Direction Of Play/ Discussion

Direction of Play left undefined?

Rich: I think the book makes it quite clear early on that all stones have power, and that this extends in one or more directions, depending on position and neighbouring stones. The direction of play is therefore the direction that makes best use of the power of stones already played. This may be too difficult or vague to grasp in many cases, but IMO it is defined.

Charles Unfortunately the 'power' concept in go is even less clear than influence, thickness - which cause us enough trouble here. Then K postulates that it is a 'vector' quantity. Then you have to optimize it, somehow, relative to existing stones ...

No, this is trying to define how you play go well, not give you a working concept. There are two attitudes I have found useful in dealing with the points the book raises:

• there is something here you have to learn to become strong, namely it is no good trying to play go well unless you want to become rigorous about working your stones to close to 100%;

and

• what you therefore do is to try to maximise some sort of coherence function, replacing the evaluation function idea that applies well to chess computing.

When you understand the first point it becomes much clearer why go improvement becomes harder at dan level - you can get less from good local pattern-matching. And the second point explains something (to me at least) about why good moves can look so odd (as one sees in Kajiwara's games).

By the way, I have probably brought this up before, but Cho Hun-hyeon is on the record as finding K's theories opaque. So in a sense I am contesting this on behalf of haengma, a rival concept, on which Cho is an acknowledged master.

Bill: Kajiwara is a good read, and stimulates thought. But he should not be taken as gospel.

Malweth: I think most books should not be - especially in the Go world. The benefit I find from "The Direction of Play" is not in the theory, but in the general idea. The partial game commentaries help instill the idea of direction and help you experiment with it in your own games. A good book as long as it's taken lightly - especially for SDK or low dan.

2012 contribution and discussion moved to Archaic/Positional Game