Game Theory and Open Source

I have been planning to write up another Work In Progress which is how does Game Theory relate to Open Source. I have been procrastinating in writing this up , but the notion of linkage between Game Theory and Open Source is important in informing other Works In Progress.

I was inspired to get cracking by a very interesting program on BBC 2 last night called, “The Trap - What Happened to Our Dream of Freedom”. (It was so fascinating that my wife decided to wash the dishes instead.) The program discussed how Game Theory has affected the way that the public perceives public services and the duty of citizens to contribute to the common good. The background on Game Theory really got me going. After the series is over in two weeks time, it might worth its own blog.

I was introduced to the notion of Game Theory at a very early age since my father studied it in the early 1960s while studying for his Masters degree in Operations Research. He would use references to it frequently as i was growing up. I took a greater interest in the last couple of years after reading a Harvard Business Review paper on the relationship between Boeing and Airbus in a potential cooperation on a Super-jumbo. That cooperation devolved into a strategy of go it alone for Airbus and a dismissal by Boeing of the whole idea to go instead with a Hub-oriented Dreamliner. A Harvard Business School case study applies Game Theory to what happens to Microsoft in an environment where all other software is free.

Open source is a disruptive force that involves millions of people acting independently to use a free piece of software. Game Theory as a model is useful for describing these types of interactions. Game Theory was invented during World War II by John von Neumann, after whom the von Neumann architecture, the same architecture upon which most computers are built, is named. It was further developed by John Nash of Beautiful Mind fame who developed Game Theory in non-cooperative systems.

There are several types of games described by Game Theory that I intend to consider in the context of open source. My thoughts on these are still not fully formed, but they include:

• Symmetric and Asymmetric - this includes the Prisoner’s Dilemma - You screw me, I screw you. Anyone know if there is any relationship between Innovator’s Dilemma and the Prisoner’s Dilemma?
• Zero Sum and Non-Zero Sum - Is the market participation of open source a net positive or net negative in value creation?
• Simultaneous and Sequential - The transparency of the open source model introduces a Sequential element to the competitive model of open vs. closed source competition.
• Perfect vs. Imperfect Information - Ditto
• Infinitely Long Games - I have no time for that.

I don’t expect Game Theory to be a predictive model, but I do expect it to be a descriptive model explaining why certain things work the way they do in open source. We now have a lot of data from Linux, MySQL, JBoss and the various BitTorrent clients. With this data, I am hoping to be able to understand the following:

• Why do people download and contribute to open source?
• What license model is best due to their various freedoms and restrictions on distribution, attribution, contribution, and forking?
• What is the best pricing model for open source support?
• How do you balance the rapid expansion of open source with the desire to make money in the creation of that software?
• What happens to market conditions of the main competitors in an enterprise software market when an open source entrant disrupts the current equilibrium?
• How does an open source ecosystem stabilize around a duopoly or single dominant player?
• Does Game Theory provide a mechanism to protect the intellectual property of those who contribute while still maintaining free and open access?
• How is mass collaboration in and out of the enterprise affected by the principles of Game Theory?

Doing a Google search for “Open Source” and “Game Theory” after the program last night, I was surprised that not more has been written on the subject of how little is easily found. Marc Fleury is one of the few who has written anything and he really only made a throw away comment in attack of Oracle against their move to support RHEL. Most other references are to open source software that supports Game Theory.

Have you found any relevant references?

Perhaps Google can make this a project in their Summer of Code, but for economists and mathematicians.

Comments

[Lars Plougmann]

Why not a predictive model?

[John Newton]

A descriptive model is so much easier to create.

A predictive model may be possible if I were doing a PhD in this. However, there may not yet be enough data to create an _accurate_ predictive model.

In creating the original business plan, we created an adoption model based upon data from JBoss which is roughly accurate. At the moment our adoption rate is exceeding our expectations.

[Jason]

Im very interested in the subject and was interested in studying at a collegiate level. I'm having trouble finding "game Theory" offered as a course of study. Do you know of its offering or would i have to study something along the lines of, public policy or economics?

[L]

RE: John Newton's comment

Game theory is a branch of macroeconomics. It is of intermediate difficulty, so a general Econ degree, with a major in Games is possible. Post again if your interested in more.

[Mohammad Hossein Vaghef]

Hi, do you find any new resources for the subject? i want to write a paper about game theory and OSS for my OR class.

[r4]

Hi,

Nice article.......Quite interesting article....We can find algorithmic game theory ....Open source software is created mostly by interested individual contributors that take pride in the software they write. They also gain appreciation and respect from the software development community. Several large vendors such as Red Hat, Novell and start-ups like SpikeSource are attempting to provide open-source solutions that include pre-configured, easy-t0-install distribtutions and support for enterprises. But as Larry Ellison of Oracle recently pointed out they own little to no IP and other players can easily enter the market. Even if companies like IBM and Oracle have not entered the open-source market, the threat from new entrants due to low barriers of entry ruin the game for incumbents forcing them to keep prices low.
r4