Automatic Stock Market - Minority Game

I finally got a working model of the Minority game in R. I had worked on getting a working model in Excel a few years back but never was able to get it to work. Since then, I've moved to R to start working on a factor model for equities.

I took a short break from the factor model and decided to pick up the agent model in R this time - and I was able to get it to work.

Agent models are really interesting - i got interested in them from two different sources. I had been playing lots of Starcraft back in the early part of this decade before when I had time. It is a game that is based on building lots of units that have simple instructions and capabilities, and turning them loose on your opponent.

Then I had also run across a paper reviewed in the economist that I tracked down. I then found the book "A new kind of science" by Stephen Wolfram. This book goes into great detail about the world of very simple programs, and how the interaction of repeating this simple program is far more interesting than the program itself.

I have built a version of the Minority Game - a working version! I can train the model on real price data and move it forward into the future. I am going to see about forcing the forward movement and how that impacts the outcome. This is especially useful for trading - it forms a whole new set of quantitative tools and methods to look at the market.