MMO long tails
There’s been interesting discussion on the post about Argentum Online* and on the post about user-created content snobbery about the quality and quantity and potential popularity of user-created MMOs.
To recap, I commented that the barrier to entry for making your own MMO has been falling steadily. Given that just about any player can tell you exhaustively and in great detail exactly how their favorite MMO should be changed to make it 1000% better, this is bound to result in quite a lot more MMOs being created.
We have of course been here before. In the mud world, there were several different “server standards” so to speak — MUSH, MOO, LP, Diku, each available in multiple flavors. Of these various flavors, it was the Diku codebase that was the easiest to set up and resulted in the greatest proliferation of gameworlds. Why? Because it was template-based, and you could stage up a world by simply copying in some data. This led to the dreaded “stock mud syndrome,” where you could check a mud list and see endless Dikus with florid names all featuring exactly the same content.
The Diku gameplay went on to be the inspiration for EverQuest and World of Warcraft, of course, which is why so many mud vets say that they have played that sort of game to death.
Now, was most of this content crap? Yes, undoubtedly. But there were also a lot of really good Dikus, generally heavily customized. The lowering of the barrier to entry here not only created a somewhat horrifying landscape of zombie games shuffling along with the same Midgard starting city and the same bugs, but also essentially birthed the default MMO as we know it today.
At the time, of course, we all decried the endless parade of uninspired essentially identical muds. The commonest forms of “innovation” were to add munchkin-friendly stacks of levels, classes, and player races.
That’s a long and historically-oriented way of saying “we’ve seen piles of easy user-created content before, and most of it sucked, but not all.” But where does that lead us in today’s market?
Back in 2003 I did a presentation called “Small Worlds” based on scale-free network theory. Frankly, it’s a rather dense presentation that covers a wide variety of topics that are all loosely linked together, and the slides don’t really manage to convey everything that I was trying to get across.
One of the things in that presentation is the observation that the population distribution across MMOs tends to look like a classic “long tail” graph. This is a pattern that pops up again and again in all sorts of places. City sizes fall into this shape. Movie ticket sales. Book sales figures. Guild sizes. Popular posts on this site. Basically, anywhere that people have the choice to attach themselves to one choice over another, we see a tendency for popular things to become more popular. And then we get the phenomenon known as “the long tail,” which is that very long area that exists past the top end of the graph, where the curve falls off so slowly that there can be tens of thousands of nearly identically-sized elements in the graph.
One of the most fascinating things about this curve is that it tends to retain its shape. There is constant pressure to “smooth out the bumps.” The number two element in a graph “wants” to be at a specific ratio to the number one element, and so on. In fact, along the whole curve, each element wants to be at this logarithmic ratio to the elements above and below it. If a game starts to rise in popularity until it challenges #1, either it becomes lots more popular than #1, or #1 will fall to become way less popular than it used to be. Over time, market pressures push the games into this relationship.
This has been observed in city sizes in the U.S.; no matter which city was the largest city in the country, it was always around 2.1x the size of the next largest city. I suspect the same is probably true of MMO populations — whatever the size of the #1 game is, I bet that the #2 game in the same territory is probably some fixed ratio away in terms of population size. The shape of the curve never changes. You can probably predict the eventual size of a game by saying “will it be bigger than Game X? Smaller than Game Y?” Then just calculate that spot on the curve, and you’ll be in the ballpark of the eventual size of the game. As soon as we saw that WoW was going to be the #1 game in the West, it was pretty easy for me to say that it wasn’t just going to be #1, but that it was going to be larger by a significant margin, because that’s just “where the open slot falls.”
One implication of this, as with all power law distribution curves, is that no matter what you do there will always going to be a set number of winners and losers. And the typical world will be a loser. It will less population than the average world. As was discussed in the comment threads on the earlier posts, the typical world may well have only 5 users. And indeed, this is what we saw in the heyday of DikuMUDs as well.
