The Game of Go as a Complex Network

I wrote some code in Python to construct a Pagerank vector of moves in the game of Go. Georgeot and Giraud discuss this method in their paper but do not include any sample code, so I decided it would be fun to implement it myself and run it on some sample data from the K Go Server.

The code on my GitHub should be a good starting point if you’re looking to construct a network of moves in Go, although there are some differences in my implementation and the one used by Georgeot and Giraud – for example, I do not take into account all the symmetries of plaquettes that they do. Here is a comparison of the top ten moves for λ1.

Comments