Introducing VALVEconomics – my new blog with research notes on digital economies

AND NOW FOR SOMETHING DIFFERENT…Some of you may have read that I recently accepted the position of Economist-In-Residence at Valve Software; the Seattle-based company that makes multiple-player video games. How on earth did this decision come about? And how is it related to my work on the Euro Crisis and Game Theory? Click here for the answer. Is there something we can learn about economies from digital communities of video gamers? Click here to read my first analytical piece that sheds light on this question. More generally, you may follow my weekly posts in a special blog which I call VALVEconomics


  • Yanis,

    we can see that you have much to learn about doing business in the USA; they do not pay you to tell everyone else; they pay you to tell them, not us. 🙂

  • Prof. Yanis,

    Is there a way to reply to your posts on VALVEconomics blog? (I looked and looked again, but failed to see how) Or is it restricted to internal discussion perhaps?

    Anyways, very interesting first article. I don’t know if Team Fortress 2 “ecosystem” suffers from bot activity on trading, like some other massive multiplayer RPGs do. I wonder if the subject will ever cross your path, so to speak. After all, they are from the same breed as the famous HFT algos (in principle, at least).

  • It is very interesting.

    Parenthetically, I always had a hunch that fixing the Eurocrisis will come down to fixing a Merkel valve that got loose.

  • There does not seem to be a way to leave comments on the actual post on Valve’s site. Regardless, I thought this was fascinating to look at and a clear presentation for non-economists to understand. So thank you Yanis.

    With respect to your final question about using the data to guide user trading, it all depends on the goal. An economist “should” seek equilibrium, right? But Valve is in the business of selling software, and many other developers seek to sell their software through Steam. Their goal is to maximize sales. Valve has demonstrated (IMHO) that they have an interest in promoting long-term customer satisfaction, so they are also interested in maintaining happy customers.

    So I think better questions would be: do opportunities for arbitrage increase or hurt sales, and how does the presence of arbitrage opportunities affect customer happiness with the Steam platform? Then we could address whether a real time “trading guide” would assist in reaching those goals, or change the market patterns altogether.

    • One small but important point: Sale maximisation and maximising long term customer satisfaction are not one and the same thing. Valve is not into maximising sales or indeed profit. At least not in the short term. And since the long term is impossible to plan for, Valve is striving to find a sensible path between sales and player utility

  • Yani:

    What would happen if in your VALVE econometric model you introduced exogenous austerity?

    Since I am not familiar with the inner workings of your model, say for example, what would happen if you restricted the ability of players to transact freely by 10%, 15% or 20%?

    I am interested in the multiplier effect. If you introduced a negative delta in your units of measurement how does that affect the virtual community overall?

  • Very interesting. I wonder, has work like this been done before? Have any other gaming companies hired economists to analyse their data? This could be a very interesting field. Blizzard’s recent Diablo 3 now has a real money auction house where players can sell their in-game items for real money. I wonder how the rules governing the real money auction house differ from those governing the auction house run on in-game currency (gold). With Diablo 2, the original currency (again, gold) suffered from extreme inflation and the players actually switched to an in-game item ( ) to use as their currency, due to this devaluation of gold. I believe this has also stopped now (the game is very old) due to the increased value of high level ‘runes’.

    If Blizzard’s real money auction house turns out to be a huge success, it’s only a matter of time before other companies start their own.. Possibly even Valve? Who knows.

  • I’m unfamiliar with the game but wouldn’t the provision of near-perfect information to Steam traders lead to the volatility of algorithmic trading with potentially hugely disruptive impacts on the game and your servers?

  • Hello. I read your post on Valve Economics about arbitrage and equilibrium on the Valve Economics blog. Though it was fascinating and informative, there seemed to be two things that most people in the TF2 community take for granted but which you haven’t yet discovered. Well, you may know the first one already. Your choice of using ‘Laser gun’ as a relatively expensive example item made me think you don’t, but it may just have been a slightly surprising arbitrary choice. You’ll probably find stuff out soon enough in any case, but I figured laying out some basics in a concise way might save you a little time.