This doesn’t, however, mean that the world must necessarily suck. Indeed, Chris Anderson argued in his book 
The Long Tail: Why the Future of Business Is Selling Less of More that one beauty of the items in that long tail is that they are very satisfying to a niche, whereas stuff at the head is moderately satisfying to lots of people. In the case of all these hypothetical user-created worlds, the people they are highly satisfying to may only be the few people who actually worked on making it. And to them, it may be just fine to have a small audience. As I said at GDC a few years ago,
The thing is that people want to express themselves, and they don’t really care that 99% of everything is crap, because they are positive that the 1% they made isn’t. Okay? And fundamentally, they get ecstatic as soon as five people see it, right?
But there’s even more to it than that. You see, as long as the network as a whole continues to grow, then a rising tide lifts all boats. The tail chunk slowly gets taller and longer. Even niche games start to grow. But if there are no niches — meaning, the games on offer are all pretty similar to one another — then the growth of the network can be capped. In effect, too many DikuMUD clones limits the total population of MMO players. People gravitate to the shiniest best one, and the tail starts to die off. The winner takes all, effectively monopolizing the audience.
The great promise of user-created MMOs in this case would be variety. The more varied the array of worlds on offer, the more online worlds in general will grow — even the clones. But if we all just start making class-and-level hack ‘n’ slash fantasy games, then there’s only one winner: World of Warcraft, or whoever supplants it. The tide of social networking always pulls towards the center.
In this sense, stuff the industry considers “fringe” is probably in many ways keeping the dominance of WoW from being even more impressive than it already is.
This all suggests that the real promise of user-created worlds may be less in the popularity or quality of individual worlds, and more in the fact that they unlock the variety that market pressures often work to prevent. We may need user content simply because it’ll be different from what we old-school folks would make — and the mere presence of difference means that our titles will also be more popular.
* y por cierto, si es que hay otros proyectos notables en el mundo hispanoparlante, aparte de Regnum que Daniel Benmergui ya me indico, me gustaria saber de ellos! Es obvio que mundos de diversas culturas tambien contribuyen a la diversidad de los mundos virtuales, y por tanto al crecimiento del genero entero.
Y ay, realmente tengo que instalar el teclado para espanol aqui en la oficina…
39 Responses to “MMO long tails”
Sorry, the comment form is closed at this time.
source:MMO long tails, Raphs Website Not exactly what I was looking for, nevertheless still a good read.
{Update: It’s great to see an entry like the one I came across on Kim Pallister’s blog yesterday (Link). I’d learned about Build-A-Bear Workshop sometime last Summer iirc [Note: as I thought…on Raph Koster’s blog – Link], and my thoughts then were very much the same as what he details in his post, and pretty much follow the kinds of things I’ve been discussing here in other entries. To be honest though, I’d forgotten about the Build-A-Bear store. Not too
Sounds like you’re talking about a time-based long tail curve, or what I was calling an “ecoToroid” (bottom of this old post – http://blog.rebang.com/?p=292 ). I’ve yet to read Anderson’s book and haven’t followed his blog in some time. Has he been continuing his time-based research?
Enjoying this whole debate. Thanks.
Interesting post, thanks for taking the time to break down more about the long tail as it relates to MMOs. I’m abundantly (painfully) aware of the Diku Clone prevalence, and given the awareness and nature of the MUDs that have managed to survive in the post-MMO, I think it’s still ringing true to this day.
It’s also pretty remarkable just how long the Diku game metaphor has managed to stand up and keep a strongarm on a number of the leading slots within that slope, both in MUDs and in commercial MMOs. Personally, I’m eagerly looking forward to seeing what comes of a truly open and available MMO framework for the masses. I think you’ll see a fair number of places that allow anyone who wants to “build”, but I think you’ll also see a number of the more popular amateur projects function more like a commercial development, with a “hired” group who volunteer their time and energy, and are recruited for the team based on their qualifications.
I’ve been preaching the “long tail” since I discovered that book. I’ve used the analogy that creating an mmorpg to rival WoW in the current market is like trying to start a retail business that rivals WalMart. You’re better off opening up a small store that targets a niche market, and targets it well. To use a user created content analogy “you’re better off opening a Build-A-Bear Workshop”.
