Network-centric Go

 

Well now, this is kind of interesting:

https://www.linkedin.com/pulse/how-networks-fight-jeff-cares

Basically, about 10 years ago the author came up with an idea to re-wire the Go board. The new Network-Centric Go board had the same number of nodes and links (381 nodes, 684 links), but were arranged according to the “preferential attachment” process that researchers have found in models of the internet and world-wide web. You can see from the diagram this kind of network has a very, very large number of lightly connected nodes, a moderate number of moderately connected nodes, and a very small number of very, very well connected nodes.

The first-move advantages are obvious: the move in each turn will be to occupy the point with the highest number of nodes connected to it, until all available points have the same number of nodes. The author claims to have fixed this and made it susceptible to analysis by reducing the board to about a quarter of its size (a 10×10 grid) and adopting a “decreasing clustering coefficient” strategy. (“Clustering Coefficient is a measure of local cohesion, representing the proportion of a node’s neighbors that are also neighbors of each other” – so it kind of sounds like he decided to play normal territorial Go).

Now he has an Indiegogo project to make this game playable online. And you can give him your money!

https://www.indiegogo.com/projects/kayanet#/