    Weapons, despite being the only items with an actual effect on gameplay, are the second least valuable type of item (the least valuable being crates). All players get random weapon drops at regular intervals, which makes it easy to quite quickly gather up all weapons one wants to use, resulting in the supply always exceeding the demand by far and prices remaining low. I can only think of two exceptions to this. Every new weapon is for obvious reasons very expensive a short period after being introduced to the game. There are also some weapons that were only available for a limited period of time (Fishcake, Big Kill, Iron Curtain etc.), and therefore exist in limited supply and are expensive. All of those do however share stats with other, freely available weapons. They’re therefore regarded as status symbols rather than weapons, and their market value follows the same rules as that of cosmetic items.

    Large parts of the TF2 economy does in fact have numéraires.The approximate price of all items can be measured in Scrap Metals, Reclaimed Metals, Refined Metals and Keys. There’s a continuously updated spreadsheet that practically all traders refer to. ( Though there are a lot of trades that don’t involve these items (hat for hat trades, for example), the people participating in such trades usually refer to those set prices to make sure they don’t get scammed.
    Mind you, there is definitely still an opportunity for arbitrage. There’s always a price margin depending on situational circumstances, something that you most likely know more about than I do. There are also a lot of players that don’t know the generally accepted values of their items due to being inexperienced or disinterested in trading, and quite a number of veteran traders seek out these players to scam them.

    There. Hope I saved a bit more of your time than I wasted. Oh, and here are two more sites relative to the TF2 economy which you might find interesting.

    I look forward to reading more of your stuff. Good luck!

  • Quick question regarding the graph showing the relative prices of metal, Bill’s Hats, and Earbuds – about halfway through, there is a noticeable decline in the prices of Bills and Buds before it returns to basically the same trend. On what dates did that happen?

    Also, will you ever make these data available or just publish your analysis? I’m entering a master’s program in economics in the fall and have plenty of free time to poke around with Stata. Thanks for writing this blog, I think the analysis of virtual economies is really interesting, especially those that aren’t directly correlated to real-life currency (ala Second Life).

  • That goal by Khedira was a cannonball!!!
    Freaking amazing!!!
    But i liked more the other one that went to the horizontal pole first from a distance.
    More natural. I do not remember the player.


    Only one thing i hated.
    Why the hell Merkel was there? Mean and audacious.
    Of all people that woman had no right to be there.
    She has blood on her hands.
    It spoiled my viewing and my TV cracked.

  • Hi Yanis,
    I really liked your blog post on arbitrage at valve economics.

    I’m a software engineer and i’d like your recommendation on a subject that lies in the intersection of programming and economics. I don’t know enough about economics to take the right steps, what should i start studying? the only economics background i have is from high school. any recommendations?

    I’m currently thinking along the lines of an economic model that has something to do with data and resource management in the cloud.

  • Hi Yanis,
    I really liked your blog post on arbitrage at valve economics.

    I’m a software engineer and i’d like your recommendation on a subject that lies in the intersection of programming and economics. I don’t know enough about economics to take the right steps, what should i start studying? the only economics background i have is from high school, any recommendations?

    I’m currently thinking along the lines of an economic model that has something to do with data and resource management in the cloud.

  • Great article but no way to leave comments, I would understand but you ask the reader for their thoughts!

    • With regards to your article on Valve. its like you say, too complex to try to price things centrally, just like the real world!

      Provide plenty of data so the market as individuals can make the right decision for them.

      It would be interesting to see if there would be a free market in banking in the game. ie will/are gamers be free to make loans of keys to other players at interest? At the very least it would provide extra liquidity to the arbitrageurs!

  • Great findings.
    But i have to wonder.Isnt this conclusion likely to affect the economy if the majority of the gamers takes it into consideration when trading?For example “how they expand when new items or sales are released”,is likely to make gamers refrain from trading even more, when there has been a long time since the last “iteam release”,and they will all have an interest in entering trade during the early days of new releases seeking more profitable trading deals.
    Isnt arbitrage potential different (maybe smaller?) once this happens?

  • Yiannis Varoufakis. Without wanting to sound presumptuous, WordPress has a feature in the discussion settings where you can automatically approve comments from users whose comments you have previously approved.

    It looks like you may approve comments manually, so I wasn’t sure if you were aware of this function.

    Again without wanting to sound presumptuous, I thought the feature might help if you were not aware of the feature, if you prefer to manually check comments I understand completely!!!!!