[…] read from Raph Koster showed up today that should be heartening to us MMO developers. Read it at https://www.raphkoster.com/2007/05/29/mmo-long-tails/ It says in a sense that as it gets easier to make MMOs (with things like Multiverse), the smaller […]
The long-tail of MMORPGs is slightly different than the long-tail for music in books:
By design, MMORPGs are more fun when there are more players. As text MUDs have shown, if a player logs onto a MMORPG and finds it empty, they’re likely to log off and try another world.
If your MMORPG is on the tail, then not only won’t there be many players logging in every day, but if no one else is in the game when they log on, then they’ll log off very quickly. It’s a “positive” feedback cycle that results in low-population MMORPGs being completely empty.
To get around this problem, long-tail MMORPGs MUST provide quality single-player content to keep players around long enough that they see other players logging on, and defeat the feedback cycle.
There’s a host of other issues that I wrote in http://www.mxac.com.au/drt/PersonalVirtualWorlds.htm , and which I’ve pointed to before.
Which lends me to believe that there’ll be a fundamental difference in “what you do” between the top-5 games and the tail. Top-5 games will be about lots of people, such as raids and enormous battles, and being better than the joneses. The tail will be about niche communities, gameplay where friends log on as a group, and/or more of a single-player experience.
It may be that MMORPGs as we know them cease to exist.
You wrote: “The more varied the array of worlds on offer, the more online worlds in general will grow — even the clones.”
This is also a consequence or implication of an organizational ecology viewpoint (Wikipedia has a decent article on this), where additional niche organizations increase the level of legitimation of the class of organization (i.e. MMOs) as a whole without significantly increasing competition (because they appeal to different niches).
Higher legitimation without higher competition leads to increased founding of new organizations in the population and decreased mortality of organizations (until competition rises enough to make it harder to found and to “survive”).
Work in organizational ecology has demonstrated similar phenomena with (non-MMO) organizations. The addition of niche organizations increases the legitimacy of the population of organizations as a whole (which encourages founding new orgs) without increasing competition much (because the niches are distinct; this discourages organizational mortality).
That is exactly what we saw with the Tiny* derivatives and talkers back in the text days.
Re: Alan’s comments — yes, and it’s validated with network theory studies from a variety of other fields; the same phenomenon is common in all preferential-attachment networks ranging from businesses to low-level physics.
The spatial distribution of players within online worlds is often cited as following Zipf’s law, rather than other distributions.
It makes sense for the same distribution to apply to distribution of players among online games as well.
[…] has an interesting post on MMO long tails and how UGC unlocks niches — this works when it’s easy to move about […]
[…] the game becomes known as more popular, more people are just going to come in. You move up on the long tail chart (nobody will understand that pingback once they see it on Raph’s […]
[…] underdogs and in fact have a role as underdogs. Take a look at Raph Koster’s post just up today at https://www.raphkoster.com/2007/05/29/mmo-long-tails/ – it’s an interesting read on this topic.Posted at 1:01AM on May 30th 2007 by MG Midget function […]
Guess I should do some reading on organizational ecology. Sounds interesting.
Very well said, Raph. I only wonder about one thing, why is it, that the communities are full of players that hate “all new MMOs” and demand the good old EverQuest, Asheron’s Call and Ultima Online times back?
Personally, as a player, I loved EQ, and I’m still looking for a new game that offers the things EQ did right (though it has never been really determined what exactly this was). But all I see are “next generation” MMOs trying everything to be “new” and “different” and, more often than not, fail with that.
While I was long enough in the MUD space and see your point, there is also a different point of view. Looking at the absolute numbers of subscribers one can clearly see that EQ Style MMOs (EQ, EQ2, WoW, Lotro, LA2) are way more successful than every other. It’s not even a close match. The niche games are exactly the “long tail” of the equation.
Actually, if I had to put my money into a bet which MMO will be successful in the future, I would put it on a EQ type game that will be released in a finished and polished state. That’s what the most successful MMOs have been in the past and today. It’s out of question, that there will be a day when the players have enough of these games. But I don’t see niche games taking over soon, and there is still a huge demand for the one game bringing back the feel of the first generation MMOs.
Currently I see an explosion of social worlds and Web 2.0 connections coming, based on nothing more but Second Life, which in itself is already on the “long tail”. I highly doubt many of them will bring in their money.
It may look good to say not to compete with WalMart but go to a niche. But don’t forget that the niche costs nearly the same money than to run WalMart. Looking for player created content may seem like a good idea then. But I highly doubt it will work out. Player created content is a great thing, but I don’t think it will head in the direction where niche games would like to see it. There has ever been player created content in all MMOs and there will ever be. But it’s not about creating a game, it’s about creating content by playing the game.
So a niche MMO should better stay away from counting on the players to build the game and start thinking about how to make it fun to support the community. (And no, players with GM rights doing events isn’t the right answer either)
There are ways for it. Though nobody tested it so far. But all needs a finished (by developers) game first.
[…] underdogs and in fact have a role as underdogs. Take a look at Raph Koster’s post just up today at https://www.raphkoster.com/2007/05/29/mmo-long-tails/ – it’s an interesting read on this topic.Posted at 1:01AM on May 30th 2007 by MG Midget13. However […]
[…] Koster posted today about the ‘long tail’ phenomenon as it relates to MMOs, obviously one of his areas of speciality. This is an idea which has been bouncing around for a […]
This isnt really about working or failing, I read the entry as a general description of the online gaming industry. If you can use this type of knowledge to develop a profitable game then thats good for you.
My approach to figuring out that WoW was going to be much bigger than the current market leader back in 2004 was a lot more mundane. A clean matter of listening to the word on the street. For every gamer who fondly spoke of any current MMORPG there were 10 wanting to talk about the “imminent” arrival of WoW. This was a consistent ratio everywhere, on the internet, on the street, in peoples home, in the stores and at the parties.
Anyhow, user created content is not to my knowledge something that is defined. Each product which allows the user to create content will carry its own definition. UO allow the users to create habitats, Photoshop makes pictures etc. WoW allows guilds.
@Alan Schwartz
The wikipedia article on organizational ecology has me listed as a ‘prominant’ ecologist so it must be true 😉
I would add that there’s been some interesting work lately linking organizational ecology to networks [albeit inter-organizational networks]. There are, as you might guess, some strong parallels between the long tail Raph is talking about and network position vis-a-vis core-periphery. Degree distributions frequently follow a power law….
Don’t be blinded by the prominence of those worlds in gamer circles: in practice, a UO-style game is one of the largest in the world (Runescape), and social worlds with pets and housing and little else are crowding the top of the Western charts. In fact, of the top five worlds in the West, I think only one of them — WoW — is an EQ type game.
It’s a similar thing to the dismissal of roleplaying. We’re told over and over again that roleplaying is a niche activity. But then we see Gaia Online, with the second most popular message boards on the entire Internet — devoted entirely to roleplaying. You read ’em and they read just like the SWG Roleplay forum did. What is revealed is that in fact the roleplaying isn’t popular with gamers but is plenty popular on an objective scale.
I don’t think that this is the case even now. The current EQ-style bar is set by WoW, LOTRO, EQ2, Vanguard — that’s tens of millions in investment. You can stage up a game for the niche in under $1m.
But then again, you might need to further quantify that statement with “online” gamers.
I’ve told this story often over the years.
Back in the day, I player D+D regularly with about 30 people. When UO was in beta, not one of them were interested because it was online. Most of them had had bad experiances in playing games over the internet with cheating and jerk “gamers”. Those who didn’t knew enough of that to cause them to stay away.
Some of them might be playing now, I don’t know. But I can’t help wondering if, if this industry can ever get a handle on cheats and RMT and anything else that causes people to “not bother”, if there isn’t a boatload of new prospects waiting in the wings.
Hell, these “mainstream” games are tailored to just the kind of gamer that keeps the rest away. Static spawns/RMT, power grinds, one hit kills, all the stuff that runs right up their alley. Little wonder if I’m right, and there’s a huge number of would be online players who won’t.
Roleplaying was exactly my point when I said MMOs are already full of player created content.
Just take any MMO ingame chat. That’s a lot of player created content. All player characters are, emotes, roleplaying, drinking with others in a tavern, having fun hunting with a group, meeting someone in the wild, all guilds and even their out of game activities, even raids are player created content. And they create tons of it on the fly in any MMO while playing it.
