Shards vs. Shardless
I’ve been meaning to blog about this for a while. This article at MMORPG is the final push over the edge.
Most MMOs have gone with the shard/server approach. Each copy of the game has a distinct community. You make characters on a specific server, and the character can only be played on that server. Each server has its own copy of the world.
Champions Online eschewed this setup, and only has one “shard.” Instead, each zone is instanced depending on the number of people playing. The main advantage of this is that you never have to worry about whether a friend is on the same server.
But I think there are some pretty huge downsides, particularly in community. On other MMOs, each server developers its own sub-community of the game. Some tend to have more RP, or older players, or more hardcore raiders, or whatever. A culture evolves. Also, guilds and individual players can actually make a name for themselves. Becoming one of the best 10 weaponcrafters on a single server is not unattainable. But to be one of the top 10 weaponcrafters in the ENTIRE MMO might be.
One compromise might be a free server transfer setup that let people move characters whenever they wanted (perhaps with a 1 month cooldown). MMO companies have become enamored with that extra source of income (server transfer fees) so that’s doubtful.
So where are we? What is best? What do you like best?
I kinda like the new way. I haven’t really had any problems with Champions. In most cases after grouping with someone and having a good run through some missions, we usually add each other to our friends list.
One of the things that always irritated me with separate servers is friends going to the wrong server then finding out too late that I’m on a different server and not wanting to restart on a new server (them going to mine or me going to theirs).
With Cryptic’s STO coming next year, that will also use the one giant server set up. From what I’ve read they will allow a larger number of people per shard/instance. Personally, I don’t think it will be a problem.
While I see the appeal of shardless, as a player I’m much more of a fan of shard-based worlds. As you point out, each server can foster a specific community. Each shard can also have different rule sets to appeal to different people, something I don’t think current games have really handled very well other than a simple PvP switch.
From a developer’s point of view, you need to look at the design requirements of each. If you want to build a huge world/galaxy, then a shardless game might be a better fit, for example. Shards may also keep people from joining up with friends if they start at different times then want to meet up later. Then again, I think features like sidekicking/mentoring are better fixes for that.
Not sure about free server transfers. If you really do start making different rulesets for different shards, that could complicate things. It’s also a potential problem that EVERYONE might try to pile onto an already busy shard because “that’s where everyone is!” and therefore cause stability problems that will simply get blamed on the developer.
1) My experience in CO is that due to the one server, I never run into the same people. I seriously cannot remember a single time I happened to run into the same person on two different days. This was the first MMO where I ever had that experience. In other MMOs, one of the main ways I would make new friends is when “hey, that’s that guy I keep running into in various zones, and he actually isn’t a kill stealing douchebag. I’m going to strike up a conversation with him.”
2) Brian: Oh man, the servers with different rulesets is SUCH an underutilized feature. With RP servers being nothing more than very minimally name policed servers, basically all you have are PvP and non-PvP servers. And even then, most PvP ends up being instance based which makes that switch even more watered down.
I’d love to see WoW experiment with a few things like a full Bind on Equip server. I think the entire MMO world would see what a tragic failure of an idea widespread Bind on Pickup is.
3) Transfers: Yeah, with different rulesets, it could be an issue. But honestly, the whole “you leveled up on an easy PvE server and now you are on PvP. No fair” is ultimately not terribly important. If they are wimpy scrubs, they’ll just get owned anyway. Given a good enough timer, or even a minor $$ or in game cost, I think it could handle the big issue with shards: not being on the same shard as a friend.
I am a strong supporter of the huge instanced world model. Too tiny or fragmented has lots of drawbacks, too. Guild Wars for example, heavily instanced, I guess I do not have to discuss the pro’s and con’s of this system again.
Guild Wars 2 seems to have server-worlds but will allow easy transfer between servers. No fee and queues, waiting times and all that. ArenaNet are pretty vague how this will work in detail, but the general idea seems to take up the points Victor Watcher made in his article and combine the strengths of both systems.
The main advantage is that I can play with everyone playing this particular game. Maybe there will be language districts like in Guild Wars, but at least Psychochild, me and Muckbeast could play together without having to pay a server transfer.
The community argument and some servers being more friendly, more pvp, more raiding… it is true that many servers have a different feeling. How people talk to each other and all that. How much open world pvp there is. I know a lot of PvP servers that are fierce and brutal compared to PvP on some other servers of the same game.
Ultimately, I think Brian is absolutely right when he points out that shards or shardless also depends on the kind of game you make. But I personally prefer the one-shard model and/or easier transfers between server types.
There could be pvp “channels”, rp channels, or entire PvP/RP/or PvE “worlds” with subchannels… so yeah, I am all for shardless systems.
I prefer the single shard model like Champions and Eve, although I don’t like the instancing model that Guild Wars used. While I agree that games with different servers can develop different communities, its nearly impossible to determine what each server is like without playing on it. I also like not having to worry about meeting someone in real-life but not being able to play with them in-game because we’re on different servers (at least without paying for a server transfer or re-rolling).
I like them both for different reasons. I like the single world for its ability to make playing with friends easy and the tendency to foster a game-wide community (key for smaller games). I like the ability to just shuffle to a different “district” in GW to meet up with friends. I also like the ability to have different rulesets on different shards.
I don’t like server transfer fees, though. A time cooldown for skipping around I can see for traffic and technology’s sake, but fees for something that could be easily automated doesn’t sit well with me.
Hey there! Just found your blog the other week — can’t remember from where, but I have enjoyed what I’ve read so far.
