


One other advantage of having multiple servers is to have different set of game rules. We might even – wait, no. I would encourage to – use a server as test bed for new features. This can help balancing before deploying to all servers. And there will be people who wants to test the new features first.

A bit on server...

I know I am staying away from the server side. However, have you considered logical vs physical servers? The thing is, as far as the client is concerned, a server could be anything from a process in a computer to a distributed system across thousand of computers.

On the other hand, changing from a server to another does not have to imply to go to different hardware.

The server will already have ways to partition people. One evident one is by their position in the world (for example, if you are in a map, you should be getting updates about the movement of people in another map). We could even consider each map a logical server.

Again, goith outside of the conventional MMORPG...

It could make marketing sense – in particular if servers are magical – to sell logical servers, which are nothing more and nothing less that logical partitions of the physical server. Think, instances. They might as well be virtual machine instances. And even let the buyer configure the server with their own game rules, perhaps even mods.

They can result in a more intimate experience for the buyers. They could be closer to good old single player RPGs... except, you know, with other people, people who the buyer invites to join.

For example, a company could be interested in buying one of these logical server for employees of our company. And we do not have to see, you know, “them” (for example people form a competitor company) in the server.

Now, of course, people are part of what you are selling with an MMORPG. It is hard to convince people to play an MMORPG that nobody plays. However, that does not mean that there is market for them. Which is why I think a game to be marketed like this would not be the conventional MMORPG, but one designed for a more intimate experience.

Addendum: Dungeons are a resource.

Some memories have come to me. Memories of rushing to get to a good leveling position in a dungeon. People in the clan would try to hold our ground by rotating people. We would funnel players as they log into the game, and have reserve tanks and healers to take turns... why? Because as soon as we leave the spot, the people of another clan would take it. And we would do the same thing.

Some days we could not find a spot. We would go from dungeon to dungeon searching something appropiate for our level... and end up killing random mobs in the overworld.

I can imagine that if the server were more crowded that could go on for weeks. And if that were the case, it could make migrating, as a clan, to another server an interesting proposition.

Original answer

Suppose that there is no limit on hardware performance and network bandwidth

You still have latency. "Bandwidth", is transfer rate... I mean, how much data can be send and received in a given period of time.

That says nothing of how much time a particular piece of data takes to reach its destination. That is, latency.

Why would there be bad latency? Because of the network. I will let Grace Hopper explain nanoseconds to you.

Now, bad latency in a game means that the response from the server takes time to reach the clients. a.k.a. Lag. It is, of course, not the only source of lag. Regardless, network latency causes lag.

So, you can have a magic server that have infinite storage space and execute instructions in virtually zero time, and the game still have lag.

Which means, that regardless of how magical the servers are, there is still value in having multiple servers distributed geographically around the world. And of course, everybody would connect to the one nearest to them.

There is of course, the chance that you will have multiple magic servers and you do load balance between them, which will prevent all uses to be put in a single one... yet, there is no need for load balancing if your servers are magical...

...except, well...

If we have a hub town with millions of characters moving around... perhaps your server is magical, but the client isn't. It is a good idea to – if not physically – logically separate entities so that the client does not have to render all that. Why? Because the hardware of the client is another source of lag.

Since, you have perfect magical server, they can decide exactly what each client should be able to see, so that clients does not have to expend time deciding what to draw...


(This may be an open question, I am not sure if I should ask questions here, but there seems to be no better place to ask. I hope that I can get some inspiration through discussion)

Welcome to Worldbuilding Stack Exchange.

In a world where servers are magical, the server can render for the clients.

is there a corresponding economic, sociological and ecological model behind it?


In this universe, computing power is unlimited. That does not mean it is free. We have said nothing of energy consumption.

I do not know if it holds in your fictional universe, but in ours there is a Maximum theoretical data density. Which means that your magical servers take an unbounded amount of physical space (I will assume they grow as needed).

Similarly, we have a theoretical minimum time. The plank time. Since we cannot really execute something in less time than a plank time... in order to execute more things in the same time, we need more CPUs.


In game

If we have a conventional MMORPG, there is a chat window. Just imagine that chat window with messages in all languages of the world going by at ridiculous speeds. Even if not separated by physical servers, it is a good idea to be able to filter by language.

There is also the problem of crowds. It is a good idea to have less channels to separate people. Not only because of performance (of the client), but because large crowds can be detrimental for the experience of the game.

Out of game

A server that can grow without bounds will certainly catch the attention of authorities. What if it grows out of a legal jurisdiction? Will different laws apply to different parts of the server?