That’s the direction I see for player created content. Much much more of this. So far we have just seen the tip of the iceberg. SWG showed player run cantinas and entertainers. Wish showed townships. EverQuest getting drunk in taverns or having a big party with lanterns in the darkness.
There is so many more to discover. And it’s so much fun to “create” content. And everybody can. Just somehow this part seems to be widely unnoticed by developers thinking about “player created content”.
Who/what is this “industry” you speak of, and why do you refer to it as if it has an opinion that means anything? 😉
Very good article.
But I think that comparing the MMO long-tail phenomenon to the music long-tail phenomeon (for instance) is a bit of a farce.
This is partly because, as you yourself would admit, MMOs are more of a “service” than a “product.”
But there’s more to it than that. MMOs are a very unique form of entertainment because the value of the entertainment is (or should be) dependent on how many people are enjoying it.
If someone releases an album, it doesn’t matter if 5 people or 5 million people listen to it, it will always be the same album and always be just as good. But the quality of the same MMO is very different when it’s played by 5 people vs. 5 mil. (As someone who was only able to play WoW at 5am in the morning, believe me, that world is much less interesting when you’re the only conscious being in it.)
(Also note that the hypothetical album might have 5 fans who love it just fine; but if the MMO has only 5 players, it’s very unlikely that you’ll keep even those 5 players for long. Much as they love the world, it’s pointless without people in it.)
What does this mean? If you’re expecting your game(s) to end up on the long tail, make sure their game design is such that they’re fun even with a small number of people playing… every MMO needs a certain critical mass of players to become fun, but you want to tweak your design to make that number as low as possible. If your game will only have 5 players, make sure they can all find each other, and aren’t playing their own separate games on opposite sides of the world.
So several folks are making the comparison to the music biz’ long tail. But the example in the article is actually that of cities, not music. Cities are “services” in the same sense that a game is here. Yes, there’s a minimum population threshold required for viability, but there’s a ton of room for “small towns.” In the mud world, we saw plenty of worlds being viable at peak concurrent populations under 100 — and I bet we see the same thing today in the UO gray shard community. The big scalar factors are actually size of population relative to size of map, and what chat facilities are available.
Just a quick report of Latin American MMO developments as Raph Requested!
Regnum Online (http://www.regnumonline.com.ar) – Argentina
3D Fantasy multiplatform MMORPG RVR-PVP + Quest Driven PVE.
As one of the developers of the game, I’m pretty proud on this one 😉 , so please Raph, check it out and share your comments with us. It will be a great honor to have you log into our game.
Argentum Online (ao.alkon.com.ar) – Argentina
2D Fantasy MMORPG Skill based very much like early UO with PVP.
As one of the first of its class in Spanish and having low hardware requirements, it grew into a sensational phenomenon reaching thousands of players. Both its server and client code are open source opening doors to MODs, TCs and Enhancements, and also helping to get many young people involved in game development.
Argentum Online 2 is its 3D incarnation and it’s still in very early in development.
Taikodom (http://www.taikodom.com.br) – Brazil
3D Action Based MMOG in Space.
An impressive looking Space Simulator in a virtual world. Still early in development, the game is undergoing a second tech test. They got some interesting investment from IBM (both financial and technical resources)
Erinia: (http://www.erinia.com.br) – Brazil
3D Fantasy MMORPG –- PVE Oriented.
After a rushed release as a P2P game and trying to switch into a F2P game, the project seems to be having commercial difficulties due to high competition with other international mainstream MMOs in Brazil.
Now, my two cents on the subject at hand: Even if I do believe that it will be easier in the future to create this kind of worlds, I would like to point out that this has not been the case so far with Latin American developments. The industry here is in its early infancy so, lacking in experience and the resources, working on any project of this magnitude was very much like re-inventing the wheel.
Budgets of less than a 100 thousand Dollars and teams as small as 8 people were the norm, and that meant missed opportunities to innovate, specially taking into account that we had to meet certain player expectations as pirated versions of L2, WOW and Ragnarok are quite popular in the region
Although games like Regnum may be lacking in amount of features, and might not represent the next true original thing in the field, I do feel however that there is something fresh about it and I think its because it’s less overdeveloped in terms of content and feels more real and more alive.
I’m tired of MMOGs that end up looking like another VDATP –Virtual Disney’s Animatronics Theme Park- where to grind. It’s time to put the focus back into player interaction and gameplay systems and that’s what we are trying to accomplish within the limited scope of our resources.
I do believe we have tons of passion for games and potential in terms of development down here. Feel free to check us out.
This is very interesting. But I question the idea of some impersonal “force” at work that is always going to do this “levelling” that you posit, i.e. every time a game gets popular, it automatically rises to a threshold and then by lock-step, as you say, “There is constant pressure to “smooth out the bumps.” And “The number two element in a graph “wants” to be at a specific ratio to the number one element,” etc. It sounds like science — but I wonder if it really *is* science that always works this way.
For one, as long tail consumers become more aware, they might create more bumps that don’t rise really hugely and don’t smooth out. More social media might make more bumps in the long tail and those games with the 5 plays could each double to 5 or triple to 15 but never turn into WoWs — then what, they are then bumpier.
Templates might also suddenly become more diverse if some new technology or capacity suddenly appears to make it all easier, the way the Sims Online suddenly made it easier to make your own comic book with the family albums.
I guess I feel that while these theories are interesting, I can’t feel bound by them and I keep wondering what unforeseen thing might occur to interrupt the theory.
Well, Prokofy, something getting “bumpier” really means that the curve isn’t as smooth. And you can see in the graphs I gave that in practice, the curves never are perfect. They are always a bit irregular. But the overall distribution is still highly approximate to a Pareto or Zipf distribution.
There is no “impersonal force” — rather, it’s highly personal forces. “Preferential network attachment” can be translated as things like “going where my friends are” and “going where jobs are” and so on. People themselves, left to their own devices, sort into this sort of distribution curve. The slope of the curve will differ, but you can find it in the sizes of clumps of people at a party, the sizes of cities, the sales of books, the number of readers on a blog, the popularity of a given comment thread, and so on — just like bell curves, Fibonacci sequences, and so on tend to show up in nature all over the place too.
Basically, though, I am arguing that it’s the small stuff in the niches that is most likely to pop up the unforeseen thing you are wondering about. The stuff at the head of the curve is very unlikely to be a risk-taking project.
I think that the curve is pretty natural, since MMOs are strong word-of-mouth products. Moreover, established players will go where their friends are. Thus, winning one player often means winning multiple. It’s a system prone to landslides.
In some ways, the results could be even more dramatic with MMOs than with some other sorts of choices, because we’re not dealing with a closed system. Players are not only brought from other games — but also from outside of the current MMO population.
[…] este artigo: https://www.raphkoster.com/2007/05/29/mmo-long-tails/ e pelo menos este coment�rio: […]
The slope of the power law describing MMO industry structure can change over time. The slope of the power law describing given names has flattened a lot over the past two hundred years. There’s some evidence that website traffic has flattened over the past decade. If you’re trying to predict the market share of the top x games five or ten years from now, that may be a significant factor to consider.
[…] on the topic with a lot of very current data and additional sources and links listed.6. https://www.raphkoster.com/2007/05/29/mmo-long-tails/ – Another interesting blog which covers statistical data up to 2007 for a variety of […]
[…] ���������� ������� ��� – � ����� ������� � ������� �� ������ �������, �� ��������� 20 ��� �� ������������� ������������ ���������). ���� �� �������� ����� ����������� ������� – ���������� ����������� […]
[…] but I'd be suprised if more than 100 had anyone on them. Not sure if you've read Raph Koster's MMO long tails article, but he made some interesting observations about playerbases sizes on muds and MMOs. If […]
[…] of players will play a minority of muds. In fact Raph Koster wrote an interesting article about MMO long tails back in 2007, in which he discussed the same trend in MMORPGs. […]
[…] […]
[…] said, "big" is a relative term. Actually it's similar to Raph Koster's observations about MMO long tails. You disregard muds that have fewer than 30-50 players. You also play IRE muds, the smallest of […]
[…] last bit is the MMO tail, when a game has passed the point of being one of the big dogs in the room to a workhorse of a […]