No, mostly I just had a whole bunch of vaguely related ideas to do with analyzing the graphs of games, but I really need to find some time to hammer out the details to get something I could actually publish.

I spent a lot of time while writing my thesis on machine learning thinking about the different approaches that have been taken to board games and most tend to involve some sort of analysis of the board, and it occurred to me that rather than studying the board it might make sense to study the graph that shows how those different board states are connected.

After all you could have two games that superficially from their boards and rules look completely different but have identical graphs. Now what if we could solve one of the games easily but the other game was hard, if the graphs are identical you could covert any game state between the two games, thus using the easy game to solve the hard game.

So then you can look at things at how to create an easy game which has the same graph as a hard game, obviously for something like chess this is hard, but you don't have to replicate the whole game graph, you can do it in chunks breaking the big graph into a series of smaller graphs and then attacking them.

There's all sort of other interesting things you can do with game graphs, for example you can start enumerating all games with a certain complexity (in terms of node and vertices) and ask questions like "are some game graphs fundamentally more difficult than others" and perhaps more importantly "why?". I think the area has a big potential for deepening our understanding of board games and also raising some interesting complexity and information theory problems.