Categories
computer games conferences games design GDC 2008 massively multiplayer web 2.0

GDC08: Virtual Greenspans: Running an MMOG Economy

Summary

Speaker: Eyjolfur Gudmundsson, CCP

I want a full-time economist working for MY company.

And: CCP staff should give more of the GDC talks, they’re good. And entertaining.

In the midst of a week of depressingly dumb comments (on the topic of economy: what possessed Matt Miller to argue against microtransactions because accountants like to see x million players times y dollar per month and find microtransactions unpredictable?), it was a joy to go to an intelligent, extremely well-informed, rational talk with valuable lessons for the future.

EDIT: photos now added inline; better quality images of almost the same graphs can be found in the official Eve Online newsletters (2007Q3 and 2007Q4)

Anyway, highlights:
– Eve now has almost 0.25 million subscribers and almost 0.5 million game-accounts
– Eve’s free-market, perfect (ish) economy only started working properly when they went over approximately 50,000 players per server: this suggests that sharded games may be doomed to never get good economies
– Trade is inevitable, as are all the side-effects such as inflation and arbitrage: your choice is whether to ignore them or try to use them to the improvement of the game
– The amount of money in your game-world MUST grow in-line with economic activity.
– An Economics Professor working on a major MMO officially stated that virtual items are real items, and gave an economic rationalization. I’m sure this has happened before, but I’m not sure if they noticed yet: sooner or later expect governments, lawyers, and accountants to be very happy indeed. Sob.

Personal extra thoughts:
– Economics is about people – just like socializing and competitive play: good to understand why these things are so important to MMOs
– Do players need EXACTLY the same amount of information that the developers have? (IMHO: probably) This is a new form of an old argument from the 1990’s: should RPG’s/MMO’s let players see their character stats? Nowadays it’s: should MMO’s let players know the metrics data that the designers are using to tweak the game?
– Dr. Gudmundsson needs to put his economics reports in a more web-friendly maner (PDF sucks, but XHTML would be good…)
– CCP needs to “get with the web” and have e.g. an “economics reports page” where you can download ALL the quarterly reports (oops, my bad – thanks to CrazyKinux for the link), as well as web-based streaming of price indices etc – even if massively time-delayed – I couldn’t find any such thing on google :(

Introduction

As a former economics professor, I’m unaccustomed to having so many people flocking into a lecture, so thank you for coming.

This talk is from myself and Sam Lewis from Cartoon Network, Dan Speed from CCP – they have the games industry background and I have the economics.

CCP: 300+ employees, and no one works there – “You simply live it!”. We work to the common goal of making EVE the best place online in the world.

Eve: 225,000 subscribers and groiwing, with 460,000 characters today, all on one single shard. And … it’s gorgeous.

Economics

Why is economics important for an MMO? Because it’s about people, and if you ever get pople coming together, you get trade, and trade begets an economy.

Principals: scarcity, economic behaviour – the individuals in the environment, and the conditions for a perfectly competitive market. I think these are the three most important principles for an MMO.

But also … fun. Economics and business is fun gameplay – the Stock Market is just one big game.

The functioning of your economy is fundamental to the success of your game, to keep the people playing together.

If you do not believe me, go to the original source, Richard Bartle (2004). [Adam: my typing isnt fast enough, not a perfect quote] “although it’s not hard to see why an economy design might fail, it’s much harder to find one that wouldn’t fail”.

Further support: Ed Castranova: “the very process of making a choice under scarcity is enjoyable” (2005).

Scarcity

Why not make an MMO where you can have everything? Real life is all about things you can’t have. Wouldn’t a perfect game give you everything in the game, absolutely everything you want, straight away?

For most people, the pleasure is in the journey, not the destination. The same is true for gameplay – we want some benefit from our hard work online, and we want to see it, we want it to be measurable.

If everything is free, the marginal utility is 0 (from economic theory). or, more simple to remember, the pleasure is in teh journey.

Behaviour

We have to remember the individuals – that you cannot make their decision, you have to allow them complete freedom of choice of what they will do.

Individuals will be evaluating options given the constraints, many not necessarily on the game world, all aspects of their time, money, etc. We call the results of these evaluations “prices”.

There are as many potential objectives as there are avatars, not just players, because as a given avatar, they have different constraints, even if desires were the same.

Market

[Adam: here he did a basic explanation of perfect competitive market]

The question becomes: do we really want a perfect competitive market?

Most MMOs have homogeneuos products – same name has exactly the same qualities. Even enhanced items honour this, because it’s a different item.(one of those)

Breaking the basic rules of economminsc often results in fun gameplay – hiding information, creating monopolies, and profit distribution held in concentrated few. Don’t people want to become the Bill Gates of their game world (not possible with a perfect economic market).

So … you really do want to break the rules. So … that’s why you want to hire me.

Advanced topics from economic theory that MMOs need to start dealing with these days:
Monetary theory, Trade theory (real world to virtual ), Labor theory (return from gameplay, benefit from extra minute spent online)

Eve history

Integrated the economic design with the core game design from start of production

High security / safe zone as centre of universe, where players start playing.

All resources are assymetrically distributed, forcing trade.

Four major trade areas have been chosen by the players, not by CCP, because it’s a sandbox game and those areas emerged from the bheaviours of the players.

Opted for player-driven and dynamic economy so that it would mould to the activities of the players. Controlled economny failed for Eastern Bloc, it will fail in your MMO.

Contracts in EVE are possible but not perfect, and the loan system too we could improve.

Here’s an example of why real assumptons sometimes fail: contract job, standard procedure is that you put up collateral that is at least the value of teh googs you’re moving. This comes down to the fact that trust is not something that was designed in Eve – it’s always easy to cheat the otehr player.

However, a small population cannot support an effective market.

So, all items INITIALLY were produced and sold both by players and by NPCs.

As population grew, items were withdrawn. Currently more than 90% of all items are produced and sold by plaeyrs. Benchmark at 30-000 to 50-000 players before the market starts to work – so … think how that works or not with sharded servers.

All major spaceships etc if there is no player making it at the moment then it simply is not available.

Example: blueprints and inventions

Higher blueprints were originally distributed through lotteries – do the quest but may or may not get BP at end.

This created monopoly for the production of certain tech 2 items. As intended!

We then introduced an invention process, so players with tech1 BPs could convert them at cost to tech2.

tech 1 vs tech 2 price changes

As soon as we added inventions, and players started upgrading BPs, the price of tech1 didn’t change but tech2 dropped by 40%.

Real-life economic theory perfectly predicted this happening.

Perfect information

Eve
– historical pricing information availalbe, but regional so you have to travel to find it all
– long term planning is possible

20 day average price, 5-day avg price, dots samples for average price that day, range of pricing that day, and the trade volume.

This was cloned from the NASDAQ reports [Adam: e.g. for Microsoft here].

WoW:
– short-term server information (48 hours max auction time)
– long-term planning very difficult

Both fit well for the function for which they’re intended. So … let intended game mechanics control the level of information available. Design your economy to go hand-in-hand with the game-design, don’t just take a “perfect” economy and try to force it on every game design.

Tritanium pricing.
– as soon as we allowed decimal changes in the ISK price,you saw fluctuations
– between 1.5 and 3 per unit
– years later we realised that in-game mechanics allowed you to refine an item and recieve at a specific price, and that that was capping the price for tritanium.
– those that read the newsletters [Adam: first one here] will know I also want to fix another example of accidental price-cap

Tritanium price graph

Supply Changes

Zydrine
– sarted at 4k isk
– then a new area opened up with an easier way to get zydrine from kills
– …so a fall in price
– until may/june 2007 when it bottomed out, so game designers tried out new drop tables and put it on the test server.
– didn’t announce anything, just put it on PTS
– players noticed, and within 2-3 days, prices on LIVE server had doubled.
– this is a good example of the perfect information leading to major price changes
– they then stabilized around 10% lower than that.

Demand changes

Torpedoes
– we made a game design change to make torpedos shorter range
– cruise missiles incresed in price by 2.5 x, currently still dropping, but around 1.8 times original
– torpedoes dropped by factor of 4, no increase yet

Metrics

Designing and running an MMO is all about metrics. But … so is economics.

Reports needed:
– total trade
– most traded items
– identifying cash transfers
– relative change in rank and magnitude of individual items

Inflation

To prevent inflation problems, youi need price indices.

We have four major price indices that we publish quarterly in the newsletter, and I hope we will eventually publish live price index information, but its just my dream right now although its possible.

When I joined, players believed there was inflation, so one of the first things I did was to start calculating inflation measures.

– mineral price index (MPI)
– primary producer price index (PPPI)- items that need minerals
– secondary producer price index (SPPI)- items that need PPPI
– consumer price index (CPI)- end items, spaceships etc

[Adam: frustratingly, CCP seems to have no URL for any of those, not even defining them, let alone giving historical data :(]

First thing we noticed: deflation in the MPI.

MPI blips

Three small blips, one for each expansion pack. So, although overall economy is defalting, when ltos of people play at expansion pack, you get a termporary blip which is what people notice.

In 2007 – overall 14% deflation due to dropped production costs due to more people producing and competing.

Think about early 80’s in airline travel in US. When that was all freed up, prices dropped by 50% or more over the next 3 years, and you started to see the low-fare airlines come in. Exactly the same thing we’re seeing in EVE: large influx of players causes drop.

But, prices have recently started to increase (last 2 months), so I’m watching this very carefully to see what’s going on and react if necessary.

GUP – Gross User Product

Used basic GDP measures, worked with the Helsinki institute of information technology to adapt to virtual worlds.

Eve GUP

At same time as subscription is growing hugely, inflation/deflation in reasonable limits, GUP is increasing slightly => the economy is healthy.

Money MUST grow inline with economic activity

Simple to say, but very difficult to implement in an MMO (in real life as well!)

Need a system in place that ietnifies accelerated flows of cash into the economny.
– changes in cash earnings per minute on-line
– by skills / levels

Need sinks so you have something you can tweak to change this

Must avoid reducing the perceived benefits of anything repetitive that earns – people build up expectations of the value and the reward, so taking actions can have extra effects if you upset the PEOPLE that drive the economy.

Inflation is not necessaruly a bad thing – if it affects everyone equally, and not redistriubting the wealth substantially, then it is fine.

Lessons learned

– Use economic theory, but adjust it to YOUR game design
– Constantly monitor the system – not enough just to make it well at first, must keep adapting and tweaking it

Future

Trying to have bigger game population than population of Iceland: i.e. we need to beat 350,000.

That will change the society greatly – and I want you to realise that MMOs *are* societies.

We’re making changes to provide support to the society as it grows in size.

Emerging society with it own social, pilitical and financial instituations.

In Iceland we’ve had democracy since 930 AD, so we’re hoping that Eve will grow and thrive with a democracy.

Still want to add Futures, but we will need much more advanced financial institutions, so that will take time to happen before futures are feasible.

Will MMOs become their own world, or will they be based on real-world institutions?

Linden turned to real-life instiatutions, last year they said no-one can be a bank in SL without being a bank IRL, to fix their in-game banking problems.

Value in a virtual world is real because the players are real.

[ARGH!]

[Adam: he also mentioned the ability to cash-out virtual items as real-world equivalents as a possible future direction. More like Magic:the Gathering’s ability to get real-world cards from the virtual world purchases, I think, than the simpler, more obvious, less powerful scheme of SL’s and Entropia’s taking out of cash.

Q: why must money grow in line with activity

If not, more people will have more money and cause extra inflation, and gradually shut-out the influx of newer players

Q: I play auctioneer in WoW. In chinese new year the prize of gold out of game doubled because the farmers were all viiting their family and not around to farm gold right then. Do you monitor the external forces and deal with them?

From the numbers that I have today it’s not substantially effecting the gameplay, yet.

Q: Do you consider OOG knowledge as a bug or a feature?

In EVE that’s anticipated and hoped-for – I cant say that officially, as its not my job, but knowing the guys that do I’d say it’s quite fun for us to see.

Q: Gold-dupe bug: how would you deal with that at the economic level?

Won’t talk about specifics. We have had cases like that. The key is making sure that your tools on CS side are good. Tracking things back, reversing things.

Essential to find out and react as SOON as possible. Spot it, flag it, track it back, and react.

We’re much less effected by these things because people are intelligent and don’t do too much, so not so bad, or they do too much and we see it very early and stop it very early.

Q: Would you subtract money from people who’d benefitted? Or would you knock the worst people and hope it smoothed out?

Yes. And I can’t say any more.

Q: Couldn’t these advanced financial systems be developed by players?

Yes, they could, but there’s an issue of trust. You need someone that the players can trust absolutely.

There are a few people that have built-up a very good reputation. On the forums they only argue with “read my record”. But that’s a very scarce resource (trusted players) in Eve – eve players are generally not trustable.

Q: Goldfarming – with economists on the team, it would seem that farmers are simply part of supply and demand. So, what is too far, when does it become damaging?

As an economist, I also believe that drug trades should be made free. So, although I might agree with you, not many players may necessarily agree with us.

You earn your status in blood, sweat, and tears – this is what hte players believe. Since they believe it, we support it – so we go after the farmers because that’s what a fundamental part of what they believe.

[Adam: seems to me he was saying that it’s a part of the fundamentals of the *society* they’ve built, since he believes that each MMO is a society, so they have a responsibility to intervene to help support the society, rather than because they like intervening in arbitrary bits of game design/gameplay (CCP infamously doesn’t interfere in an awful lot of player-player griefing)]

Q: do you actively control the amount of money that is ciruclating, and how do you add/remove?

Players mostly do the adding through doing missions.

We only monitor what is going out, and if not enough then we add additional sinks (which he didn’t outline).

If you have a very dynamic economy, then you will find that an NPC based system will tend to break. So you either have to tweak things over time or have extremely adaptable NPC systems.

7 replies on “GDC08: Virtual Greenspans: Running an MMOG Economy”

By the way, a big thank you Adan for such a well put together article. I’ve only quickly skimmed it (am at work) and it’s full of juicy information.

Thanks for all the details!

CrazyKinux

Great blog! This was an awesome read. It’s hilarious how often you see EvE in the news and in blogs like this! Just goes to show what a unique and amazing game they have made. If only it weren’t so boring. :P But hey, don’t get me wrong, I love to play EvE, and in fact I play it specifically because it is boring! (I love to do homework and study while I mine)

I don’t think Matt Miller was arguing against u-transactions, I think he was saying that it is scary to fund a large game on them alone. Nexon spends next to nothing on their game development.

Anyway, thanks for the great notes.

IMHI and IME, microtransactions are if anything *considerably* less scary to fund game development on than subscriptions, because they provide :

– much finer grained information on what is popular and what is not

– much faster (approx 300 times faster!) feedback on changes in the playerbase (takes 30 days for people to stop renewing subscriptions, takes one hour for them to stop purchasing)

– much lower barrier to purchase, so considerably easier to get people to start paying if you’re afraid about how you’re going to get the playerbase kickstarted

Comments are closed.