Back to energy consumption. We do not have a growing server that could eat ecosystems... we need to feed the beast, and power plants have an impact on the environment.

I would need to actually do the computations – I'm not taking this answer seriously – to figure out if you need a Type I civilization to have these servers.

See also: Matrioshka brain.

Let us be honest, you do not really have the problem of having magical servers. All you need to know – as far as game development is concerned – is that network has latency.

Addendum: The part about a growing server is not entirely fictional. Companies could build larger and larger data centers. Except they cannot do it as fast as this answer could make you imagine. There constraint in resources. They need to buy land, the need to hire people, they need to buy hardware... the hardware needs to be manufactured, people needs to do maintenance to it... is the business good enough to support unbounded growth?

  • \$\begingroup\$ About latency, I said that "Suppose that there is no limit on hardware performance and network bandwidth", maybe I should add an additional assumption "Suppose there is no limit on hardware performance, network bandwidth and latency". \$\endgroup\$
    – chansey
    Commented Sep 16, 2019 at 10:25
  • \$\begingroup\$ The main purpose of this question is to observe the behavior of players, not server technology. \$\endgroup\$
    – chansey
    Commented Sep 16, 2019 at 10:26
  • \$\begingroup\$ For example, a player (Player 1) can initially play in Server A. He play for one year and then quit the game. After two years later, he return to the game due to his friend (Player 2). At this time, he create a new account on Server B, because Player 2 is in Server B. But you know that Player 2 has his own friends and this friends have friends too, .... So eventually all players maybe in the same server (perhaps hundreds of years later). I hope to find the theory behind this phenomenon. \$\endgroup\$
    – chansey
    Commented Sep 16, 2019 at 10:43

it all depends on your infrastructure. we never have an unlimited player on the same server as we have no unlimited server capability.

when you hear unlimited, it means big number of players but player number always depends on server responsibility, authority and game Logic. in some games server only receives the data and broadcasts it but in some other server should do processing and....

if you have a high power server you can connect all your clients to one server but you have to solve latency problem.

  • \$\begingroup\$ I know that in general game servers have a limit on the number of people due to technical limitations. But my problem is actually not about technology, but about the flow of the player population. \$\endgroup\$
    – chansey
    Commented Sep 16, 2019 at 7:54
  • \$\begingroup\$ @chansey can you explain more? what is you rpblem \$\endgroup\$
    – virtouso
    Commented Sep 16, 2019 at 8:04
  • \$\begingroup\$ This is not a technical issue, it may be a sociological, economic or ecological issue. For example, In the real world, people always want to live in big cities. Game design is an interdisciplinary subject, but I don't know much about sociology, economics or ecology. I hope to get some inspiration from the discussion, so that I can find related resources further. \$\endgroup\$
    – chansey
    Commented Sep 16, 2019 at 8:21
  • \$\begingroup\$ @chansey if you dont have a distinct question, you are offtopic. ithink you are looking for a digital life that you can join people in there. its possible but you need tricks and techniques \$\endgroup\$
    – virtouso
    Commented Sep 16, 2019 at 8:36

Most games discourage player movement between servers by making it impossible for players to transfer their progress. So there is usually very little movement of players between servers.

The main reason why player numbers on servers fluctuate are onboarding and offboarding players.

However, when a game runs for a very long time, player numbers usually go down overall. When there are less players, then the reasons why there are multiple servers disappear. So the developers usually try to find some way to consolidate servers in order to reduce running cost and concentrate the community for better gameplay. But doing a server merge or switching off a server altogether is a conscious decision by the developers, not a natural development.


The only reason I can think of to have all players consolidated on a single server given your theoretical super server is economic. It would not make sense for a developer to pay for multiple servers, whether they are physical or logical, if all players could fit on a single server. In this situation, the game developer would most likely heavily incentivize players with exp/drop boosts to either move their existing character over or create a new character on a specific server. Game devs currently use this strategy to help spread players to less populated servers.

There are very few reasons that players themselves would want to consolidate to a single server. Taking your example of a WoW like game, a single server would make the experience of playing the game awful. Since spawns are limited, it would take substantially longer for players to complete quests that require killing specific mobs or field bosses. Many positives that are associated with playing on a single server are negated with dungeon/raid finder in WoW since those allow cross server play for shorter queue times. From a players perspective, the only reason to switch servers would be to play with a friend or for some boost to progress speed. A single server would have significant impact on the end game and leveling systems of many popular MMOs/MMORPGs.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .