Tobold's Blog
Thursday, August 07, 2014
Population management

I fully agree with Jeromai that you don't need a million players in a MMORPG for it to feel populated. Having said that, I think we need to look at the problem a bit closer to explain why people are perceiving tumbleweeds in a shrinking game. The first factor here is personal. Yes, you only have a "Dunbar number" of a hundred people you know in that game; but if let's say half of the players of a game leave, there is a good chance that also about half of the people you know leave. It is that personal loss of 50 people that affects you more, logging into the game and finding nobody in guild chat, not whatever thousands of people left the game overall. The overall number can be of importance to bloggers and game journalists, because it somewhat determines the size of your audience. It is easier to discuss a game that millions of people play than to discuss a game with a population of 10,000.

The second factor of perceived decline is technical. Too many games still run with a fixed server model. If half of the players of a game leave, you remain at least for some time with the same number of servers as during the peak, with each server having half of the players. Games with mega-servers, which simply produce less copies of each zone when less players are around, feel less empty after a decline. Unfortunately server mergers have come to stand for an admission of failure for a MMORPG, so game companies don't do them as much as they should.

The third, and largely unknown factor is financial. MMORPGs have high fixed costs and low variable costs, so they are much more profitable with more players. When player numbers decline, the economics of the game pass two thresholds: One where the profit of the game becomes lower than the cost of capital, and a second where the profit of the game passes zero and the game actually makes a loss. Different companies bail out at different moments of that process, for example NCSoft killed City of Heroes / Villains when it was still making money, just not enough of it. Most of the time it is impossible to know how many players a game needs to remain profitable, the makers of WAR once said they needed half a million players for that. But if you make with much lower development cost, you can presumably run it with much fewer players. There are even examples of companies like Aventurine who make more money from government grants and subsidies for producing a game in a poor region than they make from actual players.

If we consider 100,000 players for a MMORPG the new normal, we could certainly design games which can live with that number financially. But both the social structures in a MMORPG and the technical infrastructure of many games don't deal well with decline and change of population. So we would need to think how to handle that better. Mega-servers certainly seem like a better approach on the technical side. I wonder if we could come up with something similar on the social side. Imagine for example cities with 50 houses in the virtual world, forming virtual cities. There could be social structures like elections for major and collaborative city projects for the inhabitants to work together, just like a guild. But the number of cities would be limited, and players who quit would lose their place, so their house would be quickly snatched up by somebody else. In such a model you end up with a guild-like social structure of a fixed and stable number of participants. That is just one corner-of-an-envelope design proposal, I'm sure there can be others and even better one.

I do believe that game design can have answers to deal with decline and negative network effects in MMORPGs. If we admit that half or more players will leave in the first three months, we can design the game to deal with that. Right now many MMORPGs still appear to be designed for a World of Warcraft like growth over 5 years, but that doesn't seem to be very realistic any more.

Agreed. The traditional MMO model isn't working very well any longer.

On the social side, I think we might want to look at the stable guild structures that remain in each MMO, for one possible angle.

Why do they exist and still continue?

I suspect there has to be a stable leader / center of focus, or a leadership committee that organizes events and activities and gives members reasons to log on daily or weekly to renew contact with each other.

GW2 has guild missions in place of the bog standard raids, which are an interesting flex size activity for a guild to accomplish.

(Though they still suffer from weaknesses as critical mass failure - less than 6 is impossible for some activities, and 6-12 will struggle at others, and lack of new content introduced - nearly a year now repeating the same @#!*& things.)

The stability of leadership is a keystone weakness that no one seems to have solved yet.

Once the leader (organizer/recruiter) loses interest, or that charismatic center disappears, the network that formed around it crumbles.

Saw this problem from MUDs on: even the guild I led in its heyday sort of wore down a few leader generations down from me.

No clue how to solve this really. Incentives for leadership? (But then people scramble for the rewards without actually taking on any of the roles and responsibilities.)

Somehow always influxing new blood to replace the burned out leaders? (Assuming the new blood are any good at all and won't cause excessive drama which then collapses the guild.. always a dice roll with leaders.)

Multiple guilds/organizations so people aren't tied down to one edifice which will rock their foundations and help them quit the game when it inevitably wears out?

Another interesting guild model which maybe holds some promise is a megaserver-like model for a big guild/community.

TTS is a GW2 community that maintains 150-300+ online player critical mass to do open world bosses that require 120-150+ players.

It basically pulls and cycles from a big group of 3000+ players who applied and joined the guild, plus friends and public who happen to be online and find their way onto the open Teamspeak.

One usually sees the same 50 odd regulars at specific times plus a big mix of random people drifting in and out to fill the gaps.

Despite all that, it's still reliant on having sufficient leaders to run the events, which they are so far making up by letting people apply and running interviews, creating a sort of uber guild within a guild.

Except with the social and game design pressure to organize stuff for others in order for all to be rewarded, instead of getting to become the A Team in traditional limited-size raids, which leads to closing off and insularity.

Maybe somehow, somewhere, someone has to design some kind of minigame for MMO leaders to progress and get rewards in, so that social structures can aggregate around them.
Megaservers lead to players who don't know each other. At Wow's peak, it had a large number of servers that were each relatively small, and player movement was fairly uncommon.

A large game with a world server leads to random groupings with people you'll never see again, and all the fail that brings with it.
Can servers be organised by zone, rather than by world? Or do accounts have to be world-orientated? (I don't play MMORPGs, so apologies if it's a stupid question).

If zone-orientated, traveling to another city would 'feel' busy (or empty) depending on how many people were there. And merging servers would become more logical / acceptable.

There's something to be said about how game companies model their social content as well. A 40v40 PvP area doesn't work with a population of 100. A high skill raid of 40 people doesn't work with 1000 either. High market games need a ton of people.

There are population thresholds required to make content viable outside of just ghost towns.
There are population thresholds required to make content viable outside of just ghost towns.

Good point. And it is one that might actually come to haunt Wildstar. If you design your game for the 1%, you need to make sure that 1% of your server population is still enough to make a 40-man raid with a complicated attunement viable.
Spiral Knights has a nice system: it opens new numbered instances of the town once a certain amount of players are in it. But you can freely change between instances, there are certain number ranges for different languages.

After a couple of days I found myself changing to a certain instance to hang out with the players there. Maybe that could work in other MMOs too.
My fantasy is an MMORPG that uses the megaserver model but spawns new zones at different numbers of people depending on what kind of zone they are. So the city zones would be allowed to get fairly crowned before new copies appeared but deep wilderness spawns would get new copies after relatively few people so they would always feel like wilderness no matter the overall game population.

To combat the anonymity of a megaserver, I'd use fancy algorithms to decide what copy of a zone you're put into. Preference would be for people on your friends list then people in your guild then people you'd ever grouped with (the game could track this pretty easily) then perhaps people you'd ever traded with then the great unwashed. Obviously this could be a complicated optimization problem but just getting a "good enough" approximation of the above would work far better than randomly dumping people into zone copies.

So with this system you could have a single megaserver and never worry about your total game population in terms of how players experience it.
#3 financial: it is more extreme since costs are not linear. it takes more servers and bandwidth and network admins with 10mm than 5mm customers. But it does not take twice as many, so the overhead cost per player is less for the 10mm game than 5mm.

Markets with high FC, low VC tend to have a single winner, a dominant company. So being niche or planning to surpass WoW are more reasonable financial plans than planing on being a third its size. (Not saying that catching WoW's first-mover advantage is practical.)


Game design matters in this. If you want a several nights-a-week fixed-sized, mostly fixed participant "2009 raid" group, then population turnover matters a lot. A lot of turnover in your Hearthstone playing group is far less disruptive than in your tabletop group.


P.S.: whatever folks say about Wildstar population now I think will get amped up when WoD ships. The people who love hard, attuned 40-person raids will not leave but my guess is a lot will.
I think with scalable cloud services, the fixed ongoing cost can be made far more flexible now, supporting a far wider range of populations.

The population of a game isn't actually very important. What impacts players is their perception of the population. At a certain level you can have very few players but if they all visit the same areas and they all share the same chat, the game can feel populated.
Post a Comment

Links to this post:

Create a Link

<< Home
Newer›  ‹Older

  Powered by Blogger   Free Page Rank Tool