  • Hi, Yanis.
    I’ve just read you artice upon equilibrium in TF2. It’s a pity that there is no opportunity to reply directly to the post or write you an email. Hope it will be available soon.

    After reading the article i’ve got one question. I think that in equilibrium Rkh also equals to Rek/(1/Rhe)=Rek*Rhe. So to predict the equilibrium we must take the average of Rkh, Rek*Rhe and Rgk/Rgh, don’t we?

    Waiting with impatience for the next post in your blog. I would be happy, if it will concern the mechanisms those get the equilibrium reached, if they exist; how crafting, preferences of players, prices in the Mann Co. and items’ random appearance frequency affects the equilibrium.

    What about graphs in real time. I was always waiting for this. It will reduce some arbitrage when players know actual prices. But anyway, i can not suppose how it will affect the sales of Mann Co.

  • I’d like to attack you somewhat personally, Professor Varoufakis, if you don’t mind, and accuse you of hypocrisy.

    How can you square your praise of the enterprise and initiative of ValveSoftware – a product of free enterprise and liberal economic US culture – with your self-declared Marxism and your support for the viscerally anti-capitalist Syriza in the recent Greek elections? Do you think a project like ValveSoftware would have been possible under Syriza, whose economic spokesman, Yiannis Milios, has recently written books such as Class Theory and History, and Communism and Philosophy?

    I also note the irony of you taking up a post in the autumn at the Lyndon Johnson School of Public Affairs at the University of Texas. All Greeks know the role of Lyndon Johnson in bringing about the fall of Greek democracy in 1967 and ushering in the colonels’ regime. And will you try to sell Marxist economics, class struggle and overthrowing capitalism to the Texans as you – and many of your colleagues – have been selling to Greeks for the last 40 years, or is this rotten and destructive creed only good enough for Greeks? Will you advocate to Texans to break the law and not pay their taxes as you advocated in Greece when you expressed support for the ‘I won’t pay’ movement?

    I don’t mean to be offensive in these comments and I’m not condemning your choices of becoming involved with Valve or the Lyndon Johnson school – these, of course, are matters for you – but I am rather fed up of radical Greek academics – from the safety of their universities in the UK or US, i.e. the Belly of the Capitalist Beast – advancing all sorts of ridiculous and antediluvian proposals for Greece that they know their host societies would laugh out of court.

    • No software company is the product of free enterprise, good grief. The modern computer, its software, architecture and support systems were developed in the public sector. Even the private companies involved were on state subsidy.

      Just sad.

    • Why is it sad? Collective enterprises have always produced magnificent things: Pyramids, Parthenon, writing, arithmetic, Apollo 11…

    • By publicly exposing himself to debate, he inevitably has to evolve; like flowing water erodes the dense surface of hard rock. I sat through his recent public presentation at the LSE and did not see a Marxist; I saw an honest man describing the economic world as he sees it. Not too much different to my own view and no one will call me a Marxist. Moreover, of even greater importance, he attracted a very large audience of the young people that will have to pick up the pieces and put it all back together again. Again, I did not witness any tendency to try and lend credence to the past stupidities we saw so much of from, (for example), the old USSR. My advice can only relate to my own contact. Debate with him, by so doing, we will erode away the unnecessary to reveal the truth of our situation. That does not require any of us to agree with everything he tries to teach, (as no one has any need to agree with anything I say or write).

      By honest, open discussion and debate, the human race evolves away from rigid thinking based upon historical interludes and into a more civilised future.

    • Yanis: “Why is it sad? Collective enterprises have always produced magnificent things: Pyramids, Parthenon, writing, arithmetic, Apollo 11…”

      There is only one problem with any form of public project; the ultimate source of the tax income to pay for the project. Where Marxism went wrong is that it was always based upon the idea that the general public; outside of the central government institutions driving the public projects; were there to fund the process; come what may, at any and every cost to the general population; that the general population paid for the central government’s grand aspirations, (stemming from the people in power that were driving them); with their own freedom. That to create conditions for the central government to succeed; the ordinary people were enslaved and controlled; rather than freed to themselves prosper.