Why not just have a central character server (it can be made up of multiple servers to handle load, of course) where all character information is stored, and then multiple independent Realm/Shard/World/Whatever servers where players can bounce in and out of at will? This would need certain logistics — like a defined cap of logged in players (which already happens in every MMO I’ve played in that results in queues) to prevent sudden masses of players logging into a Realm at once, but those are technical details.
I would imagine most players would stick to one server — much like a “home”, where their primary guild/friends/community are located. You could even build incentives for designating a “home” server, but this would not prevent you from visiting other servers whenever you felt like it — whether to visit friends who ended up somewhere else, some cool event going down, a different culture/community, or just a vacation — maybe your primary server is down but you still want to play with a particular character.
On a tangential note, this even opens up the idea of personalized “Realms” that change depending on player actions, offering a different experience or advantage/disadvantage, maybe with specialized “in-game” migration or application procedures. The possibilities are endless, but that is a whole other discussion.
I like your central character storage idea, and then login to whatever ruleset. I guess the downside would be if some rulesets are “easier” for leveling up. But then again, so what?
The Guild Wars example takes us back to looking at games that are “world simulators” and games that are well, games.
The MMOs that are more LPC/world simulator style are probably better off in the multi-server setup. That allows more people to be “important” or to take on various leadership roles. It also allows each world to develop its own unique culture.
In the more game-ish MMOs (Guild Wars, Eve, CO, etc.) the one server approach may be better.
Macil: WELCOME! I am glad you are liking what you are reading. Thank you!
I prefer shardless world simulators myself with minimal amounts of instancing. I find that communities still tend to form just based on region instead of server…with the huge difference being that you can uproot and moved when the current crowd of players in that area gets boring.
Server selection ends up feeling like a huge gamble, with months of time on the line if you choose poorly (most MMOs aren’t too good about allowing server transfers). Yes rulesets are a nice thing, but they don’t mean much if that server ends up underpopulated a few months after launch.
This been said, I do think that having several shards is better for traditional “progression” focused MMOs.
Serith, how would you have a shardless world simulator game with minimal instancing? I’m not sure how you could accomplish that without massive overcrowding. I agree that one huge virtual world with no instancing and one server would be awesome. But I don’t see how that would be practical as soon as you reach a certain level of population.
The closest game to that would be Eve Online. If you think about it, Eve basically does the same thing as CO. It is one server, but instead of 10 copies of Millenium City they just have 500 copies of “space, sector (aka instance) #2687″.
I may not have explained myself well. Obviously like EVE you’d have separate “cells” of space that loaded with some level of delay when players entered them. Key difference is unlike CO, I dont’ think EVE ever has identical copies of the same sector running. So unlike Age of Conan or CO if you go to a given location to look for someone they will be there, not in instance 12 of the same area.
Basically it comes down to how you define instancing…for me that has always meant multiple copies of the same area in memory. AKA the way WOW, AOC and CO handle most areas. EVE handles the player population issue by having a really really large universe as opposed to instances or multiple shards.
While you do not technically have copies of the exact same space in Eve, its really not all that different. I mean the majority of the game sectors are borderline auto-generated (and heck, they may have been originally according to some set of algorithms). That’s actually one of the benefits of making a sci-fi, spacefaring game. Outer space is incredibly easy to create. You code a few rules, and it can generate solar systems and open space to whatever extent and dimensions you want.
I’m not calling Eve an instanced world. I’m saying that in the end, its pretty similar in a lot of ways. But instead of 10 copies of Zone X, Eve goes with 10 sectors of space that are not really all that unique and definitely not special (and most likely not hand coded). And that’s fine, because for a game like Eve, they don’t need to be.
Aside from the possibility of creating different game rules/mechanics, shards (and instances) are really just crutches to deal with two things, 1) Technology limitations, and 2) Content limitations. In other words, nobody has a single server infrastructure robust enough to deal with 100,000+ simultaneous clients, so we split the clients out onto separate servers. Nor does any current MMO developer have the army of artists required to create content for a single, vast, unique, dynamic world.
So MMOGs must be parceled out via artificial means: shards, instances, phasing, etc. because there’s just not a better way to do it right now.
And on second thought, there are really three things that shards/instances are used for. The third being the apparent assumption that, for the sake of “fairness” or “fun” or whatever, every single player has to be provided with exactly the same content as every other player. Shards/instances make that easy — take X amount of content, duplicate ad nauseum. Everybody’s happy. (Well, except for me — personally, I don’t care or want to be forced to do the exact same stuff everyone else is doing…)
So, for me, the ideal MMOG would have a single world instance, large enough to hold 500,000+ players simultaneously, without crowding; with a variety of areas that support different play styles; and, most importantly, completely dynamic content.
Yeah I know, but I can dream, right? =)
Eh… Just realized I came late to this party, didn’t I. Ah well… at least some of my thoughts are “on paper” now. =)
I don’t think shards are artificial limitations.
Did you read the part in the original post about community? In playing CO, I noticed a complete lack of community. I’m not blaming that all on the lack of shards, but I think it is a factor. When you don’t bump into the same people repeatedly, you don’t get a sense of feeling like you are in the same world as specific other people.
Also, when you are all lumped together, only one guild can be #1. Heck, only 10 can be in the top 10. Only one guy can be the “top weaponsmith”, or more realistically, only one small group of people can be the top crafters. When you break the game into servers, you give a lot more people the opportunity to make names for themselves.
[...] pointed out by Muckbeast, multiple shards allow a greater number of people to be “the best” at specific [...]