Life In 19x19

As RJ says, the large majority of go "theory" is informal. But to try to get a more solid handle on what ideas appear and how to handle it, as usual, I start with AI games (close enough to perfect play, but far enough to have interesting variations). It is strangely confusing to not need clear general thinking to play go at good human level.

with kata1b60c320s637 game 749369 https://katagotraining.org/sgfplayer/ra ... es/749369/

I have written some lengthy essays in the sgf. I hope many of the principles I explain based on weak point theory represent the spirit of my investigations and are clear. Hopefully it can be turned into a more formal precise calculation at some point. I am probably too stupid for that. I'm still thinking in terms of if there are DeepBlue-like heuristics for Go. Such are easier to understand and probably more accurate than anything in chess due to simplicity of rules. The main difficulty as always is tactics especially for eyespace.

In this sort of theory:

easy: working out which weak points are more valuable for eyespace. understanding how direction of play affects the importance of weak points

medium: understanding how local changes cause some weak points to become bigger or smaller. understand how weak points can move around the group over time and why. how best to balance weak points against each other in attack and defence. a coherent strategy requires moves to work together so if weak points are disconnected, more skill is required to use the aji. understand that the defender of connected weak points has an advantage and how to play both sides of it. understand how to create and use double threats and their types (are the threats connected to existing group or does a new group need to be created?).

hard: giving precise values for different moves, or otherwise how to use the aji and in which order.