      That if ANY central authority tries to take more than the general population can deliver; that nation falls into becoming a feudal regime; rather than remain a free nation.

      Yes, there are great projects that have taken our human race beyond the horizon; yes they stand out as a success. But, everyone involved in them MUST understand that they can only stem from a nation where the people are left to be free; under free enterprise rules, in a free market economy; where they are free to succeed themselves alongside the centre.

      Centrally inspired projects must be placed upon a vast raft of free individuals; each working towards a successful and prosperous nation. One without the other always fails in the end. Once central government falls into the trap of spending more than is available in taxation; then they fall into the trap of believing the failure to pay for their grand visions is not their responsibility, but the failure of the people to deliver the money for the project and then they turn upon the people.

      Central government must first create a successful and prosperous, private, free enterprise economy. THEN, they can build upon that strong economic foundation, those major public projects. But not before, and certainly NOT upon the idea that somehow they have a right to impose feudal conditions to enslave the people to pay for their own central prosperity. For the result is an engorged centre, fat and totally submerged into the belief in their “RIGHTS” to remain in power. That they can take what they wish without any consequence to their position in the greater scheme of things. It was the latter belief that ultimately destroyed the Soviet Union, and it is the same lack of recognition of such responsibilities; that is destroying much that was good that built up here in the Western economies from the end of WW2.

      We must all work towards a prosperous free enterprise economy; Then and only then, can we return to the grand visions inspired by the people, but paid for from the general taxation of the people.

    • A quick reminder: Marx wrote in London, at the British Library, while receiving stipends from the surplus value of mill workers (exploited by his good friend and co-author Friedrich Engels. All challenges to the ideology of the metropolis come from the metropolis. Even Nelson Mandela and Mahatma Ghandi learned how to challenge the forces of evil in imperialism’s centres. As for Valve, it is an utterly non-capitalist enterprise. I shall be writing about this exclusively

  • Interesting, however behavioral analysis presupposes the ability to interpret threw -even- pattent variables, but necessitates what kind of codex to use?
    Because receiving tons of data which are inexplicable at first, the results of computation might be possibly hard to even find a specific border of classification or meaning under what we may perceive, decisions in a digital economy.
    It’s also vital by developing millions of scenarios which find relative application and take into consideration human preferences, explained by actions, how specifically those, under stress or pressure, taken in moments, give distinctive results.
    All this research is somehow related with artificial intelligence?

  • The article was an interesting read 🙂 I’d be interested in a future article on the types of data collected and what your ideal collected data would be – and how that could feed into your analysis. A kind of data point collection and analysis for dummies.

  • Very interesting post. Would question if indeed all items can be priced via barter reduction – namely new items with unique modifications that are available at first only via crate drops. It must be more challenging to pull out the latent price for these, without waiting for a rigorous secondary market pattern to emerge. Perhaps a window when econometric interrogation, for all its faults, is the right approach?

    Furthermore, providing real-time information on the inefficiency of the market would surely drive herd behaviour, speculation and increase volatility – a time to consider the bizarre equilibrium of disinformation?

  • If Gabe is reading these comments I’d like to share an idea.
    Valve could build a direct democracy platform using decentralized p2p methods used in Tribbler and Bitcoin. Such an undertaking is already being attempted by theglobalsquare but they are first trying to make it a p2p news platform and later a direct democracy one. I’m not sure they will be able to accomplish this task.
    Imagine if people could have a legitimate say on the laws. A list of laws globally voted on that governments couldn’t shut down.
    PS big fan of source engine.

  • Yanis,
    This analysis of digital economies is really interesting. You may also want to read this issue of IEEE Spectrum (It is an easy read, you don’t need to be an electrical engineer to understand it) Among the articles, there is one suggesting the harmonious co-existence of several currencies, all digital of course. I think it will tie nicely with what you have been doing so far.

  • Prof. Yanis,

    Nice introduction into the still evolving economy of digital items (at Valve). I have question speaking about data analysis. The data contained in the Valve databases must be huge. In what sort of of form does Valve make the data available? Do you receive a large multi variate data set or do you have to formulate a question which Valve translates into a data query with just the specific results.

    If you are allowed to work on large data sets you could take a look at using tableplots to make data visualization easier. Plus it can be much more informative than a simple linegraph.